<%BANNER%>

Transient performance of steam reformers in the context of automotive fuel cell system integration

University of Florida Institutional Repository

PAGE 1

TRANSIENT PERFORMANCE OF STEAM REFORMERS IN THE CONTEXT OF AUTOMOTIVE FUEL CELL SY STEM INTEGRATION By DANIEL AUGUSTO BETTS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Daniel Augusto Betts

PAGE 3

During my PhD studies a lot of happiness and tragedy occurred in my life. For this reason I would like to dedicate this wo rk, which represents the culmination of all these things, to the following individuals: Erica E. Carr-Betts, my wife (thank you for always supporting me. I love you very much); Carmen A. Carrington Betts, my mother (thank you for being strong and fighting the good fight. You have immense strength and courage); Claude D. Betts, my father (thank you for your reason and courage); Lydia C. Carrington, my grandmother (I am so sorry that we could not save you. Thank you for teaching me that I can); and to Matilda Eva Carr-Betts, my da ughter (thank you for bringing me so much joy).

PAGE 4

ACKNOWLEDGMENTS I would like to acknowledge the following individuals who directly or indirectly contributed to this dissertation. First, my wife Erica Eva Carr-Betts was instrumental in the elaboration of this study. She funded this endeavor, and maintained my sanity during difficult times. Her support and love had no bounds. My advisors (Dr. William Lear and Dr. Vernon Roan) were very patient with me. Their instruction and encouragement fueled the hope that an end was near, even when it was not. I am eternally grateful to them. Those who worked at the Ford Fuel Cell Laboratory are also responsible for this work coming to fruition. Special thanks are given to Dr. Timothy Simons and Ryotaro Honjo, who were always available to give me a hand. Dr. Paul Erickson, an alumni of UF and an old friend, was gracious enough to allow me to conduct experiments in his reformers at UC-Davis. His collaboration made this work possible. The US Department of Education and the UF Mechanical and Aerospace Engineering Department supported my research with a fellowship. Without this fellowship I would not have been able to obtain this education. I am grateful. I promise to repay this favor by dedicating part of my efforts to providing educational and professional opportunities to others. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES ............................................................................................................vii LIST OF FIGURES .........................................................................................................viii ABSTRACT .......................................................................................................................xi CHAPTER 1 INTRODUCTION........................................................................................................1 2 BACKGROUND..........................................................................................................7 Fuel Processing.............................................................................................................7 Steam Reforming...................................................................................................8 Partial Oxidation Reforming.................................................................................9 Autothermal Reforming.......................................................................................10 Transportation Fuel Cells...........................................................................................11 The Transportation Load Environment......................................................................12 CO and Hydrogen Concentration Effects on PEMFC performance...........................16 Methanol as Fuel.........................................................................................................19 3 LITERATURE REVIEW...........................................................................................23 Kinetics of the Catalytic Methanol Steam Reformer Reactor....................................24 Reformer Performance................................................................................................26 4 ANALYSIS.................................................................................................................29 Steam Reformer Model...............................................................................................29 Heat Transfer.......................................................................................................31 Continuity............................................................................................................35 Conservation of Specie........................................................................................38 Chemical Model..................................................................................................39 Finite Difference Solution..........................................................................................41 FEMLAB Solution......................................................................................................44 Steam Reformer Model Validation.............................................................................45 General Description.............................................................................................46 v

PAGE 6

UC-Davis Reformer Data....................................................................................49 Reformer Design.........................................................................................................57 Design Generalization................................................................................................57 5 RESULTS...................................................................................................................65 Transient Efficiency....................................................................................................65 Case 1 Infinite Reformer...................................................................................65 Case 2 The Medium Sized Reformer................................................................67 Case 3 The Short Reformer...............................................................................68 Carbon Monoxide Concentration...............................................................................70 Hydrogen Concentration.............................................................................................76 6 CONCLUSIONS........................................................................................................78 Steam Reformer Modeling.........................................................................................78 Steady-state Reformer Operation...............................................................................78 Unsteady Reformer Operation....................................................................................78 7 SUGGESTED FUTURE WORK...............................................................................80 APPENDIX A COPY OF FEMLAB SOLUTION REPORT.............................................................81 B PRELIMINARY LOAD FORECASTING STUDY..................................................99 C PHYSICAL DESCRIPTION OF THE UC-DAVIS REFORMER..........................104 General Description..................................................................................................104 Pumping Subassembly..............................................................................................106 Catalyst Bed Housing Subassemblies.......................................................................109 Condensing Unit Subassembly.................................................................................111 LIST OF REFERENCES.................................................................................................114 BIOGRAPHICAL SKETCH...........................................................................................116 vi

PAGE 7

LIST OF TABLES Table page 3-1. Test parameters for Nakagaki tests.............................................................................26 4-1. Estimated maximum times for vaporizer volume flow change..................................49 4-1. Values constants used to develop virtual UC-Davis reformer...................................52 4-2. Steady-state design of 60 kW reformer with an 11.9 mm radius and a catalyst bed density of 1983 kg/m 3 ..............................................................................................63 5-1. Case 1: Summary of the transient response of the infinite reformer to changes in fuel flow rate............................................................................................................66 5-2. Case 2: Summary of the transient response of the medium size reformer to changes in fuel flow rate..........................................................................................68 5-3. Case 3: Summary of the transient efficiency response of the short reformer to changes in fuel flow rate..........................................................................................69 C-1. UC-Davis reformer subassemblies..........................................................................104 vii

PAGE 8

LIST OF FIGURES Figure page 1-1. Types of fuel cells and fuel choice used for fuel cell vehicle prototypes.....................2 1-2. Energy density of different fuels on a lower heating value (LHV) basis compared to hydrogen.................................................................................................................3 2-1. Typical polarization curves of various types of fuel cells..........................................14 2-2. Driving load of a 30 ft phosphoric acid fuel cell bus operated at the University of Florida. The bus weighs is approximately 30,000 lbs..............................................14 2-3. Indirect methanol automotive fuel cell system power delivery schematic.................15 2-4. TBB-2 fuel cell bus methanol steam reformer and steam reformer burner temperatures during transient bus operation............................................................15 2-5. University of Florida bus indirect methanol fuel cell system.....................................16 2-6. Effect of hydrogen concentration and utilization on PEMFC performance...............18 2-7. Effect of CO poisoning on PEMFC performance......................................................19 2-8. Use of air bleed to recover PEMFC performance......................................................19 4-1. General schematic of modeled steam reformer reactor..............................................30 4-2. Control volume diagram.............................................................................................30 4-3. Convection-conduction ratio in the UC-Davis reformer fuel entry regions...............33 4-4. Calculated values of dT/dt for the UC-Davis reformer at the entrance region...........37 4-5. Measured temperature distributions of the UC-Davis steam reforming test rig operating at varying space velocities and varying conversion efficiencies..............37 4-6. Schematic representation chemical reaction model in the control volume................40 4-7. Equilibrium products of the steam reforming reaction (CH 3 OH + H 2 O aH 2 O + bCO 2 + cCO + dH 2 ) with respect to temperature.....................................................41 viii

PAGE 9

4-8. Separation of the reformer volume into annular elements with constant gas properties in space for the finite difference solution method used...........................42 4-9. Experimental steam reformer developed at UC Davis...............................................50 4-10. Example of CO spike in the reformate of the UC-Davis reformer after an increase in premix fuel flow (transient condition). Also shown is a decrease of the steady-state CO concentration in the reformate with increased premix fuel flow rate....................................................................................................................50 4-11. Reformate CO composition relationship to changes in temperature in the reformer catalyst bed................................................................................................51 4-12. CO and H2 concentrations in the reformate gas in the UC-Davis reformer.............51 4-13. Estimation of catalyst bed thermal conductivity using UC-Davis reformer data.....53 4-14. Virtual UC-Davis reformer steady-state CO composition with varying premix fuel flow rates...........................................................................................................54 4-15. Hydrocarbon signal output obtained during UC-Davis reformer run......................54 4-16. Virtual UC-Davis Reformer CO spike from 10 to 15ml/min premix flow rate change.......................................................................................................................55 4-17. Virtual UC-Davis reformer temperature distribution (color plot) and methanol concentration (contour plot) for premix flow rate equal to 5 ml/min......................56 4-18. Virtual UC-Davis reformer temperature distribution (color plot) and methanol concentration (contour plot) for premix flow rate equal to 30 ml/min....................56 4-19. Reformer Efficiency with varying catalyst effectiveness at varying fuel flow rates..........................................................................................................................62 5-1. Case 1: Infinite reformer operation when undergoing step variations in fuel flow rate............................................................................................................................66 5-2. Case 1: Reformer efficiency changes due to change of fuel flow rate for the infinite reformer.......................................................................................................67 5-3. Case 2: Transient efficiency of the medium size reformer under transient load changes.....................................................................................................................68 5-4. Case 2: Reformer efficiency changes due to change of fuel flow rate for the medium size reformer...............................................................................................69 5-5. Case 3: Transient efficiency of the short reformer under transient load changes......70 ix

PAGE 10

5-6. Case 3: Reformer efficiency changes due to change of fuel flow rate for the short reformer....................................................................................................................71 5-7. Infinite reformer transient CO concentrations for various for changes in power output........................................................................................................................73 5-8. Infinite reformer transient off-steady-state CO concentrations for various changes in reformer power output..........................................................................................73 5-9. Medium reformer transient CO concentrations for various changes in power output........................................................................................................................74 5-10. Medium reformer off-steady-state CO concentrations for various changes in reformer power output..............................................................................................74 5-11. Short reformer transient CO concentrations for various changes in power output..75 5-12. Short reformer off-steady-state CO concentration for various changes in reformer power output..............................................................................................75 5-13. Transient CO comparison of various size reformers operating between 40 and 50 kW hydrogen power output......................................................................................76 5-14. Comparison of transient changes in H 2 concentration induced by step changes in fuel flow corresponding to hydrogen yields from 10 to 60 kW (lower heating value based) for various size reformers....................................................................77 B-1. Transient load profile for TBB-2 under the University of Florida Fuel Cell Lab driving cycle data obtained on 4-2-2004..................................................................99 B-2. Power Change Index (PCI) for TBB-2 under UFFL driving cycle data obtained on 4-2-2004............................................................................................................100 B-3. Multi-layer Perceptron.............................................................................................101 B-4. The MLP performance as a 10 second PCI predictor..............................................103 C-1. General process diagram of the UC-Davis reformer...............................................105 C-2. Premix reservoir and gear pump..............................................................................107 C-3. Vaporizer design......................................................................................................107 C-4. Superheater design...................................................................................................111 C-5. The UC-Davis reactor design...................................................................................112 x

PAGE 11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TRANSIENT PERFORMANCE OF STEAM REFORMERS IN THE CONTEXT OF AUTOMOTIVE FUEL CELL SYSTEM INTEGRATION By Daniel Augusto Betts December 2005 Chair: William Lear Cochair: Vernon Roan Major Department: Mechanical and Aerospace Engineering Proton exchange membrane fuel cells (PEMFCs) and, to a lesser degree, phosphoric acid fuel cells (PAFCs) have been widely studied as possible replacements for transportation internal combustion engines (ICE). These fuel cells consume hydrogen as fuel, which is electrochemically oxidized through an acid electrolyte. Because of the low energy density and scarcity of elemental hydrogen, alternatives such as methanol and natural gas have been investigated as primary fuels for PEMFC and PAFC fuel cell systems. Of these fuels, methanol is the most easily reformed into a hydrogen-rich gas. The most efficient way of doing this is through catalytic steam reforming. Therefore a clear understanding of the performance of steam reformers may lead to better integration of these devices into fuel cell engines. In this dissertation the results of studies regarding the transient and steady-state performance of methanol steam reformers for automotive fuel cell system integration are provided. To power an automobile, a fuel cell system needs to be capable of changing xi

PAGE 12

power output rapidly, to adapt to changing driving loads. Although most fuel cell systems currently use batteries to reduce the required response time associated with the fuel cells and other balance of plant components (including the reformer), the ultimate goal may be to reduce the reliance of the system on batteries (which add cost, weight, and complexity to the system). This means that the fuel cell stack and the reformer must each be capable of changing power outputs quickly and efficiently. Fuel cell stack efficiency is intimately related to reformate gas composition (especially CO concentration). Since the reformer is upstream of the fuel cell stack, it has great influence on the overall power output and efficiency response of the fuel cell system. The results of this study are based on data obtained from an experimental reformer and numerical models. Non-dimensionalization of the governing equations derived for reformer model development resulted in identification of potential reformer similarity variables. Based on these reformer sizing and scaling theories were developed. Of particular importance was the further demonstration of the insufficiency of space velocity and aspect ratio as sole reformer similarity variables. Step changes in fuel flow into the experimental reformer produced transient CO concentration spikes. These spikes have also been identified in reformer literature. Through the use of the transient reformer model, potential physical mechanisms that cause these CO concentration spikes were identified. Studies in conversion efficiency and transient and steady-state reformate hydrogen concentration were also carried out. xii

PAGE 13

CHAPTER 1 INTRODUCTION Interest in fuel cell technology for transportation has led to an ever-increasing number of prototype fuel cell vehicle demonstrations in recent years. For example, 210 new prototype fuel cell vehicles were built in 2004. This is more than the cumulative number of fuel cell prototypes developed between 1959 and 2002 [1]. Enthusiasm for fuel cells centers on their high energy-conversion efficiencies, the cleanliness of its primary fuel (hydrogen), and the cleanliness of its operational by-products (water, electricity and heat). Yet commercialization of fuel cell vehicles faces major hurdles including design of low-cost and robust systems, proper water-management design for transient operation at various power levels and conditions, improvement of stack life, improvement of hydrogen storage technology, and improvement of hydrogen-production technology. Generally, hydrogen supplied to the fuel cell stack comes from one of two sources: on-board hydrogen storage, or a fuel processor. Currently, most fuel cell vehicle manufacturers are developing vehicles that use hydrogen as their primary fuel (Figure 1-1). On-board hydrogen storage is intended to simplify fuel cell engines and reduce costs. However, even with pure hydrogen, fuel cell systems are far from achieving operational goals, especially when the power output of the fuel cell system is cycled through time. In addition, hydrogen storage technology is far from achieving the energy density goals needed for widespread vehicle commercialization. Figure 1-2 compares the energy densities of common hydrogen-rich fuels. Even with the most advanced hydrogen 1

PAGE 14

2 energy-storage mechanisms, the energy density of hydrogen is far less than that of other fuels. Currently, hydrogen-fueled fuel cell vehicles have achieved a maximum driving range of approximately 300 km [1]. Major research efforts are now aiming to improve hydrogen storage capacity. These efforts include using carbon nanotubes as a hydrogen storage medium, and developing safe hydrogen tanks capable of operating at pressures of over 700 bars. Figure 1-1. Types of fuel cells and fuel choice used for fuel cell vehicle prototypes. This figure was published with the permission of its original authors [1]. Introduction of reformation technology into fuel cell vehicles has been proposed as a possible inexpensive way to solve the energy density problems associated with hydrogen [2]. Industry analysts have pointed out that alternative liquid fuels could make use of parts of the current global fuel delivery and storage infrastructure, potentially reducing the cost of the fuels. In addition, alternative liquidfuels could still continue to be used in internal combustion engine vehicles without major modifications. Of the fuels typically considered, methanol seems the most attractive because it is easily reformed at relatively low temperatures and can be produced from a wide range of feedstock. In terms of energy density, methanol has approximately half the energy density of gasoline, and methanol is approximately 1.45 times more energy dense than hydrogen stored in advanced hydride beds (Figure 1-2). It is important to point out that the process of steam

PAGE 15

3 and autothermal reformation includes the dissociation of water to produce hydrogen, therefore yielding greater hydrogen per carbon concentration than the simple methanol dissociation reaction. 0510152025303540Cetane (diesel)Octane (gasoline)heptanehexanepentanebutaneethanepropaneethanolmethanemethanolammonialiq. HydrogenHydrideWater Carbon HydrogenEnergy Density (MJ/L) Figure 1-2. Energy density of different fuels on a lower heating value (LHV) basis compared to hydrogen [3]. Using methanol as primary fuel for PEMFC vehicles is technically challenging. Methanol-derived reformate contains hydrogen diluted mostly in carbon dioxide, water, carbon monoxide, and methanol. Among the problems is that the dilution of hydrogen produces a drop in cell efficiency. This drop in cell performance can be measured as a drop in cell voltage (from V 1 to V 2 ) that corresponds to a drop in cell hydrogen partial pressure from (P 1 to P 2 ) as shown in Equation 1-1.

PAGE 16

4 1212lnPPCVV (1-1) The value of the constant C has been measured to range from 0.03 to 0.06 for PEMFCs. This drop in voltage leads to a drop in fuel cell efficiency and operational power density range. A positive aspect of reformation is that water vapor is among the diluents in the anode fuel stream. This reduces the need for anode gas humidification, which in hydrogen-fueled fuel cell systems can be very complicated. The reformate gas also contains carbon monoxide, which has a poisoning effect on low-temperature fuel cells 1 The PEMFCs have been poisoned with CO concentrations from 30 ppm. The equilibrium concentration of CO for the steam reforming reaction of methanol at 225 o C and 300 o C (the typical range of temperature for catalytic steam reforming) are 0.8 and 2.3 mole percent, which is beyond the limits of most current PEMFC technology tolerance. Various reformate gas clean-up schemes aim to reduce the concentration of CO in the anode fuel stream. Notable among these are preferential oxidizers and specialized membranes (generally palladium) capable of filtering CO and other gases from the reformate stream. Coupled with these technical issues is the need for controlling the fuel cell and the reformer when changes in load are required, in such a way that system efficiency drops are minimized and response speed is maximized. Figure 1-3 shows the recorded efficiencies of one of the 30ft Georgetown fuel cell buses (TBB-2) during a typical drive 1 CO poisoning in fuel cells is a reversible phenomenon. Small concentrations of oxygen can be used to reverse the CO poisoning effects. Other techniques include increasing cell current to promote the electrochemical water-gas shift reaction.

PAGE 17

5 cycle. 1 These buses use methanol as their fuel and use a steam reformer to produce hydrogen that is consumed by a phosphoric acid fuel cell (PAFC). The transient changes in system efficiency shown are caused by variations in anode hydrogen stoichiometric flow, changes in reformate hydrogen gas concentration, and changes in reformate carbon monoxide concentration. Temperature variations in the fuel cell stack may also contribute to these efficiency changes. The goal of this study is to provide a broader understanding of the role of the reformer in the dynamic performance of automotive load-following fuel cell systems, and provide a basis for reformer design and modeling. Analyses in this study are derived from a mixture of physical modeling results and experimental observations from which general trends in reformer steady-state and transient conversion, hydrogen concentration, and carbon monoxide concentration were derived. 1 The TBB-2 bus is a 30 ft phosphoric acid fuel cell bus housed at the University of Florida FORD fuel cell laboratory. This bus uses on-board methanol steam reformation for hydrogen production.

PAGE 18

6 0.10.20.30.40.50.60.735455565758595105115Run Time (min) FC System Efficiency Ratio (FC Power/Max Power) Stack EfficiencyMax Power = 50kWRate of Chemical Energy in the Premix flow was calculated using the lower heating value of MeOH in liquid form. Figure 1-3. Measured transient efficiency of the 30 ft Georgetown Fuel Cell Bus (TBB-2)

PAGE 19

CHAPTER 2 BACKGROUND In this chapter, several basic concepts essential to the general understanding of the analysis and results of this dissertation are provided. The chapter describes key fuel processing techniques and compares these techniques to each other. In addition, a short description of low temperature automotive fuel cells is provided. Also, general systems integration issues of indirect methanol fuel cell systems 1 are presented. In this section, data obtained from the Georgetown fuel cell test bed buses is used to show transient interactions between the fuel processor and the fuel cell in an actual system. Finally, a general discussion regarding methanol as a potential automotive fuel cell fuel is given. Fuel Processing Because of the difficulties of direct electrochemical oxidation of hydrocarbons, hydrocarbon fuels are normally processed (or reformed) to give a hydrogen-rich fuel mixture. Reforming plays an especially important role in low-temperature fuel cells, namely proton exchange membrane fuel cells (PEMFCs) and phosphoric acid fuel cells (PAFCs), since they are unable to oxidize any other fuel but hydrogen (H 2 ) 2 There are three basic reforming methods: steam reforming, partial oxidation reforming, and 1 Indirect methanol fuel cell systems refers to the use of methanol as the system fuel. The methanol in this systems use a reformer to produce hydrogen, which ultimately is the fuel cell fuel. The term indirect is used to differentiate these systems from direct methanol fuel cell systems, which are capable of oxidizing methanol in the fuel cell without the use of a reformer. 2 Direct methanol fuel cells (DMFCs) commonly use similar electrolyte material, namely Nafion, as PEMFCs. In this text the term PEMFC refers to hydrogen proton exchange membrane fuel cells and excludes DMFCs. 7

PAGE 20

8 autothermal reforming. This study concentrates on steam reforming since, of the aforementioned methods, it is the most efficient form of hydrogen production from hydrocarbon fuels. Steam Reforming Steam reforming is the reacting of a hydrocarbon with steam to produce a hydrogen-rich gas (reformate). Using methanol as an example feedstock, steam reforming can theoretically produce a maximum three moles of diatomic hydrogen gas for every mole of carbon involved in the reaction (3:1). Steam reforming processes can occur with or without a catalyst. 1 However, most low-temperature fuel cell systems that use steam reformation make use of catalysts in the reformer. Whereas non-catalytic steam reforming requires temperatures in excess of 1100 o C for simple hydrocarbons, catalytic steam reforming can occur at temperatures ranging from 200 o C to 700 o C for simple hydrocarbons. Nonetheless, non-catalytic steam reforming is still under investigation as a possible method of reducing the thermal mass and volume of reformers. Equation 2-1 shows the general, ideal steam reforming reaction for alcohols. COanlCOlanHamOaHOHCnml22222 (2-1) l, m, and n represent the number of moles of carbon, hydrogen and oxygen, respectively, in the hydrocarbon being reformed. The variable a represents the number of moles of water per mole of hydrocarbon reacted. Greater concentrations of water (a) in the reactants reduce the concentration of CO, and increase the quantity of CO 2 in the products. Moreover, the amount of hydrogen is 1 Non-catalytic steam reformation requires very high temperatures.

PAGE 21

9 increased with greater steam. The maximum concentration of water that can be reacted in the general steam reforming reaction is nla 2max (2-2) In practice, the catalytic steam reforming reaction is dependent on three-phase chemical kinetics with mechanisms that are complex and currently not well understood. Modern steam reforming catalysts can be designed to suppress the formation of CO and promote the formation of CO 2 The most commonly used methanol steam reformer catalysts in published literature are composed of copper-oxide and zinc-oxide (CuO/ZnO). As a general rule, the steam reformation of methanol across a CuO/ZnO catalyst is executed with 1 to 1.5 moles of water per mole of methanol. Increasing the H 2 O concentration to 1.5 has been found to reduce the CO content in the products while increasing the hydrogen yield when the reformation process is maintained in the range of 250 o C-300 o C. Although higher water concentrations in the fuel increases hydrogen yield and decreases reformate CO concentrations, these benefits may be counterbalanced by the energy penalty associated with water vaporization and superheating. Partial Oxidation Reforming Partial oxidation reforming is a process in which the fuel undergoes combustion in a fuel-rich environment. The elevated temperatures generated from this combustion breaks down the hydrocarbon fuel into a hydrogen rich gas. Assuming, ideally, that the reaction products are only hydrogen, carbon monoxide and carbon dioxide, the general partial oxidation reaction for an alcohol is given in Equation 2-3. OeHCOeanlCOelanHemaOOHCnml22222222 (2-3)

PAGE 22

10 To obtain the maximum possible amount of hydrogen in a partial oxidation reaction, the amount of water (e) formed must be negligible (e 0). Taking methanol as an example, the maximum, ideal, diatomic hydrogen to carbon ratio for the partial oxidation reaction is 2:1. This is 2/3 of the maximum possible obtainable hydrogen to carbon ratio from steam reforming of methanol. It is important to note that for most applications the use of pure oxygen for partial oxidation is impractical. Thus nitrogen will be present in the products of the reaction, which further dilutes the reformate hydrogen concentration. Some advantages of partial oxidation over steam reforming are that water is not required for the process, the equipment is more compact, it allows rapid start-up, the same reactor can reform various types of fuel, and the high-temperature reactor can tolerate many impurities (including sulfur). On the other hand, care must be taken to prevent carbon deposition. Also, since the process is exothermic, relatively high local temperatures are produced, causing possible materials problems. For integration with lower temperature fuel cells, a gas clean-up step is necessary to reduce the inherently large CO content. Autothermal Reforming Autothermal reforming combines aspects of steam reforming and partial oxidation reforming. A catalyst bed, typically nickel, and steam promote the reforming reaction. The necessary heat is generated through combustion of a portion of the reformer fuel. The general, idealized autothermal reforming reaction is given by Equation 2-4. CObanlCOlbanHbmObHaOOHCnml22222222 (2-4)

PAGE 23

11 In autothermal reforming, the maximum amount of water (b) that can be reacted is given by Equation 2-5. anlb22max (2-5) Compared to the ideal steam reforming reaction, the maximum amount of water that can be reformed is reduced by the presence of oxygen, thus reducing the overall hydrogen yield. On the other hand, the addition of water to the reaction allows for higher hydrogen yields and lower carbon monoxide concentrations in the products. Autothermal reformers have had a large amount of interest because of their capacity to reconcile between the benefits and detriments of using either catalytic steam reforming or partial oxidation reforming. Transportation Fuel Cells For transportation applications and small-scale residential power production, PEMFCs and PAFCs have been most widely considered and explored. These fuel cells operate similarly, oxidizing hydrogen through an electrolyte and a current collector as shown: Anode: H 2 4e +2H + Cathode: O 2 + 4e +2H + H 2 O However, PEMFCs operate at a lower temperature than PAFCs (25 o C-80 o C as opposed to 150 o C to 200 o C). In addition, PEMFCs, due to their high efficiency over a wide range of current densities, typically have higher energy density than PAFCs. Mainly for this reason, PEMFCs are widely considered the preferred automotive fuel cell. This distinction does not mean that PEMFCs are trouble-free. Major engineering challenges are yet to be resolved in order to commercialize automotive PEMFC stacks. Foremost,

PAGE 24

12 these fuel cells must dramatically improve their operational life; they must achieve consistency in performance under mass production, which has not yet been demonstrated; they must be designed to have effective humidity control and positive water balance during all operating regimes; and their cost must be dramatically reduced. In terms of life and humidity control, PAFCs have been demonstrated to far surpass PEMFCs. However, PAFCs are much more expensive to produce than PEMFCs and their cost is harder to reduce to target levels, even through mass production. Figure 2-1 shows typical polarization curves of various types of fuel cells including molten carbonate fuel cells (MFCs), solid oxide fuel cells (SOFCs), and alkaline fuel cell (AFCs). The reader might find it curious that AFCs are not considered in this study even though they exhibit much higher voltages than all other fuel cells. The reason for this is that AFCs do not tolerate air (due to the presence of CO 2 ). This limits their use to space applications and as electrolyzers. The Transportation Load Environment The power load associated with automobiles and buses is high and extremely transient. Vehicles are driven with a rapid succession of speed adjustments, especially in city traffic. Any fuel cell engine that is going to be incorporated into a vehicle must be able to change power delivery over a wide range of conditions fast and efficiently. Figure 2-2 shows driving loads obtained from a 30 ft fuel cell transit bus operated at the University of Florida (UF) on a low speed circuit around the city of Gainesville, Florida. The transient response of fuel cell systems is limited by their capacity to deliver fuel (hydrogen) and oxidant (oxygen) to the fuel cell stack. In a reformer based fuel cell system, as shown in Figure 2-3, the fuel processing subsystem must be capable of changing hydrogen production levels at the same rate as current draw changes occur in

PAGE 25

13 the fuel cell stack. In turn, fuel cell voltage at a certain current draw depends greatly on H 2 and CO partial pressures in the reformate gas. The process of steam reforming is complicated by the endothermicity associated with hydrogen production. The composition of the reformate gas is largely dependent on the temperature at which the reforming reaction takes place. Yet, in steam reformers the heat required to maintain the reforming reaction is provided by an external source and thus is limited by heat transfer considerations. This results in varying reforming reaction temperatures during steady-state and transient operation. As an example, the University of Florida fuel cell bus (TBB-2) reformer temperatures during transient conditions are shown in Figure 2-4. 1 In TBB-2 the reformer catalyst bed is heated via an external burner that consumes stack anode flue gas (Figure 2-5). The anode flue gas contains varying quantities of hydrogen, depending on stack current draw and anode stoichiometry. Anode stoichiometry is dependent upon fuel flow rate and reformate quality. Current-draw depends on the load environment. For TBB-2 the stack power draw and the reformer hydrogen production are coupled. University of Florida researchers have found that this interrelation between the stack current draw and reforming heating has led to instances of reformer overheating and under-heating, both of which have led to drops in overall system performance [2]. 1 Reformer top, center, and bottom refers to the position of thermocouples in different regions of the catalyst bed. The catalyst bed entrance is located close to the reformer bottom thermocouple. The center thermocouple is located close to the midpoint of the reformer and the top thermocouple is located close to the exit of the reformer. The reformer burner thermocouple is located close to the reformer burner flame.

PAGE 26

14 Figure 2-1. Typical polarization curves of various types of fuel cells. This figure was published with the permission of its original authors [4]. Transient Load Profile for TBB-2 under the UFFL driving cycle-10010203040500100200300400500600700Run Time (s)Power (kW)4-2-2004 test 1 Figure 2-2. Driving load of a 30 ft phosphoric acid fuel cell bus operated at the University of Florida. The bus weighs is approximately 30,000 lbs.

PAGE 27

15 Methanol:WaterPremix TankSteam ReformerPhosphoric AcidFuel Cell StackPowerConditioning Batteries Parasitic andAux. Loads High VoltageDistribution TractionDrive Wheels Figure 2-3. Indirect methanol automotive fuel cell system power delivery schematic 20025030035040045050055060065000:00.007:12.014:24.021:36.028:48.036:00.043:12.050:24.057:36.0Time (minutes)Temperature (oC) Reformer Top Reformer Center Reformer Bottom Reformer Burner Ramp Up PeriodSteady State High PowerSteady State, Idle Ramp Down Period Reformer Burner Idle Reformer Bottom Reformer Center Reformer Top Figure 2-4. TBB-2 fuel cell bus methanol steam reformer and steam reformer burner temperatures during transient bus operation

PAGE 28

16 ReformerBurnerSteamReformerNeatMethanolWaterMethanolPremixStart-upBurner Air Air Excess Airand Water Anode Air CathodeFuel Cell Stack Anode Flue GasReformateCombustionProductsHeat ExchangePlateHeat ExchangerBypassVaporizerCombustionProducts Figure 2-5. University of Florida bus indirect methanol fuel cell system CO and Hydrogen Concentration Effects on PEMFC performance Small increases in reformate CO concentration lead to drastic reductions in fuel cell performance in low-temperature fuel cells that use platinum catalyst to promote the cell anodic reaction. In these cells, CO competes directly with hydrogen for active sites in the anode due to its high affinity for platinum at low-temperatures. Springer et al. propose that the reactions depicted in Equation 2-6, Equation 2-7, Equation 2-8 and Equation 2-9 represent well the competition for active sites at a PEMFC anode [5]. )(/COMMCObfckfc (2-6) )(/2COMMHbfhkfh (2-7) MeHHMkeh)( (2-8) eHCOMCOMOHkec22)(22 (2-9)

PAGE 29

17 Equations 2-6 and 2-7 represent the competition between H 2 and CO for active sites, where kfc and kfh represent the forward reaction-rate constants. The electro-oxidation of hydrogen and CO are represented in Equations 2-8 and 2-9, with their respective rate constants (keh and kec). PEMFC models and experimental results have demonstrated that drops in anode hydrogen concentration produce very mild decreases in cell performance (Figure 2-6). However, high CO concentrations have been shown to drastically drop cell performance, especially when diluted in reformate bearing low hydrogen concentrations (Figure 2-7). In order to further reduce the CO concentrations produced from steam reformers, many fuel processing systems use preferential oxidizers and anode air bleeds (the introduction of small quantities of air into the fuel cell anode chamber). Preferential oxidizers require the addition of air into a catalytic bed that promotes the oxidation of CO. Typically a small percentage of the hydrogen in the reformate is also oxidized. Currently, control of air delivery systems for preferential oxidizers is complicated due to the high sensitivity of the chemical rate equations to gas and catalyst temperature and CO concentration. In addition, the amount of air associated with preferential oxidation operation is very low and difficult to meter. In the overall reforming process high upstream variations in CO concentrations may lead to increased concentrations of CO leaving preferential oxidizers, thus increasing the probability of anode catalyst poisoning. This is especially problematic since CO spikes occur during periods of increasing power demand. This phenomenon will be further explained in subsequent chapters. Various solutions have been proposed for solving the problem of CO poisoning of the fuel cell anode catalyst. Of these the most commonly reported has been the

PAGE 30

18 introduction of small quantities of air into the cell anode during periods of decreased cell performance due to CO poisoning. This produces catalytic oxidation of the CO in the presence of anode platinum. Nonetheless, the air bleed also has a tendency to oxidize a portion of the hydrogen introduced into the fuel cell. This results in a further reduction in cell efficiency. Figure 2-8 shows how the effect of air bleed introduction in the anode reduces the poisoning effects of CO. Also, it is difficult to distinguish between cell performance deterioration due to CO poisoning and other operating conditions (humidity variations, temperature variations, pressure variations, cell degradation, etc). Based on this, a reformer that produces minimal increases in CO level during increased fuel flow events is more desirable than one that produces high hydrogen concentrations (within typical reformate hydrogen and carbon monoxide concentrations). Figure 2-6. Effect of hydrogen concentration and utilization on PEMFC performance. This figure was published with the permission of its original authors [5].

PAGE 31

19 Figure 2-7. Effect of CO poisoning on PEMFC performance. This figure was published with the permission of its original authors [5]. Figure 2-8. Use of air bleed to recover PEMFC performance. This figure was published with the permission of its original authors [5]. Methanol as Fuel Since most of the results of this research are based on the study of methanol reformation, a brief discussion of methanol as a potential fuel cell fuel is relevant.

PAGE 32

20 Of the alternative fuels which are being considered for vehicles, many view hydrogen as the ultimate long-term alternative fuel. However, while hydrogen has some very desirable attributes, there are many extremely difficult issues to resolve, which might make the time scale for a hydrogen economy far longer than the time scale of readily available and affordable petroleum. Fuels which can be stored as liquids offer major advantages for transportation applications. Of the alternative liquid fuels, in many ways the most promising is methanol. This is because methanol reforming technology is well developed as compared to reforming technology for other common automotive fuels. Also methanol can be produced from natural gas, biomass, electricity, coal, and any other hydrocarbon. These result in a fuel that can potentially be relatively inexpensive in most regions of the globe. In addition, methanol can be used in internal combustion engines, gas turbine engines and in fuel cell engines. Most of the methanol produced in the U.S. is derived from natural gas. Twelve major methanol production plants exist in the U.S., which in 2001 produced over 1.5 billion gallons of methanol (94.2 trillion BTU or equivalent to 753.6 million gallons of gasoline). The overall U.S. methanol consumption in 2001 was, in energy terms, equivalent to ~1,507 million gallons of gasoline, half of which was imported [6]. The price of delivered methanol has averaged ~$0.50 per gallon (~$8/million BTU) since 1991. If the price of natural gas did not figure into the cost of methanol, the methanol cost would be ~$1.50/million BTU. Currently, most of the U.S. produced hydrogen is derived from natural gas steam reforming. The U.S. consumes around 9 million tons of hydrogen (925.8 trillion BTU or

PAGE 33

21 equivalent to ~7,407 million gallons of gasoline) per year, of which approximately 17% is sold to chemical plants and refineries [7]. Current hydrogen costs at large scale chemical plants, where hydrogen is consumed on site, are generally around $5.31/million BTU. If transported, the price of hydrogen is much higher. The typical price of delivered liquid hydrogen oscillates between $1.00 and $1.40 per pound ($20/million BTU to $27/million BTU). Therefore hydrogen prices are comparable to methanol prices only if hydrogen is not pressurized, liquefied, or transported. The cost of producing hydrogen using steam reforming of natural gas minus the cost of the natural gas, for a large plant, would be ~$4.15/million BTU. This price does not include transportation, storage, compression or liquefaction of the gas. While for methanol, storage and transportation costs are not essential components of the overall fuel cost, this is not the case with hydrogen. It is expected that, due to its lower energy density, methanol transportation costs per unit energy basis will be twice the cost for transporting gasoline, in energy terms (~$6.92 per million BTU) [7]. Hydrogen transportation cost, in contrast, may vary drastically depending on the distance transported and the method of transportation. Taking current gasoline transportation costs as a baseline, it could be assumed, as an initial estimate, that because of its lower energy density, liquid-hydrogen transportation would cost at least 5 times more than gasoline transportation if similar transportation forms were used (mainly pipeline and truck transport). Thus, hydrogen transportation costs could be expected to be around $18.94 per million BTU [7].

PAGE 34

22 Hydrogen distribution costs could also be estimated by assuming that the eventual hydrogen distribution infrastructure will be similar to that of natural gas. This assumption is legitimate since both fuels are gaseous. The average distribution cost for natural gas, from wellhead to residential consumer 1 in the U.S., from 1991 to 2003, was $4.80/million BTU. Thus, by applying an energy density penalty it could be estimated that the average hydrogen distribution price in the U.S. will be ~$18.51/million BTU (similar to the previous estimation). Note however that these estimates do not include additional cost penalties associated with hydrogen transport such as boil-off, gas leakage, and special materials needs. Based on this basic analysis, it is fair to assume that methanol could be an important fuel for transportation and fuel cells since it has the potential to be much cheaper and more easily distributed than hydrogen. A study by the University of Florida found that based on future fuel cell vehicle fleet estimates for 2020, and according to various economic scenarios, methanol derived hydrogen from coal would be the least expensive automotive fuel for fuel cell vehicles in the US if environmental concerns are not taken into account [7]. 1 Residential natural gas prices were used to estimate hydrogen distribution prices because it most closely resembles the distribution requirements of the automotive sector.

PAGE 35

CHAPTER 3 LITERATURE REVIEW This literature survey has the purpose of framing the work and results of this dissertation as part of the latest research and scientific debates regarding the design and integration of methanol steam reformers into fuel cell systems. Generally, research conducted on reformers can be catalogued into two different approaches, the chemical and the physical. Chemical research has concentrated in trying to understand the chemical mechanisms and rate of reaction relations of the steam reforming reaction. A large volume of papers and presentations have been published in the last couple of years regarding this subject. Common among these is a general concern among researchers that almost every research case (even when identical catalysts are used) has lead to different rate of reaction equations, although in the same mathematical form. The physical approach to reformer research has dealt with understanding the complete reformer assembly and its behavior. The volume of work in this area is limited. In addition, most of the work has concentrated on the steady-state case. Dynamic models and studies on fuel cell systems in published literature typically rely on user-defined linear exponential decay functions for each fuel cell system component, including the reformer. These investigations concentrate on the question of fuel cell systems control. Typical of this approach is the study conducted by El-Sharkh et al. 2004 [8]. 23

PAGE 36

24 Kinetics of the Catalytic Methanol Steam Reformer Reactor Although the actual kinetics of the methanol steam reforming reaction were outside the scope of this study, a general review of the work in this field is herein presented. Of note is the diversity of rate equations that have been presented in literature. This demonstrates the complexity of the reaction and the variability that can be obtained based on catalyst selection, catalyst preparation, and reaction conditions. Choi et al., 2005, proposed that the methanol steam reformer products can be obtained by with the rate equations for 3 main reactions in a Cu/ZnO/Al 2 O 3 catalyst manufactured by Sud-Chemie. These were [9], 22222232332HCOOHCOHCOOHOHCHHCOOHCH The study conducted by Choi greatly simplified the widely cited 12 reaction rate model previously given by Peppley et al. [10]. Peppley and Choi point out in their studies the large degree of disagreement in literature regarding reaction mechanisms and the debate on identifying the physical location of the steam reforming catalyst active site. The results of their experiments were a series of rate equation for each of the proposed methanol steam reformer reactions. Based on these reactions a methanol steam reformer, a water gas shifter and a PROX reactor model were developed. The results of their models show an increase in conversion with greater temperature and catalyst weight in the reactor, however this also increased CO. The study conducted by Choi did not model heat transfer in the reactor bed or the transient behavior of the reformer. Lee et al., 2004, [11] recognized the importance of steam reforming kinetics in sizing methanol steam reformers. They also recognized the large diversity of rate

PAGE 37

25 equations found in literature. Lee, in his paper includes a list of 9 different catalytic methanol steam reforming rate of reaction equations found in recent literature. Many of these reaction rate equations are contradicting even when the same catalyst is being used. Lee observes that higher water concentration in the steam reforming reaction promotes effluent CO concentrations closer to equilibrium. The discrepancy between experimental CO concentrations and equilibrium concentrations were between ~1mol% to <0.1mol% for a water feed concentration between 15mol% to 30mol%. Finally Lee, in his study, also proposed a new rate of reaction of methanol for the catalyst used. The correlation proposed by Nakagaki et al. [12] was used in the models produced in this study. Nakagaki obtained his correlations through a series of experiments that closely resembled the test conditions in which data was obtained for this study. The study conducted by Nakagaki used a temperature controlled packed-bed reformer operated at various fuel flow rates. In obtaining his correlation, Nakagaki conducted constant catalyst temperature studies and constant wall temperature studies (the latter allowed for temperature gradients to exist in the reformer catalyst bed). Finally his correlation was validated through actual reformer operation and modeling. This type of experimentation varies from others in literature in that the effects of fuel flow (transport phenomena) and varying reactant concentrations were considered. Also considered are the effects of temperature gradients in the catalyst. The following is a general description of the study conducted by Nakagaki, et al. Nakagaki et al. obtained their correlation through experiments conducted in a packed-bed experimental reformer. The catalyst used was Cu-Zn in alumina support commercially available from Nissan (G66-B). The catalysts were 1/8in by 1/8 in

PAGE 38

26 cylindrical tablets. Various concentrations of methanol diluted in simulated reformate gas assuming conversions equivalent to 0, 10, 20, 50, and 75 percent were flowed through the packed bed reformer. The reformer temperature was maintained constant and controlled through the use of sheath heaters. This first test was used to determine the methanol rate of reaction in a constant temperature reaction. A second test controlled the temperature of the reforming reaction by maintaining a constant wall temperature. In the second test the temperature distribution and the conversion were recorded. The exiting reformate gas composition was recorded using a TCD gas chromatograph. Based on second test results, the Nakagaki correlation was refined. Table 3-1 shows a summary of the tests conducted by Nakagaki. Nakagaki also conducted modeling studies through a combined energy, momentum (Darcian equation), and conservation of species equations to further validate his correlation. However, since the application of the study was for employment in gas turbine engines, the study of CO concentration in the reformate outlet stream was neglected. Table 3-1. Test parameters for Nakagaki tests Parameter Test 1 Test 2 Pressure (atm) 1, 5, 10, 15 1, 10 Catalyst Temperature (K) 473, 498, 513, 538 -Wall Temperature (K) -518, 548 Mass Flux (kg/(sm 2 )) 0.25 0.25, 0.36 Simulated conversion percentage 0, 10, 20, 50, 75 0, 80 Reformer Performance Various studies have been published on reformer performance and design but the majority of these have concentrated on reformer steady-state performance and catalyst

PAGE 39

27 dynamic behavior (rates of reaction correlations). The work of Betts et al. [13] provided a mechanistic model for observed fluctuations in reformer CO output. The identification of CO spikes as being caused primarily by heat transfer effects in the reformer was discussed in that study. A study by Horng [14] showed for an autothermal reformer the existence of CO spikes once fuel was introduced into the reformer. However, the study concentrated in the start-up process of the reformer and did not place much emphasis on the CO phenomenon. The importance of heat transfer and temperature on the performance of methanol steam reformers was studied by Hohlein et al. [15]. Hohlein concludes that reformer heating is a very strong determinant of reformate CO concentration. Perhaps the most comprehensive experimental study on reformer performance was conducted by Davieau [16]. In this study, reformer performance was measured in terms of methanol conversion for two reformers of differing diameters. The initial interest of Davieau was to study the degree in which space velocity and aspect ratio determines reformer similarity. The reformers were operated at various space velocities. The results demonstrate that space velocity and aspect ratio were not a good determinant of reformer performance. That is, two reformers operating with the same space velocity, same catalyst loading, same wall temperature and same aspect ratio had drastically different performances. Yet, Davieau points out that space velocity is the most cited reformer similarity relation in literature. Additionally, Davieau notes that temperature distribution within the reformer is a critical performance parameter. Furthermore, it is noted that diameter influences heat transfer and therefore has an effect on performance that is

PAGE 40

28 beyond catalyst loading. The results obtained by Davieau were based on experimental observations. His study did not focus on the transient behavior of the reformer.

PAGE 41

CHAPTER 4 ANALYSIS In accordance with the objectives of this research, a general methanol steam reforming model was developed. The details of the derivations and major assumptions applied to the development of this model are included in this chapter. In addition, this chapter includes a general comparison of the steam reformer model results and experiments conducted at UC-Davis using an experimental reformer test rig developed at by Dr. Paul Erickson (Appendix C). Part of the analysis includes the development of a generalized design and comparison criteria for steam reformers through similitude parameters. Finally, in this chapter transient operation of the modeled steam reformer is presented. Steam Reformer Model Packed bed catalytic steam reformers are generally composed of a reaction chamber filled with catalyst pellets (Figure 4-1). The methanol water mix along with the product gases flows around and through the catalyst pellets. The chemical reactions occur at the surface and within the catalyst pellets. In the development of this model, it was recognized that the main driver of reformate chemistry is temperature. A control volume approach was used to obtain differential conservation of species and conservation of energy relationships. In choosing the scale of the finite differential control volume, it was assumed that the catalyst pellets and the gases exist in the control volume as a homogenous mixture. That implies that the control volume used, although small compared to the overall reformer, has length scales larger than that of the catalyst. 29

PAGE 42

30 It was also assumed that the gas and the catalyst exist locally at the same temperature. This comes as a result of the assumption that the catalyst and the gases are regarded as part of a continuum. This also implies that the surface area of the catalyst inside the control volume is very large, resulting in the gas being in thermal equilibrium with the catalyst. In addition, the flow was assumed to occur in the axial direction (i.e. the radial component of the velocity was assumed negligible). Figure 4-2 shows the control volume with the major heat inputs and outputs at the boundaries. Steam Reformer Walls Catalyst Pellets Methanol + Water Reformate Reaction Chamber Figure 4-1. General schematic of modeled steam reformer reactor Figure 4-2. Control volume diagram

PAGE 43

31 Heat Transfer Based on the chosen control volume and the aforementioned assumptions, the heat transfer equation can be derived (Equation 4-1). x TCudVdE x Tk r Tk r T r ktTCgaspgasgeffeffeffeffeff ,22221 (4-1) The variables in Equation 4-1 are defined as follows: eff is the effective mass density of the catalyst-gas mixture, C eff is the effective specific heat of the catalyst-gas mixture, T is the temperature inside the control volume, k eff is the catalyst-gas mixture effective conductivity, r is the radial coordinate, x is the axial coordinate, dis the net rate of heat absorbed by chemical reactions, V is the reformer volume, gE gas is the density of the gas, e is the void factor of the reformer, and gaspC, is the average specific heat of the gas. The thermal capacitance of the mass inside the control volume is assumed to be mostly because of the catalyst. Therefore eff C eff is approximately equal to the catalyst density ( c ) multiplied by the catalyst specific heat (C c ) and the void factor (). The use of effective properties for modeling of catalytic reactors is a fairly common practice. Incoprera et al., in their widely used heat transfer text book, Fundamentals of Heat and Mass Transfer, describes heat transfer correlations for packed bed systems using effective gas/packed-bed properties [17]. In published literature, various methods for determining packed-bed effective thermal conductivity have been explored. Typical of this work are the studies conducted by Munagavalasa et al. [18], and Brucker et al. [19].

PAGE 44

32 The convection term in Equation 4-1, xTCugaspgas , for modeling purposes was assumed to be small as compared to the radial conduction terms. Typically, temperature gradients in the radial direction are typically higher than those in the axial direction through most of the steam reformer. A possible exception to this could occur at the entrance of the reformer if the entrance temperature is significantly different than the reformer catalyst bed temperature. As an example of the disparity between the convection and conduction terms, an order of magnitude analysis of these, based on data obtained from the UC-Davis reformer (see Appendix C), is presented in Figure 4-3. The analysis is executed for a run in which various changes in methanol flow were executed. The results were obtained from temperature data recorded at the entrance regions of the reformer. A finite difference approximation of the derivative and second derivative of temperature in the axial and radial directions, respectively, were calculated using K-type thermocouple readings. The radial thermocouples were placed at the wall, at half the radius and at the midpoint, where the radius of the tube was 11.9 mm. The axial thermocouples were 67.6 mm apart. It is recognized that the significantly larger axial distances used to approximate the finite difference axial temperature gradient leads to greater averaging. This in turn may lead to an underestimation of the axial gradient in the entrance region. In spite of its limitations, this analysis provides a general relative magnitude of the axial and radial temperature gradients. During most of the run, the convection term did not exceed 0.4% of the value of the conduction term. This, in order of magnitude terms, supports the decision to disregard the convection term in the energy equation. The UC-Davis reformer is discussed in greater detail at the latter sections of this chapter and in Appendix C.

PAGE 45

33 Figure 4-3. Convection-conduction ratio in the UC-Davis reformer fuel entry regions The heat dissipated in the catalyst bed is due to the reforming reaction. Consequently, dis a function of the rate in which methanol is reacted in the steam reforming reaction, gE dtOHCHd3 This in turn is a function of the temperature (T), catalyst volume (dV c ), and the molar concentration of methanol in the reformate gas OHCHx3 The rate of methanol consumed per unit mass of catalyst (r M ) and can be obtained using Equation 4-2 ),,(211333OHCHccMxPTfdtOHCHddrrdrdxdtOHCHdmr (4-2) The calculation of the rate of methanol reacted in a CuO/ZnO catalyst for methanol-steam reforming reactions was extensively studied by Nakagaki et al. at Toshiba Power Systems [12]. One of the results of this study has provided a good correlation for r M shown here:

PAGE 46

34 3.113.0/1010 toequal sit' otherwise513K T if 10)/(1035.1513563nlmolJEmatmsgmolkxeTPkrcatonOHCHRTEmloM (4-3) The Nakagaki correlation is an empirical variation of the Arhenius equation and was obtained experimentally. Further information regarding the Nakagaki study is found in the literature review chapter of this text. Given these relationships, d can be defined explicitly (Equation 4-4). gE MrcrgrEdrrdxdrdtOHCHdEEd)2(3 (4-4) where the expression for the heat of reaction per mole of methanol consumed is expressed as E r Therefore the energy equation for the control volume can be rewritten as Equation 4-5. cMrCrExTrTrTrtT2222 (4-5) The quantity is a function of temperature, thus Equation 4-5 is non-linear. This non-linearity is the source of the complex and sometimes unexpected behavior associated with packed bed steam-reformers. The boundary conditions for the heat equation (Equation 4-5) are MrrE

PAGE 47

35 iLxingwrTtTxTTxTTRrTrT)0(0)0()(0,0 (4-6) Given these boundary conditions the energy equation can be solved by using a finite difference scheme or using finite element analysis. Also, Equation 4-5 can be made non-dimensional to yield: 2**2*22****22*2****21,,, M rww R rETTTRTForrrLxTTrxtTrxFoTrLR (4-7) For purposes of this study, the heat transfer model was solved in dimensional terms in order to relate results to real physical parameters which can be readily verified through experimentation. Continuity The fluid flow within the steam reformer tube is transient, compressible and viscous. If it is assumed that the mean flow travels in the axial direction (u x >> u r ), then the continuity equation can be written as Equation 4-8. xutxgg (4-8) where u x is the axial component of the fluid velocity vector, and g is the density of the fluid. If the fluid is assumed to be an ideal gas, then the continuity equation can be written in terms of fluid temperature. In addition, it is convenient to use mass flow rate,

PAGE 48

36 m instead of the velocity because cross-sectional averages of the velocity will be used. This yields tT RT PAxmcs2 (4-9) In essence Equation 4-9 states that variations in the axial mass flow are a function of the transient change in temperature at any given location inside the reformer. However, reformer data collected from the UC-Davis reformer suggests that tT for methanol steam reformers is small even at the entrance, centerline region (Zone 1 and Zone 2). This is because the thermal capacitance of reformer catalysts is relatively high, and because the primary mode of heat transfer is through conduction, which is relatively slow. Figure 4-4 shows recorded values of tT from the UC-Davis reformer. For illustration purposes, even if the maximum value exhibited was held at ~2 K/s (the maximum recorded from the UC-Davis reformer during reformer warm-up) and assuming that this level of heating were to be carried out through out the entire reformer, the change in reformate mass flow rate that would be detectable at the reformer exit would be 0.018% for a 1 meter long reformer, based on Equation 4-10. dxtTRTPAmmLcsinout02 (4-10) Therefore, the mass flow rate within the reformer will have very small transient variations. Despite this, Equation 4-10 was used in the elaboration of the reformer model. Temperature distribution plots for the UC-Davis reformer at varying flow rates are shown in Figure 4-5. Note: for modeling purposes the pressure drop in the reformer was neglected.

PAGE 49

37 UC-Davis Reformer dT/dt values-2-1.5-1-0.500.511.522.50102030405060708Run Time (min) 0 Zone1 dT/dt Zone2 dT/dtCalculated dT/dt [K/s] Figure 4-4. Calculated values of dT/dt for the UC-Davis reformer at the entrance region Figure 4-5. Measured temperature distributions of the UC-Davis steam reforming test rig operating at varying space velocities and varying conversion efficiencies

PAGE 50

38 Conservation of Specie The continuity equation for each of the component gases in the control volume can be written as Equation 4-11. genjoutjinjmmmtmjjjj,,,jorvolumecontrol theinside ofproduction of ratevolumecontrol theofout of flow massvolumecontrol theinto of flow mass volumecontrol thein of mass of increase of rate (4-11) The component gases in the steam reformer products are assumed to be methanol, water vapor, carbon dioxide, carbon monoxide, and hydrogen. The mass generated of component j inside the control volume can be estimated using the Nakagaki correlation (Equation 4-3) and chemical equilibrium. Thus, for a control volume at a certain temperature and pressure, the rate of methanol reacted is known and, consequently, the rate of generation of the product gases is also known. The rate of mass of specie j entering the control volume is generally known from boundary conditions. However the rate of mass escaping the control volume must be calculated. For purposes of the model it was assumed that all species flow at the same axial velocity as the bulk flow. The bulk flow velocity exiting the control volume can be expressed as Equation 4-12. outjLcsincsincscsoutcsoutoutoutudxtTTTPARTmPARTmdxtTRTPAPARTmAmu,0221 (4-12) Consequently, the mass flow rate of the exiting gas j is given by Equation 4-13.

PAGE 51

39 outjoutjcsjoutjmxuRTAPm,, (4-13) In the above equation, P j is the partial pressure of gas j inside the control volume. The term x j is the molar concentration of gas j, and is defined as Equation 4-14. ottNjjjojnnttx1)( (4-14) Chemical Model The chemical reaction for steam reforming of methanol being considered is shown in Equation 4-15. OHeCHdHcCObCOOaHOxHOHCH322223 (4-15) In a control volume a certain rate of methanol can be reacted (r M m c ). The remaining methanol and water that enters the control volume is not reacted (Figure 4-6). It was assumed that the water-gas shift reaction is not the time limiting step in the reformer reaction and thus proceeds to equilibrium rather quickly. For modeling purposes the water gas shift reaction was assumed to reach equilibrium concentrations, while the reforming of methanol was bound by the rate of reaction constraints of the control volume. Similar approximations have been made in literature and the work of Rabou et al. [20] is an example.

PAGE 52

40 Methanol (mol/s)Water reacted (mol/s) Not reacted (mol/s) CO2 H2COCV H2COCO2Methanol (mol/s)WaterH2COCO2 Figure 4-6. Schematic representation chemical reaction model in the control volume. The chemical equilibrium fractions for the water gas shift reaction can be calculated in the following way. The specie balance of the complete steam reforming reaction (e and f are equal to zero) leads to dxa 2 (4-16) 2db (4-17) dc3 (4-18) The solution for d can be obtained from the water-gas shift reaction (Equation 4-19) 222COHCOOH (4-19) The equilibrium constant for the water gas shift reaction can be approximated with the Moe (1962) correlation [21] (Equation 4-20). T Kp8.457733.4ln (4-20) where T is the temperature of the reaction in Kelvin. Given this,

PAGE 53

41 )1(2254442222KpKpxKpKpKpKpxKpKpxd (4-21) With these equations the number of moles of each chemical component per mole of methanol reacted can be obtained as a function of temperature. The number of moles of methanol reacted can be obtained from the Nakagaki correlation (Equation 4-3). The equilibrium heat of reaction and products as a function of temperature for the methanol steam reforming reaction is given in Figure 4-7. Equilibrium Reformate Composition00.511.522.533.5298398498598698798898998Temperature (K)0100002000030000400005000060000700008000090000 CO2 H2 CO Hr [kJ/kmol of MeOH]Heat of Reaction [kJ/ kmol of Methanol reacted]Gas Composition [kmols/ kmol of Methanol reacted] Magnitude Depicted in Secondary Axis Figure 4-7. Equilibrium products of the steam reforming reaction (CH 3 OH + H 2 O aH 2 O + bCO 2 + cCO + dH 2 ) with respect to temperature Finite Difference Solution The control volume analysis easily leads to a finite difference solution of the heat transfer and fluid flow problem presented in the previous sections. A grid was set up where each element has a temperature (T m,n ), in which m represents the node index in the

PAGE 54

42 x-direction and n represents the node index in the r-direction (as shown in Figure 4-8). The derivatives of the catalyst bed temperature were approximated to be r TT r TrTTTrTxTTTxTnmnmnmnmnmnmnmnm2221,1,2,1,1,222,,1,122 (4-22) Similarly the time derivative of the temperature could be written as tTTtTlnmlnm,1, (4-23) where l is the time index. x r x r x r m+1,nm,n+1 m,n x r R 0 Figure 4-8. Separation of the reformer volume into annular elements with constant gas properties in space for the finite difference solution method used Given these equations, Equation 4-5 could be rewritten as

PAGE 55

43 clnmMrlnmlnmlnmcclnmlnmlnmlnmlnmlnmlnmCrExTTTCkrTTTrTTrtTT1,2,,1,12,1,1,1,1.,1,2222 (4-24) Similarly the boundary conditions can be rewritten (Equation 4-25). inmlnMlnMinglnlNmlNmwlmTTTTTTTTTT0,,2,,,02,,0, (4-25) The term is unknown and it is a function of the finite volume temperature and methanol concentration. At first, the temperature equation can be solved assuming a temperature and a methanol concentration. For the computer program developed for this study, T 1,lnmMrrE m,n l+1 was initially assumed approximately equal to T m,n l Similarly the concentration of methanol inside the volume, x CH3OH,m,n l+1 was at first approximated to be equal to x CH3OH,m,n l Once an approximate solution for the temperature was obtained the conservation of specie equation was solved. The finite difference expression for 4-10 is given by 1,,1,,1,1,,,,,lgennmlnmjlnmjlnmjllnmjmmmtmm (4-26) The expression for the mass flow of specie j leaving the finite volume can be expressed as

PAGE 56

44 21,,1,21,1,11,11,,1,,5.0lnmlnmlnmlnmlnmlnmcslnmjlnmjTTTTTTtxRPAxm (4-27) Note that as with the temperature solution, x j,m,n l+1 is unknown. In the computer program this value was initially assumed to be equal to x j,m,n l From the solution of the content of all the species in the finite volume using Equation 4-24 and 4-27, for a given temperature field, a new value for the molar concentration of each component can be obtained (Equation 4-28). NiilnmijlnmjlnmjMWmMWmx1.,1.,1,, (4-28) This new, revised solution for x j,m,n l+1 was used to solve for the finite volume temperature and mass contents. This process was carried out until satisfactory convergence was achieved. FEMLAB Solution Besides using a finite difference scheme, the system of differential equations was solved using finite element analysis using FEMLAB Multiphysics modeler software. The FEMLAB software is designed to solve systems of coupled partial differential equations using proprietary solution schemes based on finite element analysis. The system of partial differential equations for transient heat equation (including the convection term), transient conservation of species, transient continuity, chemical equilibrium and the Nakagaki correlation were programmed into the FEMLAB solver. A triangular element grid was used to cover the two-dimensional reformer space. Appendix B specifies details

PAGE 57

45 of the number of elements and number of degrees of freedom for each of the reformers modeled. The results of the FEMLAB finite element solution and the finite difference solution were similar and compatible. However, it was found that the finite difference solution resulted in lower reformate CO concentration. This was due to problems with the inherent lower magnitude of CO concentration in the reformate gas as compared to water, hydrogen and carbon dioxide concentrations. This difference created problems with convergence and minimum error search. The FEMLAB solver uses sophisticated minimum error search schemes which include parameter step annealing and magnitude parity. Because of this computational error reduction feature, the FEMLAB solution was assumed to have a higher degree of accuracy than that of the finite difference scheme. Steam Reformer Model Validation Validation of the steam reformer model results were done through comparison of the performance of modeled steam reformers with the UC-Davis reformer. Experiments were conducted using a small-scale laboratory steam reformer located at the University of California Davis (See Appendix C). Data from this experimentation was used for model validation. The UC-Davis reformer was built and instrumented by Dr. Paul Erickson and his students. Experiments for this study were conducted by UC-Davis students under the direction of the author of this dissertation. The experimental plan was also developed by this author for the purposes of this study. Various peer-reviewed publications have been based on research conducted using the UC-Davis reformer. These include Liao et al. [22], Yoon et al. [23], and Betts et al. [24] (the latter regarding some of the results of this study).

PAGE 58

46 Data obtained from the UC-Davis reformer cannot be used as a complete model validation tool. The reformer suffers from loose wall temperature control and gas sampling limitations that inevitably lead to discrepancies between the magnitudes of the model results and the data collected. Nonetheless, important physical phenomena, such as transient CO spikes, CO concentration dependence on reformer temperature, and steady-state and transient reformate composition trends are captured in the data obtained. Agreement between the model and the UC-Davis reformer were expected to be approximate. General Description The UC-Davis reformer is a catalyst filled cylinder with a radius of 0.47 inches (0.012 m), and an axial length of 21 inches (0.533 m) (Figure 4-9). The reactor temperatures are controlled using eight electrical band heaters located on the surface of the reactor. Thermocouple readings of the internal wall temperature of the reactor and the centerline temperature are used to gauge the level of chemical activity in a certain region of the reformer and are also used as feedback controller inputs for the heaters. The reformer heaters operate in on or off mode, as triggered by computer controlled relays. A simple feedback control loop, programmed in Labview, was used to maintain a constant wall temperature. Before entering the reformer water and methanol are vaporized and superheated using a series of pipes containing cartridge heaters. The temperature of the reactants entering the reformer is recorded and is controlled via a control and data acquisition computer, which is capable to turning on or off the vaporizer and super-heater heaters. Water and methanol fuel flows are metered using positive displacement gear pumps. The

PAGE 59

47 gear pump RPM was used to measure flow. A scale, on top of which the reactants sat, was also used to measure methanol and water flow into the reformer. Reformate gas analysis was done using a NOVA Analytical Systems, Inc. gas analyzer (Model 7904CM). This analyzer was designed for simultaneous analysis of hydrogen, carbon dioxide, methane, and carbon monoxide. The instrument recorded carbon dioxide, carbon monoxide and methane using a Non-Dispersive Infrared (NDIR) detector with temperature and pressure compensation. Hydrogen was detected using a temperature controlled thermal conductivity (TC) cell. Before being analyzed for composition, the reformate gases were cooled to an almost freezing temperature (i.e. ~0 o C) and the condensed liquid was extracted. The UC-Davis reformer was designed with the capability to capture all condensed liquids before gas sampling to perform conversion studies. Due to the condenser volume, a short time response delay exists in the gas analysis data. Although the exact time response delay is a complex computation problem that falls outside the scope of this study, a general estimation can be performed. The length of the condenser tube is 18ft (5.5m). Reformate enters the condenser at approximately 300 o C and leaves at approximately 0 o C. If the mass flow rate of fuel entering the reformer is assumed to be the same as that entering the condenser, then the reformate volume flow entering the condenser can be calculated. For purposes of estimation the mass flow rate leaving the condenser can be said to be equal to the reformate mass entering it. Based on these assumptions the time it would take the time it would take to have a complete condenser volume gas change (i.e. retention time) for a premix flows at STP ranging from 5 mL/min to 30 mL/min would range between 1.3 to 0.2 seconds. Additionally, for all premix fuel flow rates, the flow through the

PAGE 60

48 condenser can be estimated to be mostly laminar due to low Reynolds number. This means that a well organized concentration boundary layer is expected to exist within the condenser pipe. Therefore, although the reformate gas composition may change, traces of previous reformate gas could still be trapped in the boundary layer, close to the walls. This may lead to a decrease in the sharpness of the data obtainable from the gas analyzer. The NOVA gas analyzer manufacturer-quoted response time is 3 seconds to 90% response. The gas analyzer takes samples at a rate of 1 L/min through a positive displacement pump. At this flow rate the condenser unit would have a retention time in the order of 1 s. Since the response time is in the order of seconds and the measurable effects (that is reformate gas concentration changes) occur in the order of minutes, these effects are recordable using the UC-Davis hardware configuration. In terms of time response, passage through the vaporizers may be the limiting factor. For the model, the fuel flow rate boundary condition is at the inlet of the reformer. This is not the case for the UC-Davis reformer. If it were assumed that only liquid fuel passes through the vaporizers, then the residence times of the vaporizers would be as shown in Table 4-1. The estimated residence times presented in this table are maximum values. For lower fuel flows the space velocity is over-estimated because vaporization occurs earlier in the vaporizer tube, which increases the volume flow rate of the fuel. Also, the degree in which vaporization occurs within the vaporizers is not immediate and this will result in a ramped increase in fuel flow into the reformer instead of a step increase as in the model. This should result in smoother, drawn-out, transient gas composition data for the UC-Davis reformer.

PAGE 61

49 Table 4-1. Estimated maximum times for vaporizer volume flow change. Fuel Flow Rate (cm 3 /min) Vaporizer residence times 5 5 min 10 3 min 15 2 min 20 1.3 min 25 1 min 30 ~50 s Experimental Method The reformer catalyst bed was maintained idle for storage completely sealed and pressurized with carbon dioxide. This process ensured minimum catalyst deactivation due to oxidation. For experimentation, the reformer was first heated using the reformer band heaters under a CO 2 gas purge. Once the appropriate reformer temperature was reached (250 o C), vaporized and superheated water and methanol were introduced. The experiment maintained the reformer wall temperature at a nominal 300 o C with fluctuations due to heater binary control. Water-methanol flows between 5 and 30ml/min were passed through the reformer in 5ml/min intervals. Changes in methanol-water flows were executed every 10 minutes. The data sampling rate was two seconds. UC-Davis Reformer Data The data obtained from the UC-Davis reformer shows the following trends: There is a temporary increase in the CO concentration in the reformate gas leaving the reformer when there is an increase in fuel flow rate into the reformer (Figure 4-10). The CO concentration in the reformate is very closely related to the catalyst bed temperature (Figure 4-11). Catalyst bed temperature fluctuations in the UC-Davis reformer were mostly because of band heaters binary operation and imposed control scheme. The steady-state CO concentration in the reformate lowers with increased fuel flow rate into the reformer (Figure 4-10 and Figure 4-12).

PAGE 62

50 Figure 4-9. Experimental steam reformer developed at UC Davis CO Spike Due to Change in Flow Rate0246810121416222324252627282930Run Time (min)00.511.522.53 Pump feedback (ml/min) CO concentration (%) Increased levels of CO are registered when there is an increase in flow rate Steady state CO levels are lowered with increased flow rate Data obtained at U.C. Davis on 2/5/2004Methanol Water Mix Flow RateCO Concentration (%) Figure 4-10. Example of CO spike in the reformate of the UC-Davis reformer after an increase in premix fuel flow (transient condition). Also shown is a decrease of the steady-state CO concentration in the reformate with increased premix fuel flow rate in units of mL/min.

PAGE 63

51 Relationship between CO production and Catalyst Temperature200210220230240250260270280290300515253545556575Run Time00.511.522.533.544.5 Reformer Exit Centerline Temperature Average Centerline Temperature (oC) CO concentration (%)Temperature ( oC)CO Concentration (%)Magnitude is depicted in secondary axis Figure 4-11. Reformate CO composition relationship to changes in temperature in the reformer catalyst bed 1 Feb. 5th, 2004 Test Run at UC Davis0102030405060708051525354555Run Time (min)00.511.522.53 Pump feedback (ml/min) H2 concentration (%) CO concentration (%)H2 Gas Concentration (%)Premix Flow Rate (std. mL/min)CO Gas Concentration (%) CO SpikeCO Spike Figure 4-12. CO and H2 concentrations in the reformate gas in the UC-Davis reformer 1 The drastic decrease in CO concentration at the end of the graph comes as a result of shutting down the premix flow into the reformer.

PAGE 64

52 It was a major goal of this investigation to be able to analytically reproduce the general operating characteristics of actual reformers. For this purpose an analytic version of the UC-Davis reformer was developed using the heat and chemical rate equations previously outlined. The virtual version of the UC-Davis reformer exhibited all of the major performance characteristics of the actual reformer to varying extents. The constants used for the virtual reformer are shown in Table 4-1 and were matched as best as possible, to measured or inferred UC-Davis reformer constants. Table 4-1. Values constants used to develop virtual UC-Davis reformer Name Value Used How it was determined Catalyst Bed Heat Capacity (C c ) 900 J/kg-K Inferred Catalyst Bed Thermal Conductivity (k eff ) 5.00 W/mK Estimated from Data Catalyst Bed Density (e c ) 1983. kg/m 3 Measured Length (L) 0.53340 m Measured Radius (R) 11.9 x 10 -3 m Measured # of Radial Elements (N R ) 5 Number of Axial Elements (N L ) 150 Time Step (t) 0.5 s The catalyst heat capacity was assumed to be equal to an average of the heat capacities of copper and zinc. ** The catalyst heat capacity was assumed to be equal to an average of the heat capacities of copper and zinc. *** The void factor used was 50% and the catalyst density was measured via a water displacement method for a single pellet. To determine the value for catalyst bed thermal conductivity (k eff ), an average was taken of the measured thermal conductivity of the UC-Davis reformer catalyst bed at different points during the warm up stage, when no fuel flow exists. Figure 4-13 shows an example of the data used a band heater zone in the UC-Davis reformer. The resulting steady-state carbon monoxide composition of the virtual UC-Davis reformer are shown in Figure 4-14. In arrows are listed the average steady-state CO

PAGE 65

53 compositions for the actual UC-Davis reformer. In this figure, note that greater flow rate results in lower steady-state CO concentrations. Error between the UC-Davis reformer steady-state CO concentration and that of the virtual reformer is potentially due to the existence of un-reacted methanol in the sampled gas. Although before sampling the reformate gas was cooled to extract condensate, the difference between methanol and water freezing temperature only permitted a reduction in temperature to almost 0 o C. The existence of unreformed methanol in the sampled gas is detectable through the methane (or hydrocarbon) signature detected by the NOVA gas analyzer (Figure 4-15). Zone 2 Thermal Conductivity200220240260280300320340012345Time (mins)Temperature (oC)0123456Calculated Thermal Conductivity (W/(m-K)) Zone 2 Surface Zone 2 Exit Wall Zone 2 Exit Centerline Effective Conductivity A verage Conductivity = 4.87 W/(K-m) Figure 4-13. Estimation of catalyst bed thermal conductivity using UC-Davis reformer data

PAGE 66

54 1.98E -02 1.98E -02 1.98E -02 1.98E -02 1.98E -02 1.98E -02 1.98E -02 1.98E -02 1.98E -02 1.98E -02 1. 96E -02 1. 96E -02 1. 96E -02 1. 96E -02 1. 96E -02 1. 96E -02 1. 96E -02 1. 96E -02 1. 96E -02 1. 96E -02 1.92E -02 1.92E -02 1.92E -02 1.92E -02 1.92E -02 1.92E -02 1.92E -02 1.89E -02 1.89E -02 1.89E -02 1.89E -02 1.89E -02 1.89E -02 1.89E -02 1.89E -02 1.89E -02 1.89E -02 1. 87E -02 1. 87E -02 1. 87E -02 1. 87E -02 1. 87E -02 1. 87E -02 1. 87E -02 1. 87E -02 1. 87E -02 1. 87E -02 1.85E -02 1.85E -02 1.85E -02 1.85E -02 1.85E -02 1.85E -02 1.85E -02 1.85E -02 1.85E -02 1.85E -02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 7 2 E 02 1. 68E -02 1. 68E -02 1. 68E -02 1. 68E -02 1. 68E -02 1. 68E -02 1. 68E -02 1. 68E -02 1. 68E -02 1. 68E -02 1.61E -02 1.61E -02 1.61E -02 1.61E -02 1.61E -02 1.61E -02 1.61E -02 1.61E -02 1.61E -02 1.61E -02 1.52E -02 1.52E -02 1.52E -02 1.52E -02 1.52E -02 1.52E -02 1.52E -02 1.52E -02 1.52E -02 1.52E -02 1 44E -0 2 1 44E -0 2 1 44E -0 2 1 44E -0 2 1 44E -0 2 1 44E -0 2 1 44E -0 2 1 44E -0 2 1 44E -0 2 1 44E -0 2 1.36E -02 1.36E -02 1.36E -02 1.36E -02 1.36E -02 1.36E -02 1.36E -02 1.36E -02 1.36E -02 1.36E -02 0. 00 00 0. 00 50 0. 01 00 0. 01 50 0. 02 00 0. 02 50 0 5 10 15 2 0 2 5 3 0 35 Prem ix Flow ( s t d .mL/m i n) C O (m o l e frac t i o n ) x CO (dry ) x C O ( d r y w / M e O H ) 1. 14% UC-Dav i s 0. 9 3 % UC-Dav i s Figure 4-14. Virtual UC-Davis reform er steady-state CO composition w i th varying prem ix fuel flow rates 0 2 4 6 8 10 12 0 1 0 2 03 04 05 0 6 07 0 8 Run Ti me (min ) Signal O u tput 0 Figure 4-15. Hydrocarbon signal output obtained during UC-Davis reform er run.

PAGE 67

55 Figure 4-16 shows a representative CO spike at the point in which the premix flow rate changes from 10 to 15 ml/min. The change in flow rate was produced at time zero. This figure is meant to be a parallel to Figure 4-10. Figures 4-16 and 4-17 show the temperature distributions of the modeled reformer for 5 ml/min and 30 ml/min flow rate, respectively. The temperature distribution is similar to those recorded for the UC-Davis reformer (Figure 4-4), where the minimum temperature is recorded at the reformer entrance region. These results support the observations made from data obtained from the UC-Davis reformer and therefore demonstrate the validity of the assumptions made in the model presented. Virtual Reformer CO Spike (10 to 15ml/min)0.01580.01590.0160.01610.01620.01630.01640.01650102030405Time After Point of Flow Rate Change (s)CO Mole Fraction 0 Figure 4-16. Virtual UC-Davis Reformer CO spike from 10 to 15ml/min premix flow rate change

PAGE 68

56 Figure 4-17. Virtual UC-Davis reformer temperature distribution (color plot) and methanol concentration (contour plot) for premix flow rate equal to 5 ml/min Figure 4-18. Virtual UC-Davis reformer temperature distribution (color plot) and methanol concentration (contour plot) for premix flow rate equal to 30 ml/min

PAGE 69

57 Reformer Design Currently, most reformer design is done based on a space velocity basis. This method is not consistent with the complex nature of the reforming reaction and the virtually infinite number of possible reformer designs. As part of this study, a set of non-dimensional reformer performance variables are introduced in order to provide generality to the results obtained from analytical modeling and experiments. This section generally describes some parameters that affect reformer performance. In particular the question of how reformer length, catalyst loading and fuel flow rate affects the reformate gas composition is investigated using these suggested non-dimensional parameters. Design Generalization It has already been established that heat transfer and chemical rate characteristics are major determinants of complex reformer performance. Equation 4-5 can be rewritten as Equation 4-29. MrceffeffeffcrE x Tk r Tk r T r ktTC 2222 (4-29) In order to generalize the design of reformers, it is useful to define non-dimensional variables. For a reformer tube of a certain radius (R), the volume of catalyst in the reformer (V c ) will determine the length of the reformer (L). According to Nakagaki and other correlations it could be said that the mass of the reformer has a potential reformation capacity, The potential reformation capacity of the catalyst could be defined as the theoretical hydrogen power (P H2 ) that a certain mass of catalyst could produce if the reformer were maintained exactly at the wall temperature, and with the inlet methanol concentration. This can be easily calculated by using the Nakagaki

PAGE 70

58 correlation. For example, the potential reformation capacity of the UC-Davis reformer can be calculated in the following way. The catalyst mass in the UC-Davis reformer is kgmmmkgLRVmccc470565.05334.00119.01983232 (4-30) If the wall temperature were to remain constant at 300 o C, then r M,max and the potential rate of methanol reacted ()can be calculated as shown in Equation 4-31. max,,3reactedOHCHn sec108921.4sec100396.1,5max,,34,3max,kmolmrnkgkmolxTrcMreactedOHCHinletOHCHwM (4-31) Since the theoretical chemical reaction for the reformation of methanol yields three moles of hydrogen per mole of methanol, the potential reformation capacity () can be obtained as shown in Equation 4-32. kWkmolkmolkJnLHVreactedOHCHH4609.35sec108921.4241620335max,,32 (4-32) Also a non-dimensional temperature, can be defined as wTT (4-33) and a non-dimensional length and radius can be defined as R rr* (4-34) Lxx* (4-35)

PAGE 71

59 Given these formulations, Equation 4-29 can be non-dimensionalized to yield VrE x VTR r R kVT r r R kVTtTCVMrcweffweffwwcc2*242*22**21 (4-36) Equation 4-38 is made up of various important non-dimensional groups. A non-dimensional conductivity, k can be expressed as 2* R kVTkeffw (4-37) The non-dimensional conductivity can be construed as a measure of the ratio between the potential radial conduction heat transfer into the reformer and the potential rate of energy absorbed by the reforming reaction. It is important to point out that the non-dimensional conduction is not actually a function of catalyst length. The non-dimensional conduction could be expressed as 22*3RLHVkTkcHeffw (4-38) For the UC-Davis reformer the non-dimensional conduction is equal to 0.135. The last term in Equation 4-36, is a measure of the ratio of the actual rate of energy absorbed due to the reformation reaction and the potential rate of energy absorbed by the same reaction in terms of hydrogen production. The potential reformation capacity, is actually a measure of the potential hydrogen production of a certain mass of catalyst at a certain temperature. Thus it would be convenient to express the last term in Equation 4-36 in terms of hydrogen production and not methanol consumption. In practical terms the hydrogen production is easier to measure than the methanol converted.

PAGE 72

60 With this formulation a catalyst effectiveness, can be defined as the ratio of the actual hydrogen power output obtained from a certain amount of catalyst (P H2 ) and the potential reformation capacity of the catalyst () (Equation 4-39). 2HP (4-39) The catalyst effectiveness is a measure of the degree of reforming activity that a certain mass of catalyst undergoes. High catalyst effectiveness is desirable because it means that the reformer is operating closer to its idealized potential hydrogen yield. High catalyst effectiveness occurs in short reformers. On the other hand, the larger the reformer, the lower the catalyst effectiveness will be. From Equation 4-40, it is also notable the time constant term, max,23MHwcwccrLHVTCTCV (4-40) The transient response in a reformer would be strongly determined by A low value of would lead to very fast temperature changes in the reformer tube leading to a decrease in prominence of the CO spikes. For the purpose of fully expressing the heat equation in non-dimensional terms a non-dimensional time (t ) is defined (Equation 4-41). wccTCVtt* (4-41) The non-dimensional heat equation is max,22*242*2*****3MHMrwrLHVrExVTRrkrrkt (4-42) It is also convenient to non-dimensionalize the fuel flow into the reformer as

PAGE 73

61 fuelfuelNDmLHVflow (4-43) where LHV fuel represents the lower heating value of the fuel, and represents the mass flow rate of the fuel. The non-dimensional flow is a ratio between the rate of fuel energy introduced into the reactor and the reactors reformation capacity. The non-dimensional flow can be expressed in terms of space velocity (S fuelm v ) as shown in Equation 4-43. cMinfuelvNDrSflow max,,3 (4-44) Also useful is the definition of reformer efficiency, ref fuelHrefPP2 (4-45) The chosen reformer efficiency definition neglects the power required to promote and sustain the reforming reaction. This definition only deals with issues of conversion, since a detailed analysis of reformer heating design is outside the scope of this study. In Equation 4-42, represents the rate of hydrogen chemical energy obtained from the reformer, and represents the rate of fuel energy introduced into the reformer. The maximum reformer efficiency for a methanol reformer is given by 2HP fuelP OHCHHrefLHVLHV32max,3 (4-46) The non-dimensional flow can be re-written as max,,,3,,3max,1reactedOHCHinOHCHrefNDnnflow (4-47)

PAGE 74

62 Given these definitions, the reformer model was used to obtain characteristic reformer efficiencies versus catalyst effectiveness curves for reformers with the same non-dimensional conduction (0.315) as the UC-Davis reformer at various fuel flow rates and different lengths. These curves are shown in Figure 4-19. Additionally Figure 4-19 shows the corresponding non-dimensional fuel flow rates as a function of catalyst effectiveness. 02040608010000.20.40.60.81Catalyst Effectiveness01234567 Efficiency (5ml/min) Efficiency (10ml/min) Efficiency (15ml/min) Efficiency (20ml/min) Flow ND (5ml/min) Flow ND (10ml/min) Flow ND (15ml/min) Flow ND (20ml/min)Efficiency (%)Non-dimensional Flow Magnitude represented on secondary axis Figure 4-19. Reformer Efficiency with varying catalyst effectiveness at varying fuel flow rates From Figure 4-19 certain observations can be drawn. First, for non-dimensional fuel flow values less than 1, the variations in reformer efficiency versus catalyst effectiveness and non-dimensional fuel flow versus catalyst effectiveness, collapse into a single representative performance curve. Second, as mentioned previously, lowered catalyst effectiveness results in higher reformer efficiency. In other words, longer

PAGE 75

63 reformers will exhibit higher steady-state efficiencies at the cost of ever increasing catalyst mass. Therefore, there are diminishing benefits to a long reformer, which will have a similar hydrogen yield as a reformer designed with catalyst effectiveness equal to 0.1. As the fuel flow rate of a reformer approaches infinity, the catalyst effectiveness will approach 1 and the reformer hydrogen yield will approach zero. These results can be used to scale a reformer using non-dimensional variables. For the purposes of this study, three reformer designs were considered: an infinite reformer with catalyst effectiveness equal to 0.05 at maximum power; a mid-size reformer with catalyst effectiveness equal to 0.1 at maximum power; and a small reformer with catalyst effectiveness equal to 0.2 at maximum power. The terms small, medium and large are used as visual clues to the general aspect ratio that these reformers would have in comparison with each other if the diameter were to remain the same. All reformers were designed to produce a maximum 60 kW of hydrogen (lower heating value based). If a reformer tube with the same non-dimensional conduction as the UC-Davis reformer were to be scaled to produce 60 kW of power, its length and steady-state operating characteristics would be as shown in Table 4-2, based on steady-state catalyst effectiveness and reformer efficiency analysis. Table 4-2. Steady-state design of 60 kW reformer with an 11.9 mm radius and a catalyst bed density of 1983 kg/m 3 Catalyst Effectiveness () Reformer Tube Length (m) Steady Sate Reformer Efficiency (%) Premix Power to Produce 60 kW of H 2 (kW) 0.05 18.05 112.9 53.14 0.10 9.03 112.5 53.35 0.20 4.51 110.5 54.28 D. Davieau [16] studied the use of space velocity and aspect ratio as major similitude parameters for reformer design. His conclusions are that these two parameters

PAGE 76

64 are not sufficient to fully characterize reformer steady-state conversion and reformate quality. These conclusions are supported by the analysis presented in this study. The space velocity and aspect ratio, although important, do not capture the governing physics of the reformation process.

PAGE 77

CHAPTER 5 RESULTS In this chapter, the primary chemical and heat transfer conditions that produce variations in hydrogen and carbon monoxide levels in the reformate are presented and discussed. Special attention is given towards the benefits and detriments associated with reformer size and catalyst loading in the context of fuel cell system incorporation. For this purpose three general reformer length with the same non-dimensional conduction were compared based on their reformate quality. Transient Efficiency Case 1 Infinite Reformer Figure 5-1 shows predicted transient changes in reformer efficiency for various changes in non-dimensional fuel flow rates for the infinite reformer. This figure shows how increasing fuel flow results in an immediate drop in reformer efficiency followed by a gradual approach to the steady-state efficiency. Figure 5-2 shows the same effects on the reformer efficiency in terms of the efficiency difference variable, ref ~ defined as .., ~ PSrefrefref (5-1) The variable ref,S.P. is the set point efficiency, which is the steady-state efficiency at a certain flow rate. In other words, for any steady flow rate the difference efficiency is equal to zero. The magnitude of the initial efficiency overshoot is summarized in Table 5-1. It is notable that the magnitude of the efficiency-overshoot increases with increased change in 65

PAGE 78

66 non-dimensional flow. However, if the change in non-dimensional flow remains constant, changes between smaller non-dimensional flow quantities results in less overshoot. The magnitude of the drop in efficiency before reaching steady-state is also recorded. This is called an efficiency undershoot. High overshoot magnitudes reduced the extent of the efficiency undershoot. Table 5-1. Case 1: Summary of the transient response of the infinite reformer to changes in fuel flow rate Flow ND Initial Final Change in ND Flow Overshoot 1.10 2.20 1.10 0.019% 2.20 3.31 1.10 0.027% 3.31 4.42 1.10 0.046% 4.42 5.52 1.10 0.064% 5.52 6.63 1.10 0.072% 1.10 6.63 5.53 0.260% Infinite Reformer Transient Efficiency122.6122.7122.8122.9123.0123.1123.2123.3020406080Run Time (s)Efficiency (%) 100 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key: 10kW: 1.10%20kW: 2.20%30kW: 3.31%40kW: 4.42%50kW: 6.63% Figure 5-1. Case 1: Infinite reformer operation when undergoing step variations in fuel flow rate

PAGE 79

67 Infinite Reformer Transient Efficiency-0.3-0.2-0.10.00.10.20.3020406080100Run Time (s) 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key: 10kW: 1.10%20kW: 2.20%30kW: 3.31%40kW: 4.42%50kW: 6.63% Figure 5-2. Case 1: Reformer efficiency changes due to change of fuel flow rate for the infinite reformer Case 2 The Medium Sized Reformer As with the infinite reformer, the medium sized reformer was subjected to transient variations in fuel flow. The changes in efficiency of the reformer are depicted in Figure 5-3 and 5-4. Figure 5-4 shows these changes in terms of the efficiency difference variable. Table 5-2 summarizes these results. As with the infinite reformer, large changes in non-dimensional flow result in large overshoots, and changes between lower non-dimensional flows results in lower overshoots. The magnitude of the efficiency undershoots for the medium reformer were significantly lower than those of the infinite reformer. In addition, as changes in non-dimensional flow rates occurred between higher non-dimensional flow rates, the efficiency undershoot was decreased and even disappeared.

PAGE 80

68 Table 5-2. Case 2: Summary of the transient response of the medium size reformer to changes in fuel flow rate Flow ND Initial Final Change in ND Flow Overshoot 2.22 4.43 2.22 0.066% 4.43 6.65 2.22 0.129% 6.65 8.87 2.22 0.211% 8.87 11.08 2.22 0.292% 11.08 13.30 2.22 0.372% 2.22 13.30 11.08 1.406% Medium Reformer Transient Efficiency121.4121.6121.8122122.2122.4122.6122.8123123.2123.4020406080Run Time (s)Efficiency (%) 100 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key:10kW: 2.22%20kW: 4.43%30kW: 6.65%40kW: 8.87%50kW: 11.08%60kW: 13.30% Figure 5-3. Case 2: Transient efficiency of the medium size reformer under transient load changes Case 3 The Short Reformer As with the infinite and medium sized reformer, the short reformer was run under transient load conditions as shown in Figure 5-5 and 5-6. In contrast with the larger reformers, the short reformer did not exhibit efficiency undershooting, except between step changes in non-dimensional flow between 4.44% and 8.88%. Besides this effect, the

PAGE 81

69 results follow a similar pattern to those obtained for the infinite and medium reformers. The results of the transient run of the short reformer are given in Table 5-3. Medium Reformer Transient Efficiency-0.200.20.40.60.811.21.41.6020406080Run Time (s) 100 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key:10kW: 2.22%20kW: 4.43%30kW: 6.65%40kW: 8.87%50kW: 11.08%60kW: 13.30% Figure 5-4. Case 2: Reformer efficiency changes due to change of fuel flow rate for the medium size reformer Table 5-3. Case 3: Summary of the transient efficiency response of the short reformer to changes in fuel flow rate Flow ND Initial Final Change in ND Flow Overshoot 4.44 8.88 4.44 0.36% 8.88 13.31 4.44 0.73% 13.31 17.75 4.44 1.06% 17.75 22.19 4.44 1.27% 22.19 26.63 4.44 1.29% 4.44 26.63 22.19 9.33%

PAGE 82

70 Short Reformer Transient Efficiency112114116118120122124020406080Run Time (s)Reformer Efficiency (%) 100 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key:10kW: 4.94%20kW: 8.88%30kW: 13.31%40kW: 17.75%50kW: 22.19%60kW: 26.63% Figure 5-5. Case 3: Transient efficiency of the short reformer under transient load changes Carbon Monoxide Concentration Generally, temporary increases in CO concentration in the reformate gas occurred during periods of increased premix flow rates. This phenomenon occurs because lower fuel flow into the reformer leads to higher average steady-state temperature. This increase in temperature results from less fuel being reformed, thus lower overall associated endothermicity. High temperatures promote greater CO and water formation in the water gas shift reaction. High temperatures typically occur near the exit of the reactor. When there is an increase in fuel flow into the reformer, the average temperature of the reactor does not drop immediately and this contributes to better transient conversion (i.e. increased efficiency), and higher transient CO concentrations. As time progresses the average temperature of the reactor drops because of the increased endothermicity

PAGE 83

71 associated with the increased level of methanol reacted. Eventually the lower average reformer temperature results in a lower steady-state CO level at higher fuel flow rates. An off-steady-state CO concentration is defined as steadysteadyOffCOCOCO (5-2) Short Reformer Transient Efficiency-0.200.20.40.60.811.21.4020406080Run Time (s) 100 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW ND Flow Key:10kW: 4.94%20kW: 8.88%30kW: 13.31%40kW: 17.75%50kW: 22.19%60kW: 26.63% Figure 5-6. Case 3: Reformer efficiency changes due to change of fuel flow rate for the short reformer Figures 5-7 through 5-13 show the transient CO concentrations and off-steady-state CO concentrations for the infinite, medium, and short reformer triggered by step changes in fuel flow into the reformer. The CO spike is defined as the initial increase in CO concentration that occurs after a step increase in fuel flow occurs. This spike occurs before there is a gradually decreasing approach of the transient CO concentration towards the steady-state CO concentration. In all instances, the larger the non-dimensional flow change the greater the

PAGE 84

72 CO spike. However, between equivalent changes in non-dimensional flow, larger CO spikes occurred between lower non-dimensional flows. For example, in the infinite reformer, a CO spike 328 ppm larger is recorded for a change in non-dimensional flow from 1.10% to 2.21% as compared to a change from 2.21% to 3.31%. CO spikes, as defined here, are a measure of the transient deviation of the CO concentration as compared to the steady-state condition for the same flow. This deviation was highest for small reformers, which in order to obtain a certain steady-state hydrogen flow rate required a large change in non-dimensional flow. However, if Figures 5-8, 5-10 and 5-12 are examined it can be seen that the off-steady-state CO concentration also includes the unsteady, smooth drop in CO concentration associated with the increased fuel flow into the reformer. Changes in flow between low non-dimensional flows caused greater disruptions to this smooth drop. In a fuel cell system these flow changes may cause variations in fuel cell performance and a degree of CO poisoning during the transient event if appropriate actions are not taken. From this standpoint, fuel flow variations in reformers operating with high non-dimensional flows (i.e. small reformers) should exhibit greater linearity in power output during load changes. The short reformer, the change of non-dimensional flow from 22.19% to 26.63% exhibited a very mild CO drop disruption. In Figure 5-13, the CO concentrations associated with a similar change in reformate hydrogen power output, namely from 40 kW to 50 kW (lower heating value based), is depicted. For this case, the short reformer produced a lower steady-state and transient CO concentration change than the larger reformers. This occurs because the catalyst

PAGE 85

73 effectiveness of the small reformer is much higher than for the larger reformers, which leads to lower average catalyst bed temperatures. Infinite Reformer Transient CO Concentration0.01700.01750.01800.01850.01900.0195020406080Run Time (s)CO (wet fraction) 100 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key: 10kW: 1.10%20kW: 2.20%30kW: 3.31%40kW: 4.42%50kW: 6.63% Figure 5-7. Infinite reformer transient CO concentrations for various for changes in power output Infinite Reformer CO Spikes-5000500100015002000020406080100120140Time (s)CO Concetration (ppm) 50 to 60kW 40 to 50kW 30 to 40kW 20 to 30kW 10 to 20kW 10 to 60kW ND Flow Key:10kW: 1.10%20kW: 2.21%30kW: 3.31%40kW: 4.42%50kW: 5.52%60kW: 6.63% Figure 5-8. Infinite reformer transient off-steady-state CO concentrations for various changes in reformer power output

PAGE 86

74 Medium Reformer Reformate CO Concentration0.01650.01700.01750.01800.01850.01900.01950.0200020406080100120140Time (s)CO Concentration (wet fraction) 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key:10kW: 2.22%20kW: 4.43%30kW: 6.65%40kW: 8.87%50kW: 11.08%60kW: 13.30% Figure 5-9. Medium reformer transient CO concentrations for various changes in power output Medium Reformer CO Spikes050010001500200025003000020406080100120140Time (s)CO Concentration (ppm) 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW 10 to 60kW ND Flow Key:10kW: 2.22%20kW: 4.43%30kW: 6.65%40kW: 8.87%50kW: 11.08%60kW: 13.30% Figure 5-10. Medium reformer off-steady-state CO concentrations for various changes in reformer power output

PAGE 87

75 Short Reformer Transient CO Concentration0.0150.01550.0160.01650.0170.01750.0180.0185020406080100120140Time (s)CO Concentration (wet fraction) 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 50 to 60kW ND Flow Key:10kW: 4.94%20kW: 8.88%30kW: 13.31%40kW: 17.75%50kW: 22.19%60kW: 26.63% Figure 5-11. Short reformer transient CO concentrations for various changes in power output Short Reformer CO Spikes050010001500200025003000020406080100120140Time (s)CO Concentration (ppm) 10 to 20kW 20 to 30kW 30 to 40kW 40 to 50kW 10 to 60kW ND Flow Key:10kW: 4.94%20kW: 8.88%30kW: 13.31%40kW: 17.75%50kW: 22.19%60kW: 26.63% Figure 5-12. Short reformer off-steady-state CO concentration for various changes in reformer power output

PAGE 88

76 Comparison of CO Concentrations for Different Size ReformersOperated Between 40 to 50kW Steady State Power Outputs0.01550.0160.01650.0170.01750.018020406080100120140Run Time (s)CO Concentration (ppm) Medium CO Short CO Infinite CO ND Flow ranges: Short Reformer: 17.75% to 22.19%Medium Reformer: 8.87% to 11.08%Infinite Reformer: 1.10% to 6.63% Figure 5-13. Transient CO comparison of various size reformers operating between 40 and 50 kW hydrogen power output Hydrogen Concentration Transient and steady-state changes in hydrogen concentration for all reformers were such that the performance of PEMFCs or PAFCs should not be drastically affected. Even under the harshest transient conditions studied, in which a change in fuel flow was induced to generate a steady-state hydrogen yield between 10 kW and 60 kW, the reformers hydrogen yields did not deviate much more than 5% (Figure 5-14). The lowest hydrogen concentration was slightly below 60% wet molar concentration, and corresponded to the small reformer producing 60 kW hydrogen yield.

PAGE 89

77 0.590.600.610.620.630.640.650.66050100150200250Run Time (s)H2 Concentration Infinite (10kW to 60kW) Medium (10kW to 60kW) Short (10kW to 60kW) ND Flow Changes: Infinite: 1.10 to 6.63Medium: 2.22 to 13.30Short: 4.44 to 26.63 Figure 5-14. Comparison of transient changes in H 2 concentration induced by step changes in fuel flow corresponding to hydrogen yields from 10 to 60 kW (lower heating value based) for various size reformers

PAGE 90

CHAPTER 6 CONCLUSIONS Based on the experimental and modeling results presented in this study the following conclusions may be expressed. Steam Reformer Modeling A method of modeling a methanol steam reformer was proposed, demonstrated and compared to actual reformer performance The reformer model results matched general reformer behavior observed in actual reformers Based on this model formulation a new set of reformer similarity variables were proposed and the use of space velocity and aspect ratio as reformer design variables was demonstrated to be insufficient Steady-state Reformer Operation Non-dimensional flow, catalyst effectiveness, and non-dimensional conduction were introduced as potential reformer similarity variables. These variables effectively characterized the steady-state reformer performance of reformers. Based on these variables, the results showed that the steady-state CO concentration decreases with increased non-dimensional flows The steady-state H 2 concentration decreased with increased non-dimensional flows Higher non-dimensional flows resulted in decreased steady-state reformer conversion Unsteady Reformer Operation Transient changes in fuel flow into steam reformers produced non-linear variations in hydrogen and CO concentration in the reformer outlet These variations are mostly caused by transient changes in reformer catalyst bed temperature 78

PAGE 91

79 Step increases in non-dimensional fuel flow produced a transient increase in CO concentration followed by an eventual drop in the CO concentration This transient increase in CO concentration was greater when large changes in non-dimensional flow occurred Step increases between large non-dimensional flows produced lower transient increases in CO concentration than the same step changes between lower non-dimensional flows. The implications of the results and methodology presented in this study are that steam reformer chemical behavior, in the transient and steady-state case, has been expressed in mathematical, non-dimensional, terms. This leads to the proposal of similitude expressions that can be used to better characterize, compare, understand and design packed bed steam reformers. Also, this study demonstrates the mechanisms that produce carbon monoxide spikes during period of flow changes. Generally, from a design standpoint, there seems to exist a balance between the design of an infinitely long reformer that would have excellent efficiency (conversion) but would also have high outlet CO concentrations in the steady state and transient regime, and a short and stubby reformer with low efficiency but low transient and stead sate CO concentrations for the same fuel flow rate. The optimum point will depend on the fuel cell and its tolerance of CO. Design should be made such that during the highest changes in non-dimensional flow, the CO spike does not exceed a maximum concentration that would drastically reduce fuel cell performance through poisoning.

PAGE 92

CHAPTER 7 SUGGESTED FUTURE WORK The analysis provided in this study is largely based on experimental results obtained from a single diameter reformer (UC Davis reformer). The numerical models developed generally matched the performance of this reformer. A greater body of experimental data, with reformers of varying lengths and diameters should be obtained to determine the practical limitations of the analysis presented here. In addition, a set of non-dimensional performance variables has been introduced to the field of reformer design and analysis. These variables worked extremely well at describing reformer performance based on the models developed. A study should be conducted to evaluate whether these variables (or a variation of these variables) are capable of determining similarity between distinct reformers. Expansion of this study into reformer controls design, fuel cell controls design, and overall fuel cell system controls design is a logical expansion of this body of work. Understanding the transient nature of fuel cell systems is critical to their effective control. It is hoped that this work forms the foundation for the development of controllers that are capable of producing fast transient load changes in future fuel cell systems with minimal efficiency penalties. 80

PAGE 93

APPENDIX A COPY OF FEMLAB SOLUTION REPORT 81

PAGE 94

82

PAGE 95

83

PAGE 96

84

PAGE 97

85

PAGE 98

86

PAGE 99

87

PAGE 100

88

PAGE 101

89

PAGE 102

90

PAGE 103

91

PAGE 104

92

PAGE 105

93

PAGE 106

94

PAGE 107

95

PAGE 108

96

PAGE 109

97

PAGE 110

98

PAGE 111

APPENDIX B PRELIMINARY LOAD FORECASTING STUDY The lose of efficiency due to the sluggish response of the fuel cell system when operating in a transient load environment may be mitigated by the capability to preempt the onset of large changes in load. For a vehicle this may be achieved by learning the driving style of the vehicle driver such that periods of increased acceleration may be anticipated. Figure B-1 shows a typical load profile for a drive cycle. This load profile was obtained by driving TBB-2 in actual traffic conditions. Transient Load Profile for TBB-2 under the UFFL driving cycle-10010203040500100200300400500600700Run Time (s)Power (kW) Figure B-1. Transient load profile for TBB-2 under the University of Florida Fuel Cell Lab driving cycle data obtained on 4-2-2004. 99

PAGE 112

100 In order to test this hypothesis TBB-2 was driven under a standard route (UFFL cycle) and the transient load, L(t), was obtained. Given this load a power change index, PCI(t), was defined (Equation B-1). )()()()(ottLtLtLtPCI (B-1) In Equation B-1, t o is the sample interval in time. Thus L(t-t o ) represents the previous load data. For the data obtained from TBB-2, t o equaled 1 second. Notice that PCI is equal to 0.5 if there is no load change and that its value is bound between 1 and -1. PCI for TBB-2 under UFFL driving cycle-1-0.75-0.5-0.2500.250.50.7510100200300400500600700Run Time (s)PCI Figure B-2. Power Change Index (PCI) for TBB-2 under UFFL driving cycle data obtained on 4-2-2004. By looking at Figure B-2 it is evident that most of the time PCI oscillates slightly from 0.5, meaning that relatively small variations in load are occurring. Furthermore it is shown that large fluctuations in the value of PCI occur in a recognizable pattern. That is,

PAGE 113

101 every large decrease in the value of PCI (from breaking or stopping), there is almost always a consequent large increase in the PCI value. f(net) f(net) f(net) f(net) Hidden LayerInput Layer f(net) f(net) f(net) f(net) Output Layer x (t-20)x (t-19)x (t-18)x (t-17) x (t-0)y1y2y3yN f(net) f(net) f(net) f(net) Hidden LayerInput Layer f(net) f(net) f(net) f(net) Output Layer x (t-20)x (t-19)x (t-18)x (t-17) x (t-0)y1y2y3yN Figure B-3. Multi-layer Perceptron. Out of ten runs PCI data from one of the runs was used to train a multilayer perceptron (MLP) whose input were 20 PCI data points. The MLP used had 10 output processing elements, whose output should approximate PCI(t+t o ), PCI(t+2t o ), PCI(t+3t o ), PCI(t+10t o ). Figure B-3 shows a diagram of the MLP used. The output of each processing element (PE) is the hyperbolic tangent of the sum of the multiplication of the inputs to a gain or weight (Equation B-2). See Figure B-3. ijiijwxPE,tanh (B-2) Therefore, the output of an output layer processing element can be written as shown below, where k is the output index, i is the input data index, and j is the hidden processing element index (Equation B-3).

PAGE 114

102 kjkijiikoutputwwxPE,,,tanhtanh (B-3) Training of the MLP was done using the error back propagation algorithm. An error was defined as the difference between the MLP output and the desired MLP output (Equation B-4). For the purposes of these experiments, the desired output was the future data, x(t+kt o ) that the k th output of the P.E.should approximate. koutputoikPEkttxE,,)( (B-4) From this error definition a mean squared error (J) can also be defined (Equation B-5). NiMkikENJ112,21 (B-5) In Equation B-5 N represents the number of data points introduced into the MLP (for our purposes N=20). M represents the number of output processing elements in the MLP (in our case 10). The mean squared error is used as a cost function whose minimum is found with respect to the weights using a gradient descent algorithm. The algorithm chosen was the error back propagation algorithm. The trained MLP was tested with new data obtained by driving TBB-2 under the same UFFL route but on a different day and under different traffic conditions. Figure B-4 shows the performance of the MLP as a load predictor. Generally, the trained MLP was able to predict large increases in load.

PAGE 115

103 MLP Performance Test-1-0.75-0.5-0.2500.250.50.7510100200300400500600Run TimePCI Actual Run Prediction (t+10s) Figure B-4. The MLP performance as a 10 second PCI predictor.

PAGE 116

APPENDIX C PHYSICAL DESCRIPTION OF THE UC-DAVIS REFORMER The following is a physical description of the UC-Davis reformer test rig. Part of the information presented in this Appendix was provided and prepared by UC-Davis students and Dr. Paul Erickson, director of the UC-Davis Hydrogen Production and Utilization Laboratory (HYPAUL). The UC-Davis methanol-steam reformer used for the experimental portion of this study can be divided into subassemblies as shown in Table C-1. Figure C1 shows a general process diagram of the UC-Davis reformer. Table C-1. UC-Davis reformer subassemblies Methanol Steam Reformer Elements Subassembly Premix Reservoir, Pump, Scale Pumping 3-Stage Vaporizer, Super-heater Vaporizer Steam Reforming Reactor Catalyst Bed Housing Condenser, Condensate Trap Condensing Unit General Description A premixed concentration of methanol in water is mounted on a scale. This premixed fuel is pumped through three stage vaporizers which ensure only vapors are introduced into the reformer tube at the user specified temperature. In order to ensure this a superheater is also inline. The superheater ultimately controls the fuel temperature before introduction into the reformer catalyst bed. 104

PAGE 117

105 Figure C-1. General process diagram of the UC-Davis reformer After passing through the catalyst bed, the gaseous species were directed into two tubes via a system of valves. One route was for analysis and the other was for exhausting the reaction products, reformate. Both routes lead to the condensing unit where the reformate temperature is reduced in order to separate liquid water and unreacted methanol from the gas mixture. A condensate trap is used for this purpose. The generally dry reformate gas is then routed to the gas analyzer. The system also includes the capability to drain the catalyst bed through a series of drain valves and allow for carbon monoxide purging and preservation of the catalyst bed. An air purge was adapted after the catalyst section to ensure retrieval of all condensed species in the condensate trap.

PAGE 118

106 Pumping Subassembly The pumping subassembly (Figure C-2) is initially composed of a 4 liter (1 gal.) polyethylene carboy reservoir containing a user prepared molar concentration of methanol in water. For the experiments conducted a methanol concentration of 1.5 moles of water per mole of methanol was used. The tank was allowed to rest at room temperature. Appropriate premix concentration was verified through the use of hand-held specific gravity sensor (Anton Parr model 35n with a resolution of 0.0001 g/cm3). The premix reservoir rested upon a 4100 gram-Ohaus scale with a 0.1 gram resolution. The scale had a 9 pin bidirectional RS-232 port, which allowed the user to electronically record the scale reading during operation. The premix was then drawn out of the reservoir by way of a gear pump through a 0.635 cm (0.25 in) I.D. vinyl tube. The gear pump head (model EW-07002-25) and driver (model A-75211-30) was manufactured by MicroPump and enabled the user a resolution of 0.1 ml/min with a premix flow rate range from 2.6 to 85 ml/min. The pump driver was equipped with a frequency output signal, which allowed the user to correspond a frequency (or gear pump RPM) to a flow rate. The user can then electronically record the instantaneous flow rate and control the pump with a voltage signal during operation. Calculations for the mass flow rate could be verified by both the recorded pump flow rate and by recording the change in mass of the scale divided by the time the experiment ran (both were recorded via the computer control program).

PAGE 119

107 Figure C-2. Premix reservoir and gear pump Figure C-3. Vaporizer design The vaporizer subassembly, composed of a 3-stage vaporizer and a superheater, is shown in Figure C-3 and Figure C-4, respectively. The premix exited the gear pump through a 0.635 cm (0.25 in) O.D. stainless-steel tube (0.124 cm (0.049in) wall thickness) and then entered the first vaporizer of the 3-stage vaporizer. All 3 vaporizers were built with identical dimensions. Each vaporizer was made of a 20.3 cm (8 in) stainless-steel pipe (nominal Dia., schedule 40) threaded at both ends. Caps for the pipes were machined to adapt a 0.635 cm (0.25 in) tube on the top end and a 0.635 cm (0.25 in)

PAGE 120

108 cartridge heater on the bottom end. The energy for vaporization was supplied from 120V cartridge heaters. The first vaporizer contained a 24.1 cm (9.5 in), 525W cartridge heater, while the last two stages contained 12.7 cm (5 in), 400W cartridge heaters. Each vaporizer was monitored for temperature by two, stainless-steel-sheathed, ungrounded K-type thermocouples. The temperatures of the cartridge heaters were monitored by a 0.025 cm (0.010 in) diameter-thermocouples. These thermocouples were located external to the flow, attached to the base of the cartridge heater (Figure C-3). The exit temperature of each vaporizer was monitored by a 0.159 cm (0.0625 in) thermocouple. These thermocouples were located internal to the flow, at the joint union between each stage (Figure C-3). At the exit of the second vaporizer a 0-103.4 kPa (0-15psi) pressure gauge (identified as PT in Figure C-3) was installed to monitor the pressure near the beginning stage of the reformer. Several stainless-steel, high temperature quarter-turn valves were installed throughout the vaporizer subassembly. At the inlet to the first vaporizer, a valve was adapted to drain the condensate in the subassembly when necessary and at the exit of the first vaporizer a valve permitted an inlet for catalyst reduction gas. When the vaporized premix was not up to temperature, an exhaust valve at the exit of the 3-stage vaporizer prevented the flow from entering the superheater and catalyst bed housing, as shown in Figure C-4. A CO 2 purge and check-valve were located at this position as well, to purge vaporized premix out of the reactor and lock oxygen out when not in operation. The superheater housing material was a 30.5 cm (12 in) stainless-steel pipe (nominal Dia., schedule 40) threaded at both ends. Caps for the pipes were machined to adapt to 0.635 cm (0.25 in) tubes and could be removed. To prevent condensation build-up on the bottom of the superheater, which would funnel the liquid premix directly

PAGE 121

109 into the reactor (thus damaging the catalyst); a recessed tube was inserted approximately 7.6 cm (3 in) inside the superheater. All pipes were sealed on the vaporizer subassembly by either welding or a sealant (Resbond 907GF), which was capable of withstanding temperatures up to 1288 o C (2350 o F). All tubes were sealed using Swagelok compression fittings. External heating was applied to the superheater using four nozzle band heaters (2.5cm (1in) I.D., 5.1 cm (2in) width), each with a 120V, 275W rating. To evenly increase the temperature distribution throughout the superheater, a highly thermal conductive aluminum tape was wrapped around the exterior. Three 0.159 cm (0.0625 in), stainless-steel-sheathed, ungrounded K-type thermocouples were strategically adapted to the superheater to monitor performance. Two thermocouples were integrated inside the superheater to monitor superheater internal and exit (just before the reactor) temperature. A third was placed externally on a nozzle band heater, permitting the user to monitor the temperature of the heat source. Insulation for both the 3-stage vaporizer and superheater composed of an alumina-silica material capable of withstanding temperatures up to1260 o C (2300 o F). Because the superheater utilized external heating, it required extra insulation composed of inorganic-volcanic rock fiber capable of withstanding temperatures up to 649 o C (1200 o F). Catalyst Bed Housing Subassemblies The housing material for the reactor was a 61 cm (24 in) stainless-steel pipe (nominal in Dia., schedule 40) threaded at both ends, as shown in Figure C5. The upper cap was machined to adapt a 0.635 cm (0.25 in) tube, as well as a 0.635 cm (0.25 in) MNPT pipe on the side. The MNPT pipe created a pathway to a Kistler pressure transducer. The cap at the exit of the reactor was a specially machined reducing T fitting. One end of the T attaches to a 0.635 cm (0.25 in) tube, while the other attaches

PAGE 122

110 to the 1.9 cm (0.75in) pipe. External heating was applied to reactor A using eight nozzle band heaters (2.5cm (1in) I.D., 5.1cm (2in) width), each with a 120V, 275W rating. In a similar fashion as the superheater, a highly thermal conductive aluminum tape was wrapped around the exterior of the pipe to evenly increase the temperature distribution throughout the reactor. An array of seventeen 0.159cm (0.0625in) Dia. stainless-steel-sheathed, ungrounded K-type thermocouples was used to monitor the temperature within the reactor. The thermocouples were attached to the reactor housing using 0.32cm (0.125in) MNPT to 0.32cm (0.125in) pipe fittings by Swagelok. All fittings were sealed using a chemical resistant-Teflon tape or by welding. To observe the temperature of the heat bands, eight 0.025 cm (0.010 in) Dia., ungrounded K-type thermocouples were placed between the heat bands and the exterior reactor wall. The reactor pressure was monitored using a 0-103.4kPa (0-15psi) pressure gauge and was located at the exit of the reactor (Figure C5). Insulation for the reactor was composed of a 3 thick calcium silicate material with a temperature tolerance of 649 o C (1200 o F). In both reactors, a stainless-steel mesh (256 squares per inch, 0.015in wire diameter) was placed at the bottom of the reactor to provide a base for the palletized catalyst. The catalyst used in this study was a pelletized commercial-grade copper-zinc catalyst on an alumina substrate. This catalyst (FCRM-2) was manufactured by Sud-Chemie and is recommended for an operating temperature range of 250-280oC (482-536 o F). Stored in an oxidized state, the catalyst needed to be reduced before steam reformation could occur. Reduction was done by flowing diluted concentrations of hydrogen in nitrogen for a period of 10 hours. Successful reduction was accomplished

PAGE 123

111 once temperature variations within the reactor were negligible. Information including the mass of catalyst loaded and the length of filled catalyst bed housing was recorded. The catalyst was cylindrical in shape and had dimensions consisting of 0.47 cm (0.187 in)-diameter and 0.25 cm (0.100 in)-thickness, as stated by the manufacturer. Schedule 40 SS Pipe0.824 ID, Nominal12 Length, Threaded Ends Figure C-4. Superheater design Condensing Unit Subassembly The condenser housing consisted of a modified 113liter (4.0 ft 3 ) refrigerator manufactured by Haier. The condenser was designed with two intakes, one for reactor exhaust and the other for analysis. To increase conduction and gas path length, the reacted species or exhaust entered into 0.635cm (0.25 in)O.D. copper tube coiled inside 10.2cm (4in) PVC canisters (2 canisters on the analysis side).

PAGE 124

112 Figure C-5. The UC-Davis reactor design Here, water from an ice bath doused the coils via a submersible pump; lowering the gas temperature from 250 o C (482 o F) to 0 o C (32 o F). The decrease in temperature promotes a phase change, causing water, methanol and other relevant species to condense. The condensate was then trapped in the container. The condensate from the exhaust side was removed from the trap via a drain located on the outside of the condenser. To acquire the condensate from the reactant species for analysis, the container was removed from the unit. The dry product gas on the analysis side was then routed to the gas analyzer, while the exhaust gas was directed to the fume hood. Two 0.159cm

PAGE 125

113 (0.0625in), stainless-steel-sheathed, ungrounded K-type thermocouples monitored the condensing unit temperatures, one inside the condenser, and one located directly in the flow at the exit of the analysis gas side.

PAGE 126

LIST OF REFERENCES [1] M. Cooper, Fuel Cell Market: Light Duty Vehicles, Fuel Cells Today, London, U.K., 2003. [2] D. Betts, V. P. Roan, J. Fletcher, Proc. Int. Mec. Eng. Cong. Exp., ASME 2000. [3] G. Thomas, Hydrogen Storage Overview, Sandia National Laboratory, N.M., 2003. ( http://www.eere.energy.gov/hydrogenandfuelcells/pdfs/bulk_hydrogen_stor_pres_sandia.pdf ) [4] F. Barbir, Clean Energy Research Institute, University of Miami, Coral Gables, FL. [5] T. Springer, T. Rockward, T. Zawodzinski, S. Gottesfeld, J. Elect. Soc., 148(1) (2001) A11-A23. [6] Office of Fossil Energy, Todays Hydrogen Production Industry Aug. 1 st 2005, U.S. DOE, 2005. ( http://www.fe.doe.gov/programs/fuels/hydrogen/currenttechnology.html ) [7] V.P. Roan, D.A. Betts, K. Dinh, A. Twining, T. Simons, U.F., An Investigation of the Feasibility of Coal-Based Methanol for Application in Transportation Fuel Cell Systems, Advanced Vehicle Development Program, Georgetown University, Washington, D.C., 2004. [8] M. El-Sharkh, A. Rahman, M. Alam, P. Byrne, A. Sakla, T. Thomas, J. Pow. Sources, 138 (2004) 199-204. [9] Y. Choi, H. Stenger, J. Pow. Sources. 142 (2005) 81-91. [10] B. Peppley, J. Amphlett, L. Kearns, R. Mann, Appl..Cat. A: Gen., 179 (1999) 21-49. [11] K. Lee, J. Ko, D. Kim, Appl. Cat. A: Gen., 278 (2004) 25-35. [12] T. Nakagaki, T. Ogawa, K. Murata, Y. Nakata, J. Eng. Gas Turb. and Pow., 123 (2001) 727-733. [13] D. Betts, P.A. Erickson, T. Simons, V.P. Roan, Proc. SAE Pow. Fluid Syst. Conf. Exp., Doc. Num. 2002-01-2855 (2002). [14] Horng, R., Ener. Conv. Manag., 46 (2005) 1193-1207. 114

PAGE 127

115 [15] B. Hohlein, M. Boe, J. Bogild-Hansen, P. Brokerhoff, G. Colsman, B. Emonts, R. Menzer, E. Reidel, J. Pow. Sources, 61 (1996) 143-147. [16] D. Davieau, An Analysis of Space Velocity and Aspect Ratio Parameters in Steam-Reforming Hydrogen Production Reactors, U.C. Davis, California, 2002. [17] F. Incoprera, D. DeWitt, Fundamentals of heat and mass transfer, fifth ed., John Wiley and Sons, New York, 2002. [18] M. Munagavalasa, K. Pillai, Int. J. Heat Mass Trans., (2005). [19] M. Brucker, J. Majdalani, Int. J. Heat Mass Trans., 48 (23-24) (2005) 4779-4796. [20] L.P.L.M. Rabou, Int. J. Hydrogen Energy, 20 No. 10 (1995) 845-848. [21] F. Bustamante, R. M. Enick, A.V. Cugini, R.P. Killmeyer, B.H. Howard, K.S. Rothenberger, V.M. Ciocco, B.D. Morreale, S. Chattopadhyay, S. Shi, J. Am. Inst. of Chem. Eng., 50(5) (2004) 1028-1041. [22] C. Liao, P.A. Erickson, Proc. of the ASME 2005 ASME Summer Heat Transfer Conference, Doc Num. HT 2005-72043, (2005). [23] H.C. Yoon, P.A. Erickson, Proc. of the 3 rd Int. Energy Conv. Eng. Conf., Doc Num. AIAA 2005-5567, (2005). [24] D. Betts, W. Lear, P.A. Erickson, Proc. of the 3 rd Int. Energy Conv. Eng. Conf., Doc Num. AIAA 2005-5603, (2005).

PAGE 128

BIOGRAPHICAL SKETCH Daniel Augusto Betts Carrington is the son of Dr. Claudio Daniel Betts and Carmen Alicia Carrington Betts. He was born in Panama on August 7, 1975. Daniel graduated with a B.S. in mechanical engineering from the George Washington University in 1997, and obtained his masters degree in mechanical engineering (thermal sciences concentration) from the University of Florida in 2000. Daniel is married to Erica Eva Carr-Betts, and they have a beautiful and intelligent daughter, Matilda Eva Carr-Betts, who was born on September 28, 2004. 116


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101119_AAAAEC INGEST_TIME 2010-11-19T23:15:59Z PACKAGE UFE0012167_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 69193 DFID F20101119_AACSPE ORIGIN DEPOSITOR PATH betts_d_Page_117.jp2 GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
f13ebea58355b8feba5eb228f14dacc8
SHA-1
f9313e27539e8e0edbc3e2b9e825b69885a74f3f
90934 F20101119_AACSOP betts_d_Page_098.jp2
43e1ddc5513775d2da7903b2fa671d5c
127440e35e873d556ede6ffbf33ab9c980ce2e9d
114307 F20101119_AACSPF betts_d_Page_120.jp2
1f8b3292a4024e6b61346a5aec52abac
c5b146739c179ec336a79d56565d906889a3c472
108379 F20101119_AACSOQ betts_d_Page_099.jp2
40378c63de937f12458ea152c047ed08
bc0647f32268f16d477cc90a3caf4739b4b9d5e7
108018 F20101119_AACSPG betts_d_Page_122.jp2
47c5de66a68d01f5153515df14fe6b2b
a30e9dc49bc8498cf3d98c95157e66c9784eda48
60632 F20101119_AACSOR betts_d_Page_100.jp2
d727c3af49827935d74ece6205672a14
d31cee44cda76f1524707b7bee745a0395976b3b
66638 F20101119_AACSPH betts_d_Page_123.jp2
c17d7723f2fe771b0f0a2af0e3dcc95d
0a23d2e953c9eeffce420e9c5220f33e5df26162
47094 F20101119_AACSOS betts_d_Page_101.jp2
d7053a67362479ace055f72ed612e81e
8481b06d05aafa70155ce99d3c523393ff888d6b
78255 F20101119_AACSPI betts_d_Page_124.jp2
9d31f08f1300bad82c1ae58600cf9380
561c1a3d0e9447352d6905165f8393bb08794858
657333 F20101119_AACSOT betts_d_Page_102.jp2
36a594e8d38aad181811c419213edad1
80abe3b39e444b5ea4656d12bb5184b95ec2a929
16283 F20101119_AACSPJ betts_d_Page_125.jp2
68137dd23e01036d6bd7d2c274ac6f24
4fcb6fc2ab5714f808e5a55a0b4994b8ab8d4ff1
120226 F20101119_AACSOU betts_d_Page_103.jp2
448b5c095f925409a8f212bc4d268fc1
ac0c3dd8e614bf387d7e279627e4236bd94ea423
68208 F20101119_AACSPK betts_d_Page_127.jp2
f92c4b82162e759025a2094b9b05b42d
badb590a30b0ecd5aa26c6725b4cf51b8070ebb7
100650 F20101119_AACSOV betts_d_Page_104.jp2
a765bda9e2539c92bfdd6faadf9a52e3
25463f7d3f0a72b0697b8238525025de9dc52588
112501 F20101119_AACSOW betts_d_Page_105.jp2
254721c35b4151cab338b23200ed8b85
26e7cd7270f0a63718903e4ddd5a022f4233f90a
36246 F20101119_AACSPL betts_d_Page_128.jp2
4fc0eec324d3550f16ab4524e0aee42a
2ee834a32cc85fff7d04ca80d688691fa7d0cb28
95683 F20101119_AACSOX betts_d_Page_108.jp2
b8c2e874688b49750fd79d9f59f3601e
b239edcc07abe0a637c7b053e720a8f479fa9b49
1053954 F20101119_AACSQA betts_d_Page_025.tif
18076452150faa0746e06b10c23abdb5
65257cc370c3f3f5630e78e02fce3d1fd018a8be
F20101119_AACSPM betts_d_Page_001.tif
d616980e2f268e27a1b98503e050601a
59fdf2a891ceb234ace6d56595c981c52af1d18f
105628 F20101119_AACSOY betts_d_Page_109.jp2
b81a381c5e8f37e7c075d05c1f9b77a5
3a2389b61954b065a74b59c1a8c187ddd4a55d0f
25271604 F20101119_AACSQB betts_d_Page_026.tif
64685102f99b6b2d4a629fe09f4780bf
d279edf9509978220e06ad261b4c91ad2db2e99d
F20101119_AACSPN betts_d_Page_003.tif
cc41c6f2f03866327c2eeba932a9e689
13df1b78df156cfffda70a7cc558ac9d79292a7f
37228 F20101119_AACSOZ betts_d_Page_110.jp2
3b6ab0062a3113ac785c8308c6cfba00
98e8fdab537875f005715ac598a7f47fef15be1a
8423998 F20101119_AACSQC betts_d_Page_027.tif
7cc86d4ea0dbcf0a3084dd6fb11f54a7
f22928c54eef83917d67706195d15bc9f369e529
F20101119_AACSPO betts_d_Page_005.tif
de1a59e16eca7a7f74f79d3ab8647eae
731e41f8f6426531b2269185a8751e9db51dad91
F20101119_AACSQD betts_d_Page_029.tif
0e9dc76bac0acef82e56862213ffe03b
fbdc03816833e83fae8fa9388bde3ccce1d77761
F20101119_AACSPP betts_d_Page_006.tif
b107bdf02f8439aa69d17f7bc01bd49b
bd6f013b3b3f141b79f3fed64af8ca2e48eb9806
F20101119_AACSQE betts_d_Page_030.tif
d7dd2aa4a3a72ae3f7a364c323e531db
2d56bdb35b757e7c08b216655f5364410605e828
F20101119_AACSPQ betts_d_Page_007.tif
bc469e311e010ffd9f484e1eb3b7b3bb
ca1d19cc3aa12362156c8d465f551ebe33f40a6d
F20101119_AACSQF betts_d_Page_031.tif
57a49d5c7d898a43fcf5d95eee3a3503
0aabb14b6d8bc1c6415c1e5a6586c35e7e3f20b9
F20101119_AACSPR betts_d_Page_008.tif
ce53c8e2b29ecfbd5567afca5e0707cc
f487f0542771699ca7fbf0435b794943e876419e
F20101119_AACSQG betts_d_Page_032.tif
06b13bb8f693bd5a02859bc8a9aa61c5
639642b09c630157283cccca1b7cf38f2328376d
F20101119_AACSPS betts_d_Page_013.tif
dc9d79481e4dab0477a68025d66880b3
9b8daa5386587a45ae2ac6a404380c0ab192b48b
F20101119_AACSQH betts_d_Page_033.tif
b60e2090849d395e64f1791e7515f3ad
bb647a3746968ea787b9dea6d505f6bab407db39
F20101119_AACSPT betts_d_Page_014.tif
997b560d8fdb8a2db835798f2c01bd7d
7e5f8d83f244a52203cd73e1df47197fcedb4b50
F20101119_AACSQI betts_d_Page_034.tif
150133a0f740d0105ea47be6d6894e6d
d50065fc974c2719487143a03c746bb6561cb600
F20101119_AACSPU betts_d_Page_015.tif
64a1eac4140919d13dc8dc2c729de014
8758f557dda30a5eea8127211577fd92eb7bfb73
F20101119_AACSQJ betts_d_Page_035.tif
0163befa9355b9986d0d280f6cd03dde
979513bcdbffe36dfd94ca161f856dd6ab7cae1e
F20101119_AACSPV betts_d_Page_016.tif
9e40ea7f273727c98099f64edb87abe5
3b141f706469a1a414531ffb010efd725d3d8ef1
F20101119_AACSQK betts_d_Page_039.tif
c3aef13a9a3aced999ae7a6bd0596c59
fec3944874f6f7bbf3ff105e92576495bbfa9d5e
F20101119_AACSPW betts_d_Page_018.tif
bac742089d7f35a1c7d80649ec83d2f1
6d1ced2f39687cce0662d333adeac26b88aef1b6
F20101119_AACSQL betts_d_Page_044.tif
4978b16945ec7860a1dd6422ce61341c
1a1ad9ea8fbfac969ee9312b026ccda41d8dc05f
F20101119_AACSPX betts_d_Page_020.tif
96445620c6a385c4f946ad42b4423724
8d448dc5b4b9430c2d775cab3b5268c190e97678
F20101119_AACSRA betts_d_Page_068.tif
7e3d9896254c170a3308a63153e26649
b368dcdb36229068c48232a8d7d9e850f5d2acbe
F20101119_AACSPY betts_d_Page_022.tif
15c1ace3b026f2079f6600bd6f659e49
671acc42fa5bd4c30b187b3e6f7f96a5c295f0fd
F20101119_AACSRB betts_d_Page_070.tif
be757874dca721aa657048a5a94d5b3e
4b9f4442f0d4e96e01cb399cecf233add7bf31f8
F20101119_AACSQM betts_d_Page_045.tif
1461ae596a0969a7a92cd3ae0d1ed490
7cdabfd8354ddd8071993222409d975a19782b0d
F20101119_AACSPZ betts_d_Page_024.tif
f706a94166a98957868bdec272fea049
18757eddfe4f154755f9e146b6f43d4b09b6bad7
F20101119_AACSRC betts_d_Page_071.tif
93401414cc21febe70c0c6987160ed7b
3969555e1ad1ef8af3b8f3deecbc4a16f8074ba7
F20101119_AACSQN betts_d_Page_046.tif
99f2f7080a9141e3431735409084a06a
86180cad83b6db06d6da23b199a7fe52f51df417
F20101119_AACSRD betts_d_Page_073.tif
3f2f7d17e66208a603cb63092064a4e8
944be160b979c913c1c496feb884278dd5851f64
F20101119_AACSQO betts_d_Page_048.tif
687e864cb114bb1e9e7d1361bbe3bd2b
c0ca15332006e939a298264a532f81c65fba3d6c
F20101119_AACSRE betts_d_Page_076.tif
144e5beada3c15620e22ec61a89d4a34
03bdb590dc3bda63eca11be5871ac6f9b3defcb1
F20101119_AACSQP betts_d_Page_049.tif
68e18d560c4cc967b470d6de0b35c8bb
fe622ac5300665f2c32c6687ea4bc416e7c96707
F20101119_AACSRF betts_d_Page_079.tif
d3aa8ea6b956b6fd6bfdf34adefb4ea3
6e75209f65b35597e0bd4d6a4ff2c1378f0c4568
F20101119_AACSQQ betts_d_Page_051.tif
f1090b52b93b2dd6e190c04e26a20ceb
a5ea61b69789fdec7ae64756d3b07535ae717f9e
F20101119_AACSRG betts_d_Page_080.tif
38c9ccb06a2fd1f6e5ef83ce85fa2071
cb9d71e287e5073bcd22d7a3542c4f6cef5a0e5c
F20101119_AACSQR betts_d_Page_053.tif
59a81ac5b75f4cbe7698462557545cf9
9d9110273ca7b671d05b1cdf8087eb879e5e31d9
F20101119_AACSRH betts_d_Page_085.tif
e83ddeecf0ec8f0dc3d5a83b1b5ad066
c07b3eeb2128572bc9713b08bbdc8907afa73260
F20101119_AACSQS betts_d_Page_054.tif
52c1cb822847ec9d7423e0f0683e0fe2
76fdf17ac215ac439dc51bc80ed6cd6ca881f2cf
F20101119_AACSRI betts_d_Page_086.tif
2527516b29952fc2ef26c8f1122fe5df
628a23b9af721805bb17407117cbbdc3e72e3a6b
F20101119_AACSQT betts_d_Page_056.tif
3b53e97d38aaceeed5422beb97d1697e
b68fa65f3216e31d26b2d43dce5e03bb171314f6
F20101119_AACSRJ betts_d_Page_088.tif
52f108dd4115ba9224891ade6be0b61c
8ff1039a248db84b16aafd52466324411ed1fde8
F20101119_AACSQU betts_d_Page_058.tif
27f697b1315b642f12d30a20de347acb
b855cfd239f9162e376e85df2320b43ef58dea4a
F20101119_AACSRK betts_d_Page_089.tif
2823e5553fbb1e6fb2c255b3c1a2f521
e4696f4710391e082f3aa3dffe83b4b45abdb638
F20101119_AACSQV betts_d_Page_059.tif
cc94aa00fd86c03628e4dc0c40ab1ce2
b086c99e72dc2eca8622910a3456f74ef613f55a
F20101119_AACSRL betts_d_Page_090.tif
d299f9569f3e91b642c6f980f3f70203
4baead1885998455cf8d13e3e351557eaae0be71
F20101119_AACSQW betts_d_Page_062.tif
01f97d251bd5e77208a67dcddfb4a5b6
59154c5f03f0d1afdd29271c6838de0b50c3d478
F20101119_AACSSA betts_d_Page_113.tif
836fe54236c19f724b34f1e0fa8a7729
7a70226e455083d02149d1f0bc62f793e015475e
F20101119_AACSRM betts_d_Page_091.tif
e2634d75d174420adb4ee7ecdd5b6de1
f06c834000c80e5634ded97f2be2efc51f8af96e
F20101119_AACSQX betts_d_Page_064.tif
bf4b0f079599a3f4be17308b1ccf088e
8f84611c6aa7a36c4ff024a373c1b2fc1e297984
F20101119_AACSSB betts_d_Page_114.tif
ae1f79a31c62dc28bcb6164f49bf3c41
b4e78c5cc8e97579e93be2c21b83a99908e72c3c
F20101119_AACSQY betts_d_Page_066.tif
70745f94e9814734c3ca45f8f5004ce2
9dfd8c51ad296e8165450ed41f141805ed6be74d
F20101119_AACSSC betts_d_Page_116.tif
d0f0fe1b145278ea284fc41e325c06f6
0d9c85356425f07e8aec06c6c85e1f36c922042d
F20101119_AACSRN betts_d_Page_093.tif
bcc65d790160407249927a84f91d9b75
971aea082f1d39a3bb14f22a89c13a2d1221fca1
F20101119_AACSQZ betts_d_Page_067.tif
ad77fec82742c70e4f222a7d7eb0e4a5
7d66549469e060fdb873b28aae9b6f40d6e83960
F20101119_AACSSD betts_d_Page_117.tif
733bffcb1a942f8f9b7ed7b371b916db
b75100b823e1addce901c46b3a12f616389c4604
F20101119_AACSRO betts_d_Page_094.tif
0e5234c3cde90987ddc0ea5cd029a37b
cc640ed8d6c6b37520d6c7af298fe4cc4f916155
F20101119_AACSSE betts_d_Page_118.tif
b0dcdfcdd0077eeaba0952a81aee085d
242e2ca78ef60ef7a1590be78f1887d18f44c5b3
F20101119_AACSRP betts_d_Page_095.tif
9d6abe2725ae6fcac086816315441ce1
b5728a981240ebfcc45f4dd7dbd4c86458d631f1
F20101119_AACSSF betts_d_Page_119.tif
a26fc789566f810a1def3e318db835b7
02a7cd733afb321eadef736962b20d4c75145fd8
F20101119_AACSRQ betts_d_Page_098.tif
63782a2d74ce0f69653edcd16ce80529
a5028ddc2061db63873721c0e5d42c6e65b2ba29
F20101119_AACSSG betts_d_Page_121.tif
7fff874f7c895636ee87e3e82507be11
5c4282084ce32be2afd6689fbcecda564ac38778
F20101119_AACSRR betts_d_Page_100.tif
ae73a8a1076ec870761901c750ef0b50
658f0e72ed98ddbf1911112a3fd053f6ace11fea
F20101119_AACSSH betts_d_Page_122.tif
2e1c3af628e87d491a61de1cc4c0c465
17425c870ca70265a476dd865e35f28dccabc947
F20101119_AACSRS betts_d_Page_101.tif
b115421cdbc734f2d855279f0317a9b4
639b19636f7ed7a64b58fcc0329307bbe8d38c93
F20101119_AACSSI betts_d_Page_125.tif
f596b2c73cba079e3399da802f53f031
e4071d793d272fedd7463c6e400de093c3706442
F20101119_AACSRT betts_d_Page_102.tif
caebeb8951fb402fbc8c32b21fe2a9bc
76c0f8001c512c58ab84f26bc4ccf826947600d9
F20101119_AACSSJ betts_d_Page_128.tif
b71726dfdf4a87f692f039ecec266a74
d5af2b12a97a65a10525236a9bb02610cd827f8d
F20101119_AACSRU betts_d_Page_105.tif
a7e0062720886cb7338294ada8db1c66
0d16f8e211fc966c670f192f709c5d3b7e1316e7
F20101119_AACSRV betts_d_Page_107.tif
bca43e54eaa81f532eed304091d8a74b
a01882985bf7d6fc12091d20ed492ee2552bc75c
8938 F20101119_AACSSK betts_d_Page_001.pro
ff5502218d221bcd6968d300fa764237
23573a2354708bc6b0f1a340a25e2f6c66002e1c
F20101119_AACSRW betts_d_Page_108.tif
b04ab892521862729a1b509733b97a13
3f2eba9649f4bfef33560698b693b35f5c68fc54
76009 F20101119_AACSSL betts_d_Page_005.pro
b32cd1982714bee1ee7ffcd20fe20b9b
e2842c2ac58848c64057180c0cb8df3ff2e479f3
F20101119_AACSRX betts_d_Page_110.tif
e01369f93a1a258c6db10564dea48626
bb1e2fd522e9d5ae2d1f0dd7d6af3da39a31cc6e
40905 F20101119_AACSTA betts_d_Page_023.pro
be6016ff941e5f94154e3b7253388bc8
f58065b788814bae5809da3c16bcfb1c66b12478
64342 F20101119_AACSSM betts_d_Page_006.pro
3573e3c30dde00e0f13523d981606dbf
1deaeaccf37fa848eceec7ead6515d6c6b18c02c
F20101119_AACSRY betts_d_Page_111.tif
d7530e6347ce515ed2152ca940b5d62d
f33230294b35840f2a3b85f14aa9340de6247151
52153 F20101119_AACSTB betts_d_Page_024.pro
64ca7f338c7df8ee08b4e11f3bdf550d
8a603b6c760e0bce9894db797f121d6e4afbb1ac
32539 F20101119_AACSSN betts_d_Page_007.pro
94add78e3848b8600bf2dbe78f6afa7a
a96fa442b1b98cb40485a6cf0411de83bef3808f
F20101119_AACSRZ betts_d_Page_112.tif
9db86f3fa41c05bdba43321048549085
dee39bc1a372631ab28fce3fc22c927e2a2a571c
15724 F20101119_AACSTC betts_d_Page_027.pro
9fdb7297175c849032a7db36fceb83ca
68772e7741797fd1ea3bc58f53df9c2e41ba8d30
30452 F20101119_AACSTD betts_d_Page_028.pro
de1c0e545a44a0369df55167aa758237
5994f967e4ed74ce1ef59d3129da6e4c22dd3678
57509 F20101119_AACSSO betts_d_Page_008.pro
45aeb6e2dd3994ba05c293142b5b0a36
224e8ed6b4dfbbb4ee7cf7157dfff923b932c791
50427 F20101119_AACSTE betts_d_Page_029.pro
f178bb6320cb165248a86f572008a288
73a4b5758e484f9a3d9da9ed3574e79145b9c1f1
74005 F20101119_AACSSP betts_d_Page_009.pro
855ce39e150eb375e8a85dff3a5dcd15
5d1c22f0c9ef3fde686779fc210a65606108c239
39386 F20101119_AACSTF betts_d_Page_030.pro
6241c5379ad1e8af885d38bff290e2f3
2bb178fedf64905efe66aedcafd3ff0cbfceb58d
74178 F20101119_AACSSQ betts_d_Page_010.pro
ad87fbd2b6be99e20528796f9b4dce55
afde8b66d24080f8156e658f1890042289df7b1d
47422 F20101119_AACSTG betts_d_Page_032.pro
ee6717abf3405ccc56662b9f8cf828bd
2eebccb07fd9e0368d4d2e4e1943c3e5633fd1c7
39498 F20101119_AACSSR betts_d_Page_011.pro
c89f7603e44351e7d9cea914b757dcc7
2f3627f2e6209319d65b0d052de8a640726dc8db
45554 F20101119_AACSTH betts_d_Page_033.pro
997144f0ae8b0544ce22f0d7a9a8ceee
ab44a9c8bb8808ef0488cdd69d036b33198b6d6f
44203 F20101119_AACSSS betts_d_Page_012.pro
b1999cf2fd5bf0287dc14effe8330ea4
1effb7147b1cf6f08b5d676f2660769185b67c2d
50726 F20101119_AACSTI betts_d_Page_037.pro
df8d79fe1cbf4e70ddd1c779c4a34ed7
1ea6bb9778184c29cab5d72037f870cd3068d350
44264 F20101119_AACSST betts_d_Page_013.pro
8f47b1d132e117585d03657656a7a2bf
728f8784aaf6ebe187402b871ef411a86e0d0e43
48038 F20101119_AACSTJ betts_d_Page_039.pro
3c179b81ba085529b14f4ccc7de23d7d
3680c3abeafeadb074c8d9120522902970721c5b
44186 F20101119_AACSSU betts_d_Page_014.pro
068dd51bd1e261cc553a44af483bb95c
3feff4b2a865019056e0e1164f316fbba1dd34fc
4650 F20101119_AACSTK betts_d_Page_040.pro
b89b52692b566de3cabee60ab2e25ef1
abb8688af2d8d16ab9806a417e5f3d12254ebd4c
22815 F20101119_AACSSV betts_d_Page_015.pro
be9699b8d70a9e59c626a61cae769d18
55246d7094e6ac04f020faaafa02a00cfb29d38e
22727 F20101119_AACSTL betts_d_Page_042.pro
63562afbc63f08a3568dfb93334d5923
bfd45cd386cf55de87e614c9e4b609b6ee4351c1
30278 F20101119_AACSSW betts_d_Page_017.pro
4a6e87ca5fecc2bea282fbc01f0d9645
de3944ddc9a98a93155b1615ed3616b587a06bd0
46780 F20101119_AACSUA betts_d_Page_058.pro
3cddd8b8cad9c80708ebcfa89a884690
c385a7c03fd856b501e24a4a758b762e5c46bd4c
42895 F20101119_AACSTM betts_d_Page_043.pro
d6aa5d60bc6759ced59b831c354f3b14
fa6a9a54e849037abd4709ecd700efe87d3e60f1
49330 F20101119_AACSSX betts_d_Page_019.pro
2fd1a8021c48985798ba61113633064d
24431f2c465a49f67147665838506f8296b200a6
50820 F20101119_AACSUB betts_d_Page_059.pro
e4d5c9dfa3e11661c43bde535d6df8ba
7d6219349ccb616094b261ca7bae149e7435730b
51439 F20101119_AACSTN betts_d_Page_044.pro
98de16aa7aab27590b27fe5f5516cddb
0c849e5f2f21d17717ad936a6694025e17a3cb03
43969 F20101119_AACSSY betts_d_Page_020.pro
ac5095140ebfd0c9ba79b502a91b66c5
945834b31d2f90324685a0fa891fcb5193e0e802
48767 F20101119_AACSUC betts_d_Page_060.pro
ac0a82526b635c29417d8e7b4d254b8e
c2d4a3348e9d019e9ba86d5db4d7d7b575c086cb
31702 F20101119_AACSTO betts_d_Page_045.pro
02f7ab15d06e4eda052ce98d2ec6d0d7
0d92b1abfa70ae0a27c0b46358cc83b9890ae831
44438 F20101119_AACSSZ betts_d_Page_021.pro
828493a4c9f6256a82bc117630ad13f5
51e5d1c1f5605c1d83761646de7bab8e9e41f9a6
45270 F20101119_AACSUD betts_d_Page_061.pro
b888807b191f775513dfe7dad563c388
3025f5fad05069c4caf509884b18c15876ba00fc
22956 F20101119_AACSUE betts_d_Page_063.pro
4afbb5041714679b96b52592af034fe4
5184d2a8a76c5e04647ddce312c9452877d6c695
28938 F20101119_AACSTP betts_d_Page_046.pro
93a13a4e40894a5ac5968fca18595fbf
59d184d9aaac96c165ecff4e28557f1bf4507736
51531 F20101119_AACSUF betts_d_Page_064.pro
690dd4cb23003923905899859669a293
ecd21db3861dd133c8e7475aaa7480b392350b4f
30997 F20101119_AACSTQ betts_d_Page_047.pro
1513515ab99182c4eb1481ee98d87896
90bf4cf119336d41f71467424e9ca364841fdea3
31827 F20101119_AACSUG betts_d_Page_065.pro
b102a3029c71e4e29c2cef1a83d7171a
40a3f37a5483668c76bc96eb8d8198d7a26aa095
42526 F20101119_AACSTR betts_d_Page_048.pro
7343ac80226a366a7af2651306ae3c0c
3438bace5689b1eb8ddd0ebdb2979444086d5898
22044 F20101119_AACTAA betts_d_Page_058.QC.jpg
b7c33e2a27f734617dc28a9dd49c16b7
d1d0b8fc8da85d861c0a173095beec7961b92115
13455 F20101119_AACSUH betts_d_Page_066.pro
1e27d2e0bf1359d546f042b3c0bc1a06
f76dcafd3959759a0488d285efa7d492f68ec477
20047 F20101119_AACSTS betts_d_Page_049.pro
641a9cbfa2cc5b7be07ff33f18de71f6
b3e11c3439147a83905c15e707e374a5f20a5e80
5790 F20101119_AACTAB betts_d_Page_023thm.jpg
fd594e935d88e5b9b59007e70abcd692
d53e3fdf88405138fa6121a7b68d07dd02723e43
26873 F20101119_AACSUI betts_d_Page_067.pro
d0a29e6d6d5f6bea1b8638d465622a4f
cb0e0b13da96fc5194c4706936335d116f27a168
38614 F20101119_AACSTT betts_d_Page_050.pro
d38975a3514423ba91ba42e40ee39d39
4840fa80c56da9f087dda87a999788001f8ddd01
22598 F20101119_AACTAC betts_d_Page_032.QC.jpg
425c0af171b0f7e91bfc3f6f9a06f14e
e0f066cab08f27764fd7e55ca606f1589dadbc08
45563 F20101119_AACSUJ betts_d_Page_069.pro
6e521f2995ec0eaf7886aa2cb7246fcf
98aa742c93c5242e75bc0f2d34c914195cc31452
29107 F20101119_AACSTU betts_d_Page_051.pro
78014a9ac153069d3a68dd15746372a9
3329325306312f47f3cccdfece1d5e8ccb22ec9d
22436 F20101119_AACTAD betts_d_Page_008.QC.jpg
2f149f648e5dfe59b17217fb0a820ec4
513c4a447d539b9f42ff26701e2ed93fe10e1148
29760 F20101119_AACSUK betts_d_Page_070.pro
37b980f0327003e425f2467fd9738651
b35f4882cce5b1c9b73c26220e5b919182ec0312
19136 F20101119_AACSTV betts_d_Page_052.pro
66a9b19683cdbff38b6f6051f868910a
5d999d89f182e731bbf2ebe42ebdfe8e6f15ac94
15037 F20101119_AACTAE betts_d_Page_107.QC.jpg
375a6f45528ef8472aa5635cd38a3f6a
fcb99785ad6c7ff5594da285ad068628508b4b27
39328 F20101119_AACSUL betts_d_Page_071.pro
03cab05708b0762f109215869151d16f
277ff0ca1b9fc1a8caa4e23fce6bb44bbd7416a8
45081 F20101119_AACSTW betts_d_Page_053.pro
58dc304a358eb87690c24cbc3e24165f
5c101d661c6c7adf3dabd62e7c4b83cef13c730c
11990 F20101119_AACTAF betts_d_Page_066.QC.jpg
5e40bdae6ac8cf78301ce31d4426eeef
dc69db67dd13e30df8883aef2a00c795e825586c
34292 F20101119_AACSUM betts_d_Page_073.pro
31b25240ed09565033dbef6ba8228175
41facfcf2af59cd81360d0c876c10ec7c3012669
19717 F20101119_AACSTX betts_d_Page_054.pro
3afe3f5742309569c54645593a4ed02d
8ae36ec63539509e880a0c911ed0045f7b0c8cbf
19312 F20101119_AACTAG betts_d_Page_083.QC.jpg
6c5ac27dfffac63968dc4b9c8fb5abfb
cf9e71d9b1e56c2d1399a2db1347267b4a2e6070
34262 F20101119_AACSVA betts_d_Page_092.pro
fe995aa67ae9ae62432a19c7e4db5f32
105290255344d9cc0a2db0ffe441b8e47694cf33
40797 F20101119_AACSUN betts_d_Page_074.pro
aab25291d2065bd44a2f34407e4905ef
ad19530a0683c1050f1ec363646032e761909b40
31035 F20101119_AACSTY betts_d_Page_055.pro
27db33365a3fb0a534355622fb78c1f5
e344971e5ba295c22e858f10b58f633dbdd43efa
5253 F20101119_AACTAH betts_d_Page_063thm.jpg
cdad0576e985a27b1c578579b8f2745e
6ea85d82ca77add3bca4749e4f3ea1d0738c081b
18785 F20101119_AACSVB betts_d_Page_093.pro
901ada202c9e8a7469f6244934e84074
ee7e2a376183e09b44fa36f7bdd55d14f155e3eb
51954 F20101119_AACSUO betts_d_Page_075.pro
ec1ca2f821203eff7238a8f84c96d854
bbc2d8971e16ed04934246434807955b5e337e31
46798 F20101119_AACSTZ betts_d_Page_057.pro
a725c8b294e4cbbb17554a53b859628f
9ebe46bf86b9b5d5fdeb0cd057d5cce6f6c4fd88
4478 F20101119_AACTAI betts_d_Page_006thm.jpg
ac60b138505f3a78d3c92892c4bc4e3c
65c89bc946ec25d9067c2745f0c407141d602b39
37925 F20101119_AACSVC betts_d_Page_095.pro
0fbecae2a37f0380cb190486655e88ca
3ff2452022fd240f6d56c8a0f6baf74bf15b02b7
7982 F20101119_AACSUP betts_d_Page_076.pro
325847ab800a60076a3027e2882dac9d
22b4198fcedabe3ecb659c1951a84b6503041445
15514 F20101119_AACTAJ betts_d_Page_087.QC.jpg
27a26470510c5f4891824e6e662446f0
01d4ad084f9cdb7df3add28ffd6c4d8584ba14fd
31337 F20101119_AACSVD betts_d_Page_098.pro
d0f6f0a4cb6d66f293ea26be91fd2a0c
2495efd0e4e51557e89791254a224d51d634f6d1
5134 F20101119_AACTAK betts_d_Page_065thm.jpg
3e2fad257b24ab88d65e919ccbb47dc2
4ce95934490f74204131ab19e17ec678b9cf9781
28639 F20101119_AACSVE betts_d_Page_100.pro
8e92db95155069eb821a86996e44de00
5b63a7b64ad5db734c4c021ecbdd2e53450e5e30
34620 F20101119_AACSUQ betts_d_Page_078.pro
24e5c9264e7ba98ff9e7a65a56015ae0
8fceaaff774b893768bfd3b00e6c23c3c18534f7
3992 F20101119_AACTAL betts_d_Page_102thm.jpg
c11336f29bcd761b61d8d8127933645f
af4e7e89b7302b80bb502901271a381f42636d88
20709 F20101119_AACSVF betts_d_Page_101.pro
113f983ff4ffca2771337da9ea71fd32
510bc6fe3e8a3beb02e7b8626498c498843b1059
32086 F20101119_AACSUR betts_d_Page_079.pro
29aa2131a26f594018d07180adeaed26
ce50b1ce466d06c4ac32200f1035ef17770883de
14255 F20101119_AACTBA betts_d_Page_070.QC.jpg
321198ddfafc400f69932023a2b13dc0
57ef6c284b42e556493bc8818e342b622fb7e4a7
4383 F20101119_AACTAM betts_d_Page_051thm.jpg
de052cc5515fd68946ca68520b8b9520
e497ab6546dfc1e9b13d7b7aef1d965adda37497
19156 F20101119_AACSVG betts_d_Page_102.pro
4bd10478b0ceca85fea8fd27f5983cf5
6b44a2d06b130a87e045fe9979adec20cf19ff87
38650 F20101119_AACSUS betts_d_Page_080.pro
21b569f327f05b3e63314dcba29ecf06
6edb92025c7a0b77e269339d56a326100d880112
5630 F20101119_AACTBB betts_d_Page_061thm.jpg
8e8ce9c83b4f9042a93ea452a338b060
2721f26023667b85be580084afaa078c9be55fb2
4898 F20101119_AACTAN betts_d_Page_086thm.jpg
6cd91e380d43e2bfd3b828559421dc23
e11996462578547377ed7c44ea39653cc694446a
34199 F20101119_AACSVH betts_d_Page_105.pro
4b0003627a65c2b00c57a3eb21934e1a
4b671967cc26ead387580128802164a5117fcdd0
22325 F20101119_AACTBC betts_d_Page_036.QC.jpg
61de6801f637893186901b75162a4611
5424ce1ee30f0cb682afa51a9b7b7277c5541212
41075 F20101119_AACSUT betts_d_Page_082.pro
d2b7a4f0252f1956bb2ea4319c64374a
9b462441b575db25c4959947e0dfbd7e9e62a485
14409 F20101119_AACTAO betts_d_Page_046.QC.jpg
43270fc0295bb8a06001a71685e0029a
cd4ca1fe5f3f7b3009eefd4bfa06f857006b1a80
37352 F20101119_AACSVI betts_d_Page_106.pro
7b30fcc8ddac276cc190437c312e9d42
7d426fae5665e29731096abdfa872b5d7fd0fe10
3050 F20101119_AACTBD betts_d_Page_096thm.jpg
743bcb4f67d174d5f97340800817212c
54614b58072f5b956bb36483b5fd121ee3e31d6e
37240 F20101119_AACSUU betts_d_Page_083.pro
cc073b924ebaece441dc1f451f56bbfc
12a4fb27c9dcf1a8ba1032b992e565ca4f8b0c79
17358 F20101119_AACTAP betts_d_Page_105.QC.jpg
bf9ce997118c4fd6dbca637a9dc0e720
e40d165333fc0f7bd53ee92a810b0a732ebda0c2
37823 F20101119_AACSVJ betts_d_Page_109.pro
2045078154bdeff9ee97f4cc62f2ee83
35559c2e4bc7e6ff4b2f25dae16308422845865c
17823 F20101119_AACTBE betts_d_Page_079.QC.jpg
6885b94c362e9dfe2c5bc4dbf6d27ed6
8ff0597f3f17513d2afa022b7080ccb6991c5c18
18767 F20101119_AACSUV betts_d_Page_085.pro
e80a5648916b420acbcdbdaa7e0ddacc
97b377ec41ec440471ccfe791e857226f3f6c871
6324 F20101119_AACTAQ betts_d_Page_039thm.jpg
b33e7eeed2369b1b7af5a3b0052b40c4
5e541045d456d81cbbcf32b35180f2b7a7603803
10273 F20101119_AACSVK betts_d_Page_110.pro
952d91a6db667b08893b9d53fd29020f
7caaf9276402327d2898e5bfad8a21934506be0e
5898 F20101119_AACTBF betts_d_Page_082thm.jpg
31a2f68402d5906a1797e7a4cce17151
8c6c2615fd9135133c01244588d15f35745aaed2
25951 F20101119_AACSUW betts_d_Page_086.pro
7a1cff7dd358538fed2fee477b1a0ca8
bf7547b5eaaa9c17ab89718247155891778f95f2
5970 F20101119_AACTAR betts_d_Page_126thm.jpg
b645bed76ea9d8efe45f22fe9fa7597b
f19475a4a423a10e0b6d3dbd540c0c218a6ae9af
22744 F20101119_AACSVL betts_d_Page_111.pro
a488d4be7cc6cbed1310e53ff7d5f15f
69c4932c7c744bf65e5c15b6433b6be1a028c064
5716 F20101119_AACTBG betts_d_Page_053thm.jpg
4131ef8bc7ddec206cd0b39d6bf00c7d
12487b733975e123bdc4c97c052441e5977bdca4
31153 F20101119_AACSUX betts_d_Page_087.pro
9ffc7b99168aee3d9a2f835208f5dd34
7526f582db7f7ef7822ca82221e8d8853f6a9a60
821 F20101119_AACSWA betts_d_Page_003.txt
3ea081c2f6c3de05be65e7e6fb9be472
e36b34b40fb6a5d08890e9f57b8c6da7081bc2ba
6465 F20101119_AACTAS betts_d_Page_064thm.jpg
af5e511e90adc91fa78d285fd929c907
62ba783de481ea21d060a3b5bde903c37e82555b
28907 F20101119_AACSVM betts_d_Page_113.pro
594bc068d35823418d4d06b462540a52
7a877c3034e6eaf1fc019a96535e2659d71390e2
4320 F20101119_AACTBH betts_d_Page_098thm.jpg
81aea06db2a4a96c8c22658c62d37132
6f509b227f05a7501470d273748efed88bd24053
30113 F20101119_AACSUY betts_d_Page_088.pro
8176ff31787ca9d105a7037ca4107e89
a38e363b8e308c0ed6542f93db7f326ffff108c6
1433 F20101119_AACSWB betts_d_Page_004.txt
25cf5b202423a0efb15da89555381147
61bfd2b509bde5c57e137916329880e8aef495bd
5619 F20101119_AACTAT betts_d_Page_112thm.jpg
a22f7e4482dbd2d53d5123032e0673d9
9915eae71a86421b125e56f91140cb8297087f0a
32967 F20101119_AACSVN betts_d_Page_114.pro
a8723de25c5932f3e723fc260536778a
c62c41fa8fd4f7240b1bbe8729694e425d4af77a
5112 F20101119_AACTBI betts_d_Page_045thm.jpg
b439a47173c7400e866c678321ce0dcc
183cdd4ff5705497e92407eaba95c579b95d537b
41948 F20101119_AACSUZ betts_d_Page_091.pro
f27cf127d9fb1ea87b8fb0111bf41e5f
1ae5939c82ccafb39c9e7379ccb3fed5a03c42ac
3173 F20101119_AACSWC betts_d_Page_005.txt
15f75a09f1413517a303102faa9c0a81
94cf854a53b816df772804c7b3bd742b1563434e
2251 F20101119_AACTAU betts_d_Page_110thm.jpg
1a646deb758e28e8643b1130676fd446
4b820d5916b348198a5a8966cccc0bcbc3902a80
7148 F20101119_AACSVO betts_d_Page_115.pro
70651fde84d6c3b37f7caad69a67ae65
d69456308810c01b78847c4caca2e9beda04740a
16472 F20101119_AACTBJ betts_d_Page_085.QC.jpg
45c2ae9ea3d1624a5d64090899ca7a03
765f0b30581254ecb1c717a1b7b88f8982ce916c
2603 F20101119_AACSWD betts_d_Page_006.txt
f151690bb613c157bf8f62a18a453edf
89e4f20572afacfb200c31ea2b928d8259fae74c
14276 F20101119_AACTAV betts_d_Page_062.QC.jpg
8cc4584534f044bbf9ad002514bfc8b4
04a0b01168946a1c72678489e77e86acca0445a4
32927 F20101119_AACSVP betts_d_Page_116.pro
329e6e8dca3bc60d95048121580761ed
02d02e5575beb7e385dfef8e5dcef0efdc8fd08c
3993 F20101119_AACTBK betts_d_Page_066thm.jpg
81a1fbf0e57579db38b00b17a578fdbb
d96b2b81c60ac9c41e1ba8143019b4f2ed7202e2
1386 F20101119_AACSWE betts_d_Page_007.txt
2cef067d21369609b52725ca5dc48107
6ea68adbe01417905061c7da850c88621cbf1e79
7665 F20101119_AACTAW betts_d_Page_001.QC.jpg
a012cfef8a33d4d386e5e53a7d9afd7e
1297932280e6e6f7d9b293e55cc09e2728711ce3
41083 F20101119_AACSVQ betts_d_Page_118.pro
471a6aaf35a481765bcdcc89cd463291
18495ecd0cadd7fcabfa2de687438a621ca94872
10716 F20101119_AACTBL betts_d_Page_094.QC.jpg
2c4011aba7d4d9cc60ed962a0688267a
2a0285765ba86fb348e7162f8e1fbdbd67a6610f
2302 F20101119_AACSWF betts_d_Page_008.txt
d77dd26be4f44c1325f3aa4efae62a13
1c18bfc1c9fd86b0a64b55efda534ec08d7372f4
22495 F20101119_AACTAX betts_d_Page_060.QC.jpg
0fa71a61689d17d7c47b507a42b6b01d
c59e62c301dd0e17215aa186b694dddd6d5fb696
12677 F20101119_AACTCA betts_d_Page_007.QC.jpg
0ee864c89c891e8cc483ee960689b4cd
8cd9eab9c8943c486aa447769e869753012195d2
6828 F20101119_AACTBM betts_d_Page_110.QC.jpg
ed2e1c0db66eab2b9c4b1c8c899691df
a348052c488c34b6c3a7a298a110572955e98189
2949 F20101119_AACSWG betts_d_Page_009.txt
cccd12ecdbb43e50caddffb9c85685e0
e3f754e75f2e087572e5a19305dcad5efafd9588
23507 F20101119_AACTAY betts_d_Page_029.QC.jpg
39730e27fd8f7e439af8c20923141b5f
66959bcd8d6affe40b5de45806643ce7a8c44d1e
19791 F20101119_AACSVR betts_d_Page_119.pro
521af74d7a5c8a9d609e36a731eb0d44
20bb20b54df719d0cf07e66f9eeacb891e134d16
26782 F20101119_AACTCB betts_d_Page_009.QC.jpg
bfde702fe774991ff8d27c9de71cee9d
cc374d8bbf2663c5d3b5a32b41217cc25cbffc9e
5775 F20101119_AACTBN betts_d_Page_083thm.jpg
52047ca29470e759887db71ade016232
5219f2749b95f0c4eb0d992948312b5c9391a4c8
2951 F20101119_AACSWH betts_d_Page_010.txt
bbd881abbe717468731d96f77fda8ca3
c11c7313c2d1ba1aa78eb66ee40d5c13eda1be60
5796 F20101119_AACTAZ betts_d_Page_030thm.jpg
4678c9ad8e1fea2fe0f55a39e999cde6
731967127de128bceed8aed63f7fdf11b2816b30
53161 F20101119_AACSVS betts_d_Page_120.pro
954fe2d2f82ddf99301475fdedd6866c
487af8602de137a65332d7d533fc83a358dd8ea8
6887 F20101119_AACTCC betts_d_Page_009thm.jpg
40fdbd048bb228d1312c7248c94be093
287a43eaba93a895b8f3818d6a90f9bd80708cdd
21029 F20101119_AACTBO betts_d_Page_016.QC.jpg
1901fb1e22210ce56c9cfb181477c968
e1415d4f8c68299a675463176033bb2a2e82210f
1726 F20101119_AACSWI betts_d_Page_011.txt
b5c9231d29f58911e94a7609160b3bce
f3adbef2674831bb6b7a9a1aba12e50fbc00a9ca
52164 F20101119_AACSVT betts_d_Page_121.pro
e752dc4cd51b85068d64d94b8c01b22c
a820dfb1e0d34ace22ff0b6c14a04d64b5076a68
19477 F20101119_AACTCD betts_d_Page_011.QC.jpg
06d69ac47b37a8401b7990cb5d3bb8ec
6bc4418058672acebeb00e18540d17a5b59f5a26
3017 F20101119_AACTBP betts_d_Page_089thm.jpg
7e491e7522422fda35be14d951910280
8236813cb9d450d3b11b4a423f752af66287eeb5
1930 F20101119_AACSWJ betts_d_Page_014.txt
f682a145d4e667b9047b0a89a1ecc46a
57106689f978ff79ca4d6b396d6978e5071395a6
49202 F20101119_AACSVU betts_d_Page_122.pro
1f114c3284ff1748ec320665392af5c0
71b6eb12d8c54ac02c9ca885e6bc7c6acf38e276
6377 F20101119_AACTCE betts_d_Page_014thm.jpg
c4dc3ee6794243406fb2196926c9edf2
d1ecdd9c913aa287d2a1e9ff890c1024cb9a46c4
24473 F20101119_AACTBQ betts_d_Page_075.QC.jpg
ecedafd42d10ab9903d7ff3a3a0bd87e
b9079e69bb947d12270de398e9f23007348d3455
1873 F20101119_AACSWK betts_d_Page_016.txt
884a1e89395f1f84100a53b690278cc0
25f60d0e524088fd0e1181ce1c6a9df59dd91fae
19478 F20101119_AACSVV betts_d_Page_124.pro
63972b4da0ab677bc8fee76227c1ff47
3c0eb36a374385bf6d0ea12dacec2ba0df0273a0
14536 F20101119_AACTCF betts_d_Page_017.QC.jpg
f53ebbcf4beae7da6509d3bb3d630ca0
616a09370ef870f6390add7a5fc43b5d60190679
17757 F20101119_AACTBR betts_d_Page_080.QC.jpg
0e169ff95489780c276f8938270969cb
7f542224a689126383a439e233e22a3491b75114
1237 F20101119_AACSWL betts_d_Page_017.txt
105b4cb0aeb39c1ed8ddd25b4a156d3a
9497a025eed618b706109ced44c8f44f0539c295
42714 F20101119_AACSVW betts_d_Page_126.pro
713f342f514a44ffefabc7d4b0c16676
c4439d30cfab9b40e2b3385e0f9a5c6ff67c8287
21469 F20101119_AACTCG betts_d_Page_020.QC.jpg
0f7777b779e358edd2ea813ee7f60f9d
cf956176c88496b11fe6290d03119f10b4e485b3
19322 F20101119_AACTBS betts_d_Page_030.QC.jpg
bc0e51581d0f550ad2cf4bd50769acbc
156339250957fd52330b565b2f5fdaab58521e8d
1628 F20101119_AACSXA betts_d_Page_035.txt
739300bbc2baffa5b1817e846d7512cf
cc9fb6df08d87a3333d4550504b5593da230524c
654 F20101119_AACSWM betts_d_Page_018.txt
da8b826c946ce52003e9f680900456e2
a301752625f24598cd1fdd74da16ca4a4bed6f22
31654 F20101119_AACSVX betts_d_Page_127.pro
4a4a70ef2c797d6c91783866f24bd747
9eec27e9889ac3b1c3ffb658a25ddfd844a0e60d
21492 F20101119_AACTCH betts_d_Page_021.QC.jpg
d4a431fe418dedfb22adaa5e80d695be
de20873dfd8ce0c8887c7a0b98071d41952f425f
5041 F20101119_AACTBT betts_d_Page_080thm.jpg
aecb814d21807731b228b39fab2cc958
9c754326cbf5d7ef88c439c9c9ea0bcf72fe3776
2002 F20101119_AACSXB betts_d_Page_037.txt
4c29575cb914a937bde0680ff90bde72
fca4f88b17909980b3cd4b5588403af68ea70e4d
2039 F20101119_AACSWN betts_d_Page_019.txt
4e4de5f6073a9214a0ce8fdd5615da22
5c8ad4a11de71fa60cae9d98cf732ad808877d76
14820 F20101119_AACSVY betts_d_Page_128.pro
e51e3fde88b4d17b52a51d9928e45867
44b815cb2f0e1f8ab6551ae28e6a7520252a4818
21502 F20101119_AACTCI betts_d_Page_022.QC.jpg
d57175626eef1d9f1e7b94581c1438e5
97af2f3bb3045c5b14a3cd6c08f914ac2d0a32de
14613 F20101119_AACTBU betts_d_Page_073.QC.jpg
56af7a8c72ee6dc9b4adc47b4844f20f
713fe637487200d44d16328170c85139d11a3061
1880 F20101119_AACSXC betts_d_Page_038.txt
81967afda0ce9bd5275467296d372ce8
c021f74206a778a7687773575ff532b25ae210a8
1807 F20101119_AACSWO betts_d_Page_020.txt
d832db1bc1c53b9ee5eb067b46f0742d
266c9d66288de38468842142dc0455fae5591af3
476 F20101119_AACSVZ betts_d_Page_001.txt
65232c980a04b751b997a6cb3a636ba4
d589f73c0874a0a6fde4d35856ee6491bc9d50b1
6078 F20101119_AACTCJ betts_d_Page_022thm.jpg
53c65b70c03cd70380ab9b9f04823cf8
71c7aaf04e6093c5a2b6e835eb8fe58e21f96e7d
22141 F20101119_AACTBV betts_d_Page_126.QC.jpg
a0d07810bb996f69302be78dacd35848
a39b7a8f6f3d6867f02411f6c636b30bb7baea72
1897 F20101119_AACSXD betts_d_Page_039.txt
5dfac9e3f7bd1e740fc3fc25af34730b
ce46f3284172d20111c860ac51866784f759414a
1889 F20101119_AACSWP betts_d_Page_021.txt
0c9bf3b390f34d59e512b95fc74b4387
f49e1acb6864cb5fe0138b957f4b056e6929a822
18811 F20101119_AACTCK betts_d_Page_023.QC.jpg
bd436adf68129a1e6d41791309039ea9
e578db8556022808f65ab1489afab8a995986c2d
190963 F20101119_AACTBW UFE0012167_00001.xml FULL
024d7109f88ddc1ec8537cc779882d10
75a6fda07d98c038deb2839060ae0b53c5a91013
226 F20101119_AACSXE betts_d_Page_040.txt
30c71ab448ec407fd73d2344e50f84c1
01f0ac010e133f1a90ee751f810e727fb09f8432
1718 F20101119_AACSWQ betts_d_Page_023.txt
d49f5e17d02bbbce962d9ec5aef77eb5
7697c77eb518edd4153023f88983bc2f7cdf79de
24127 F20101119_AACTCL betts_d_Page_024.QC.jpg
17406418cbd661eb01dbd9eb9b88a9de
9b0a1051455f3caf2cf9100eb60ea6656149bb13
5097 F20101119_AACTBX betts_d_Page_004thm.jpg
40266ae2eb5ad01ad7e78f76dca84d6e
ea7c4c463ad8083407bc57c1274d034d057acd43
1783 F20101119_AACSXF betts_d_Page_041.txt
575b387665c7840d454ab4e11356b38a
aa4018be4fafecd27e0fb8355e72b32607f44a99
2070 F20101119_AACSWR betts_d_Page_024.txt
d0ef65d4959b01603457a7d653104d2d
9b4b12411c2d583d5d8db87ed001a772a6399dc4
F20101119_AACTDA betts_d_Page_040thm.jpg
e0417a002671034188196e8e7b74e0cb
000620dfb7b48e949a063a1194ca4c73bd0c6a24
6558 F20101119_AACTCM betts_d_Page_024thm.jpg
01d2c2232403bac1ad5811f085f4cba8
3b1fa5d388829c341996860215e755a768db5dda
16945 F20101119_AACTBY betts_d_Page_005.QC.jpg
ffa0b127226042646ddfcce27809a28b
b98d26b17b42fa8f710218400effd2dc57f5f699
1078 F20101119_AACSXG betts_d_Page_042.txt
ee82b2363acd4b17c443ddf1f6de8ebd
b36225dff0347a29bbab4ef14490ff0ade51e110
4525 F20101119_AACTDB betts_d_Page_042thm.jpg
2f6e9acdfcdeab766f0a393a0ee1a96d
6bf270c100d09830027af42d6382991d17394821
15646 F20101119_AACTCN betts_d_Page_027.QC.jpg
fbb5446600ed7a1687cad21756205df6
4aecc9a15c70747088edb6862ed64ef2e32cfc51
4796 F20101119_AACTBZ betts_d_Page_005thm.jpg
3b9ca62bdd32eb156b5b4edb1124d4d8
217e86e971d4515ac4d4614a006bbb5488b82c60
1978 F20101119_AACSWS betts_d_Page_025.txt
b9efc4b9d9eecd4aeb53be598dcfd32b
a1deabba05d38f1d738e731ace8a9aa93925e2e9
1387 F20101119_AACSXH betts_d_Page_046.txt
f8983ccdf2bd45af1d532a67a74b4e75
c612870fdb5970b687487377dbac3b0179fe4307
23662 F20101119_AACTDC betts_d_Page_044.QC.jpg
688e1b6115993f6005585a45da73aaa1
83ae56837c6b72d5321604877e76921f92f700b4
4604 F20101119_AACTCO betts_d_Page_027thm.jpg
728ed3c4b5341a1b6a8fe4916beda257
cbb0530dfd6e4236a8a7619492160f370d8b6fcb
754 F20101119_AACSWT betts_d_Page_026.txt
27d222bdcd35692653e95ed82bb34b5e
a69c6a72a24f45759e482e9085c0b77e08dad432
47608 F20101119_AACSAA betts_d_Page_117.jpg
57faa3f897582e93abf9dcd0aa8f71a0
b16cb095dbe51de59dd44dbc7bbb10587da0125b
1441 F20101119_AACSXI betts_d_Page_047.txt
052f302e07ca8178efd5115fc2676f66
de513720d5ebda5f66094991870ed17562cadf8a
4447 F20101119_AACTDD betts_d_Page_046thm.jpg
d19ebb024d289ddf733ecdd5853cf6e3
eaaf2d560967693b59fe59a3e5868df7b5f0c9d1
6408 F20101119_AACTCP betts_d_Page_029thm.jpg
3805e8f28b5485a56fbcfdac8627dc59
501bbc1298a3c3186f5e28dacd9633375917b074
777 F20101119_AACSWU betts_d_Page_027.txt
11a0a8644c91013b01b1c8c8ee0afa18
fab52f32cb7d4ad15a5098c70753a127875c0094
610620 F20101119_AACSAB betts_d_Page_093.jp2
492b32b37434f57d815ee75aeb83d740
3056f058f8fdd9a79887a8850e512f77eb5a512b
1816 F20101119_AACSXJ betts_d_Page_050.txt
37a08a4c40e889e0bf75ed5430cd695d
6a240145b4b0bd0f388b4e832d4d0f585734a5ff
15250 F20101119_AACTDE betts_d_Page_047.QC.jpg
94394a90da499b802d5d51582caf9489
7565a158234b5172b3acc995bbc9f868bb845fcd
21533 F20101119_AACTCQ betts_d_Page_033.QC.jpg
1c6e49d64cc8b973da791efc45123005
c67b9243b1cf27d93283989f9dc6def3fd612c41
1557 F20101119_AACSWV betts_d_Page_028.txt
c4fbdb8a4fc6b290e9e2207f2b7b4ede
9ea01131f08cfc9ba037341a52f60c2d5543a25c
F20101119_AACSAC betts_d_Page_097.tif
c5d2e34bbf13b19e3d4de6697cdd0641
f18488cef9473a6898f26b61479ef53524de8b6b
1314 F20101119_AACSXK betts_d_Page_051.txt
f7d2898927e9042d55ecb47b7c1bde21
6ff6bdfc33989447131b9fa658b7b47eee1ea1bc
20045 F20101119_AACTDF betts_d_Page_048.QC.jpg
02732a4a63929c428fbe121b3e1cbe14
b4d889b3c5d3fe659631b9d42a8c8bbd6645e759
5989 F20101119_AACTCR betts_d_Page_033thm.jpg
1974d6585cd91bd258dd5b3d628d1e90
fccff07c0c03f21e7094585c43b7bc8577b7795a
1859 F20101119_AACSWW betts_d_Page_030.txt
36845652ff93f75692568b486b8e62f3
3aff0ee600fcbaf75e308c50aff316d6dce533d3
F20101119_AACSAD betts_d_Page_038.tif
3e41c36d8221220d88078c0765674f81
5586b674b58f8e2a6f165e0a4909c73c755d48a6
1172 F20101119_AACSXL betts_d_Page_052.txt
ef8d2719151990fdfdb6c3f30b149dee
bbbcfc8b35b36e22995e74301f00d0ca11a1d5eb
5840 F20101119_AACTDG betts_d_Page_048thm.jpg
aa8bbe8b92dc40e1e173fadc460196af
a3bad6038ad415946c17360c9f9ae8bdedc02a4a
18015 F20101119_AACTCS betts_d_Page_034.QC.jpg
8e8669adcf23f44665897fa98cc1075e
43f37837282d07da08e6d748607f5969f4430ce2
1853 F20101119_AACSWX betts_d_Page_031.txt
d892ec71c48a7a028921f8c4e68fe5a7
b796c78f0388cb799d1fc9d2bfd7b4c4306b3fae
110888 F20101119_AACSAE betts_d_Page_024.jp2
d2d7d0ada9f523a23e994c913c82d176
eb1ce5f03dae0ab6144c7ce8831dfa87b1d47985
2187 F20101119_AACSYA betts_d_Page_075.txt
9d32c433dc60de0edb7614785cfa321d
94b8bfdf487b450cbd9bd5f54a670c060943a3fb
1113 F20101119_AACSXM betts_d_Page_054.txt
dce910041613fb964ffcfea56bf7cec6
fe782be90078109f13a6a87bbfcf19f2d15f5d67
16128 F20101119_AACTDH betts_d_Page_049.QC.jpg
7c5713d09b8aab2999b0c189e31b9891
0dc3c2ab2cfabdfafa52639cc7cbcc1f58e73747
5251 F20101119_AACTCT betts_d_Page_034thm.jpg
7fc815925a4ec85f39105bbd311f963b
a8a031ed28b59da71725a487db2e6201ab32ea95
1884 F20101119_AACSWY betts_d_Page_032.txt
ce209ee35a4218454a6431206e82bde5
6861b81380f554f15fc8260f2aa91182303051b7
54793 F20101119_AACSAF betts_d_Page_034.jpg
64753da71db5c0a912b2f081982f6b08
1f481d601190df9cad8fc583d2bf46e0963e2634
361 F20101119_AACSYB betts_d_Page_076.txt
6564b51903f1a202bcf4ed4632f33e28
2d8351a2f4666471788d2b8cfced5613302674af
1586 F20101119_AACSXN betts_d_Page_055.txt
f169001344c2bee8a4cced7d2d999aa5
6e66cdf51c172eb0ba2144c746ff32d814689c9d
14360 F20101119_AACTDI betts_d_Page_051.QC.jpg
8b62205057a12a00ae2c70c2a9472397
f72426a2787f5a1c592e18e99bcf5985ede96187
5472 F20101119_AACTCU betts_d_Page_035thm.jpg
8b8e07aaabd39ba6b0f993e2158e0514
8a8f066a56ae845a4dcbdc34d9781bef185f8f14
1539 F20101119_AACSWZ betts_d_Page_034.txt
e202e7e8deabc86d985c81b619d891d2
8ee5e2c5d777091cc17967edeb0891798eb87710
43819 F20101119_AACSAG betts_d_Page_022.pro
ea007a1beb847dc73cae1022a817592d
7d96b2268e6bd3afb6b40e2078c48cbe7aed0614
1734 F20101119_AACSYC betts_d_Page_079.txt
1898f9762104b1033be1782b8da629e9
9449bc02e8ab15e2e7b24b833f9c00e650238a92
1875 F20101119_AACSXO betts_d_Page_057.txt
70396b14c20d653fbb32bb2992b4f7ab
e6b77740bd8073e6e71c45c405766ba86e76adab
20432 F20101119_AACTDJ betts_d_Page_053.QC.jpg
3637f807650d593f5aab4cbe50a327a1
311c661d3bfb873fb8a7603a4dcae5ab66416f45
6293 F20101119_AACTCV betts_d_Page_036thm.jpg
c131aabd2e9e73eac7bb82bf09344abc
b84746d32c6f114dccefd764ac9fadc25ebef1f6
F20101119_AACSAH betts_d_Page_115.tif
82f91f19d1ff05e9fe7551baf17c6fcf
9dc053f2d805b2d17d0ce732592ec1e208eafe89
2054 F20101119_AACSYD betts_d_Page_080.txt
64fbf29881d401b95ae9adc605264a0a
af476f1bc07a60328f7c5ac9c34783f47bd839f5
2006 F20101119_AACSXP betts_d_Page_059.txt
33031f0edc6079f87ac2cff92b8d2d5f
ba21ca98e253e8bbf285e3cafb4b1c598aa40060
11539 F20101119_AACTDK betts_d_Page_054.QC.jpg
cc68a194ffe71e432509cecec5c52972
b362e0f5e62f32fc1ed0f005a312db5e170572c9
23517 F20101119_AACTCW betts_d_Page_037.QC.jpg
e3312dffdd698cb56e31838f14668693
45ab341d42d13e98c345c3d12a6b119ecac7b223
15222 F20101119_AACSAI betts_d_Page_099.QC.jpg
a600dbfc9f22d559b1c1aee248b0618b
d764d8f7fd06c40a441558dadc454439d9c22dd2
2092 F20101119_AACSYE betts_d_Page_082.txt
7e0a7c1e0e30b97120deef7d6c5e6392
1b41f493252e4192c9d279858b2b2d07634b19e1
1895 F20101119_AACSXQ betts_d_Page_061.txt
5ac7a8b9296ec650adf2e92fed675327
c43f833f034b79d64351483622ec0a398156f64a
4449 F20101119_AACTDL betts_d_Page_055thm.jpg
ffbca63704133564e445979a0a3648bf
eed0f621f5e6c093f2bb61da5f70daf2e7c802cf
6399 F20101119_AACTCX betts_d_Page_037thm.jpg
10863f9fc67565865f40e9608620759b
71af16db97b5d3aa1f6494691495365f578afb0e
12069 F20101119_AACSAJ betts_d_Page_093.QC.jpg
573ad21021c2ae72a428b81596a63fc8
67e5f5485408aa8c0c45275ac815a2344c7b9946
933 F20101119_AACSYF betts_d_Page_085.txt
8c7fe484c59fe1a52a88c9c887dd1a96
af30c80eb6446b38dd5ad1f93d6e8ca25b0160d7
1757 F20101119_AACSXR betts_d_Page_062.txt
e67eceb31d5846b41dd0e798287ea780
3581bccb128a63fb716c81786b44c088527116fc
17724 F20101119_AACTEA betts_d_Page_078.QC.jpg
1f03aaf705ea536ed4bf2ca57a94aae4
92134b014ef1bffcc11fee64d74183ccf27930c3
18947 F20101119_AACTDM betts_d_Page_056.QC.jpg
7313128d0360db2227ebb3481f677d5e
cfa465100a1b2b25d6f5ecd999b616554cc57f4f
6393 F20101119_AACTCY betts_d_Page_038thm.jpg
7ffb56ff1507dcedcc686df3ca11e9cd
683d8a0dbc36b3dc658d00a765da9938b00624b0
F20101119_AACSAK betts_d_Page_057.tif
e4e2fca9f635fa09f8cc76c2081d7957
b2b585e8e87be705f42be216dfd366cbf5f6af49
1263 F20101119_AACSYG betts_d_Page_086.txt
a7ef95416411e26dbd02503325b75f10
f5f87d9a399e81c4fe819ba3a0b1e3d576294007
2292 F20101119_AACSXS betts_d_Page_064.txt
c48af6d0254bad560419ed1b66f34e67
2412eb4ecae8fad73433a7ba826b0b17095ee201
4981 F20101119_AACTEB betts_d_Page_078thm.jpg
d7777185868abfd711e5402dd4316834
dfac20ae697430902a9a31784c617f0e875cf5d5
5499 F20101119_AACTDN betts_d_Page_056thm.jpg
2f4a121c8cc551f6d0a40999fee987da
895c433737856764f812c03213633d9edb778c28
22280 F20101119_AACTCZ betts_d_Page_039.QC.jpg
9f76e5709af47ea57a8427ee5e8340cf
8182340f0c464f507407855af0b4ecb2ee3f602e
F20101119_AACSAL betts_d_Page_060.tif
f7e8482801f9409e47bfa5b34cad34c2
54f7f0008e2c17f0cd20ff73261b6ec389e0c888
1885 F20101119_AACSYH betts_d_Page_087.txt
fe38299e5dd69eafd4e59b0403f3eee8
d0d7038c1de97fcaddd1411174c3f701dd271183
14529 F20101119_AACTEC betts_d_Page_081.QC.jpg
eb78e52f8dea15c25b3c5a39cd37445e
9275cc21be0747ac6d9b306c9c4f7ee6d45b26a2
24074 F20101119_AACTDO betts_d_Page_059.QC.jpg
3551ac9659558dc1cf89bda6cc6d364d
8a3cb98e817916b9a80af2233445f38d3a8145ae
F20101119_AACSAM betts_d_Page_021.tif
7ffeb74aa176ff2d9fc9979685ea55d1
42dbbb7211451e351118c8b147e48478350adb27
1583 F20101119_AACSYI betts_d_Page_088.txt
afd2dee63d922b22db8615b2213079f6
af8703760218f0725d6b44159e11726a740c6d37
1627 F20101119_AACSXT betts_d_Page_065.txt
5f82d2c8f5b6dd048a8fd914c4087cdb
8317a938bb07bd0f14f87ac66132a9ffdf7ddf1c
F20101119_AACSBA betts_d_Page_043.tif
35f7a5277975d0f4af86dc3efd0910a0
73767fb27ad721c9952e60e142d26c4f34be31ae
20764 F20101119_AACTED betts_d_Page_082.QC.jpg
aa570ba2b3bcd9e6b82fe55a9e235bff
ae609be3995bbad383a09e499518dcb69219b15a
6371 F20101119_AACTDP betts_d_Page_060thm.jpg
ed55547249a9b7e5fa2ad1a0aa2bbdd5
80f40bc3233a56973cb2025e98805b304c757111
20005 F20101119_AACSAN betts_d_Page_015.QC.jpg
15f178607333f86d580d2c2273a8044a
6a342e06acc59c094e7dd05cce1b72426c8bea8b
1724 F20101119_AACSYJ betts_d_Page_090.txt
9514e8e53f9000a3985b5ca449fb05df
69e2e8c1ff6214072f7e244186bd9c84d2de7fe4
F20101119_AACSXU betts_d_Page_068.txt
22cf8324a5a5b54c6dc1d50abed62a36
3622620be0909f5d803f57d599eeb2bd4871c82e
76739 F20101119_AACSBB betts_d_Page_092.jp2
9601673f8a8f71d142ffcf88b407b9a8
8a3c505acefa29671fef3186c1708452b0cce77e
16292 F20101119_AACTEE betts_d_Page_086.QC.jpg
29425d73eafc36b965c28578c59b59bd
897a9c47282e36a9a5be87318bde207f3a4789d4
16702 F20101119_AACTDQ betts_d_Page_063.QC.jpg
60dcd4c01c4352ae4bdcc9d63e61bc3f
23b7bde25b1929bddac77d43c7473817b2efbaa5
6035 F20101119_AACSAO betts_d_Page_020thm.jpg
1efb99fc432743a495cf80e557a27fa1
898dd86fa6ee14c249f22e3db4750b7f962afe0b
1453 F20101119_AACSYK betts_d_Page_092.txt
53f68c435f6ee69fa770827d9109ce02
69035f44c8462c84e9a77dcb3bd56894ee8fdf17
1924 F20101119_AACSXV betts_d_Page_069.txt
e270daa701766ddd6be260996662927b
153ba1766e6c6358340a6d5bc5e8dd9b87478edb
17566 F20101119_AACSBC betts_d_Page_111.QC.jpg
895b8b43bb1baeac6ccd8995989c683c
1b0363935cabdd1f4f48e07c438069a29206af67
4514 F20101119_AACTEF betts_d_Page_087thm.jpg
e8f9216874348be43bc1c8a8edff5ab2
6919d0dadb1fd8e7c7bcd7b9df1ecdf86646e83b
23960 F20101119_AACTDR betts_d_Page_064.QC.jpg
73e7373c8b7f8cf0c06a31d843eb51f2
fcc74e6972539b71ce44dd2994b8908bf82a1b78
6090 F20101119_AACSAP betts_d_Page_076.QC.jpg
22b9f11e09e99036480d04302cc6ac0b
9c6315c2578a7e6cf84c18b12d79e980ea043b3b
766 F20101119_AACSYL betts_d_Page_093.txt
a9cfa14a5260432ed2a7fafa356474b5
7f82ef047b5171a38f8bf06e850f57cf25afa5cf
1588 F20101119_AACSXW betts_d_Page_070.txt
a53d99ae2e1d6d2ab6adc77f64da372d
9bbe2278a1e050db8b062b4fc46d490640be2314
1813 F20101119_AACSBD betts_d_Page_033.txt
6dd3206a339d8aa559d04e9d83bf3d4f
9e3a65ac9d3ffdd63984d4b306665ea20bfa7bf9
15829 F20101119_AACTEG betts_d_Page_088.QC.jpg
39ac7574c49f8bb7d55fb96fafccfa6d
0db77d6f6dc623dcc3a718965728ce1f32fd7cb5
4992 F20101119_AACTDS betts_d_Page_068thm.jpg
7ec747f5b6454352c39c6acdd419cd15
786c0ca48e9e982455e957b22f850f2d951a148c
F20101119_AACSAQ betts_d_Page_106.tif
7fca76abe5db1b65f250c2d86b5490d9
eebf8736dc533e0160d7dab36fe4c39ca05df0f7
1266 F20101119_AACSZA betts_d_Page_113.txt
d348c15ceb21490bbdd7b5746def6c8b
add03d74cf588e1542bda3cb6cfaf9707f20338e
804 F20101119_AACSYM betts_d_Page_094.txt
7c43e897afab0e495fcf247bf89a1030
6ece0d34ea148d892896afcac0747a81fee7d577
1761 F20101119_AACSXX betts_d_Page_072.txt
b879a80566951e5a8187237598bf1bcd
0439a3c79ae5207eb7f395dabc61b366f3fb5591
10828 F20101119_AACSBE betts_d_Page_003.QC.jpg
0eb12043931c3bcdcbf5eea01df8add7
5e1ef117858d13c58091c160abdc18ebbfde2e9b
4688 F20101119_AACTEH betts_d_Page_088thm.jpg
580334252485e4be9df712eaf6a488e9
1a1f936b25771bbca104f342a418fa2f7830740e
20690 F20101119_AACTDT betts_d_Page_069.QC.jpg
7594c23a622c7c0c96f5dfd6bda57380
b1ed6c6bd40b14d52ba123970a1e34a1d39e099d
4676 F20101119_AACSAR betts_d_Page_108thm.jpg
8e676efb6ef872f64be317b00ed34c0a
9d6a346b4d6c71402c29a8c23ae60c4ab9d6b3e8
490 F20101119_AACSZB betts_d_Page_115.txt
a9f61f4863f2b6e8a8de6115ea66c669
1cbb1a77ca8c680a0c2b927612e3dd2da3e18fd6
1464 F20101119_AACSYN betts_d_Page_095.txt
8d7d0346a86a62337c93fc3034e5d5c9
8bb543ab3c541534c44f4a6d0c64849d66d6944a
1633 F20101119_AACSXY betts_d_Page_073.txt
844703a382b5ad6ff49c4467b46bf26c
d1c2b6295d92c94acb42a653d8287e1ae01b26f2
25834 F20101119_AACSBF betts_d_Page_031.pro
a75b0e29a38581fb3335348d983a0c7a
d58beb173239cdca52de3255f7f8b7dea404c1ab
9201 F20101119_AACTEI betts_d_Page_089.QC.jpg
383e7e9aa891894ff75df736ab9db1b3
41a618752124b47b52ebc0aba4c9523cb82bf4d2
18411 F20101119_AACTDU betts_d_Page_071.QC.jpg
acadffeec3e85ef8a828eaeabb77053a
925ceb518d431f0fa4008b5719fdfdeee27ffa40
1303 F20101119_AACSAS betts_d_Page_002.pro
822b94919ebff89a3fd5e0668a9df61d
267eddc1e888742829963ef8b9a180a14b2e844d
1435 F20101119_AACSZC betts_d_Page_116.txt
8df6b1f2d66da8d3e605dc0fdba019b6
38148f8236953a6091e5fea92a0d9540a6c8e1a3
568 F20101119_AACSYO betts_d_Page_096.txt
ccedf83c18e9810490d580d234878c89
7f0d23da29c6148c2dfe3067aa1249706e646769
2313 F20101119_AACSXZ betts_d_Page_074.txt
226fc6b7b5802ca69dde73b8d3621bdc
f01f964872e19f4eb8ca422d90e1ec8039fd2a6b
F20101119_AACSBG betts_d_Page_082.tif
6430749ab07ce6532a1b20aaf754fa56
a400b047a233312c547753c3bef1403d90f9c7bf
5244 F20101119_AACTEJ betts_d_Page_090thm.jpg
97c6b966170fbe36b522b7a338a78e31
55a8d884543a5c2fc761b984aedcd5bbe49c03c7
5526 F20101119_AACTDV betts_d_Page_071thm.jpg
9160e07ed8e2dd8495d2bd056b198a39
ab086256764229b68649f2cf0be225d1a6fa878c
9883 F20101119_AACSAT betts_d_Page_018.QC.jpg
2d240a6ec76ed15d04864fb1838b767d
f4ed6fc52a45da11cc08d738efc5cea74815f4c7
1058 F20101119_AACSZD betts_d_Page_117.txt
d2096aba972a21d166882944dba4c63c
47570f4f9a8c7c154bb41074be67e1428921494e
1114 F20101119_AACSYP betts_d_Page_097.txt
121b77f17ff31212441a352ad091841b
80850d50d611a784a6f3ab79f2f31830d60ed2c4
1012804 F20101119_AACSBH betts_d_Page_015.jp2
e5065623ea760da6d562352f93921cd8
5da219404f3a2219a078fd347e42a389374360f5
19957 F20101119_AACTEK betts_d_Page_091.QC.jpg
6079e18dfe11591c84afd7e32e0deec9
2cf13f3550882a3c73fcbc07b021b3867aa72c58
5175 F20101119_AACTDW betts_d_Page_072thm.jpg
cd9e04a246fca73aa0b72fb0d5fef8fc
c501a91ed27feb1324ea6d0c3fa3487fc2b5b7c2
4374 F20101119_AACSAU betts_d_Page_052thm.jpg
9237a5f393f2065677f56874eb671b40
04e39c949229ded244f98403a192e238776911b5
1672 F20101119_AACSZE betts_d_Page_118.txt
5c1317e2a5fa9ad0a69392e3ffb7df9d
c92ed2f3c5a6438a0cbd100a615bb23499cba6ea
1249 F20101119_AACSYQ betts_d_Page_100.txt
5f24ac984a324868d7533d7687af7f6f
d84308212da7bf4ce9b190647f1c11e55b023c8b
2906 F20101119_AACSBI betts_d_Page_128thm.jpg
08d8bdf846e4c22f5bbef0f9beee9806
138e61bd2b7c804c6d18d22e505e90a410f3a145
5561 F20101119_AACTEL betts_d_Page_091thm.jpg
787f041d54fde7de2d1c8a68b000fee4
9b312183c8173a88f200634b32dd0e627269bad1
6151 F20101119_AACTDX betts_d_Page_074thm.jpg
3ae34b900a2275afbe9488a881b4ad74
d1a50dc6fe0449f4549f856fa2c610ede62283f6
4361 F20101119_AACSAV betts_d_Page_047thm.jpg
5f30d8a1abe7c59b98fbc84b6d17dbfb
1a97a1f17bdc496974308228a795778a0ba55e05
780 F20101119_AACSZF betts_d_Page_119.txt
12cb3473b3b6420875b281be5b737820
4f981c51b35cf0b6099c61e807ff28b82db64189
802 F20101119_AACSYR betts_d_Page_102.txt
dff448a62ff4488c1899a8efe14eb439
efd827d7b9a8783a72ef63268731187e01190c50
F20101119_AACSBJ betts_d_Page_015.txt
e467e2f9022281c1dda7040994bfb5b8
252ce134364aa1a746735a4ad09fc06340bab950
F20101119_AACTFA betts_d_Page_116thm.jpg
65eee2e37cabce77f4157e7ed9792ccc
62565e51589fbfbadea0e2fc8720d927758882a3
3564 F20101119_AACTEM betts_d_Page_094thm.jpg
e7b225bd7a255a123a6a87e5857c3c7e
759617cae747712ab4f02692f58edaefe3c94746
18166 F20101119_AACTDY betts_d_Page_077.QC.jpg
ee5560eb06ee386a4047833a2017c70b
f7b9d0c8ebebff01f922ff3804654bb6a49d598f
113235 F20101119_AACSAW betts_d_Page_121.jp2
581402bda64e28a222772bf2150b24e3
015d57393baf179bb2becf1ad1460ad635b02def
2075 F20101119_AACSZG betts_d_Page_121.txt
adb7a2aa16e9d85c3f0dbe177ac940a5
706be14f55edb8cf78aedeb2360bfe0ff50edbae
1819 F20101119_AACSYS betts_d_Page_103.txt
04d38e279004214192384054c5fb7bb0
9111c2a89966ff95fcb93b916950dc1574cb7233
4713 F20101119_AACSBK betts_d_Page_107thm.jpg
10f67f8e1d5fd69eebfee82a7f1b7c58
9b8d5ebdf4f6ae55c3ed53886756b5259cd5858c
20455 F20101119_AACTFB betts_d_Page_118.QC.jpg
c060365d87a81a9a92376fbdc9b09f3c
b50b9e3f44f52cd1b78903b5fa599f43dd9a7812
9238 F20101119_AACTEN betts_d_Page_096.QC.jpg
f5a2448625e16494ead66bc0e5083b88
ef361646669af480ea366bc9c6884fc085b5a158
5239 F20101119_AACTDZ betts_d_Page_077thm.jpg
d0c5501ef8886ec7e312e47a9021fd9d
8bc973367d5be44d7772dc8165deb12cb989eb72
2076 F20101119_AACSAX betts_d_Page_044.txt
faf39f51563d037a6de42b4d7e247135
24eeaa76a2fe3f8aca9bebdbfafc97b1a6e4517c
1825 F20101119_AACSZH betts_d_Page_126.txt
8bb0308c4fb33cdb20251adc0a5666b8
21e631329125208cc0028c7a91b6eb0433116c8c
1414 F20101119_AACSYT betts_d_Page_104.txt
3ef41ec00efbdef93422c580e0ec80d2
707d825b38c66bfac678f59ca14c1b62bd9895f2
6273 F20101119_AACSBL betts_d_Page_084thm.jpg
8947fe037f40105a626dea078a0484c7
82980503efc76d3cc8f8396f6a40350aac78f68b
5662 F20101119_AACTFC betts_d_Page_118thm.jpg
df301980c3f5428747d840793a428055
930b4461019e3beae3295c0046559d932f709e24
10275 F20101119_AACTEO betts_d_Page_100.QC.jpg
ec156f83d50adbf0e0e63f96167ab613
e3b9c51773123637285274821aea458bd821da97
53506 F20101119_AACSAY betts_d_Page_063.jpg
a45840f76fa8fb65f55e1d00e4f7c9b2
f446ce49c7929ced23815ee6f5a24e37044a35e9
F20101119_AACSZI betts_d_Page_127.txt
fc135813e4886f661480f180dfb25eac
1a5f8ce05671f4ce0461d51d8ec274c1401abcaf
46971 F20101119_AACSCA betts_d_Page_052.jpg
4cb2d0718749e459d08144bab5b67ef0
d2be32aaa89c0d85f94ed08e5c60ebfec2393a84
65239 F20101119_AACSBM betts_d_Page_028.jp2
592bd8b90cdaa68be1981049b7af3581
ff630276a4af40adccf36bd3d04ac805ae5def59
4798 F20101119_AACTFD betts_d_Page_119thm.jpg
2d57d04620a37de4d2910f0fdce3cd52
f0038bcbc0a4ea3f8e92e6f8b825187360ce8ac5
3375 F20101119_AACTEP betts_d_Page_100thm.jpg
d1ab71a9ba873a73034056c029a36d35
d3165ab60f74ce6cc049a4d39471dd8e8d2c01db
42528 F20101119_AACSAZ betts_d_Page_098.jpg
4dcd6be705fee9d5d7676e28f96524a9
5e4a66e1b2dbd2a18c1fed41575d1c4364c22c09
637 F20101119_AACSZJ betts_d_Page_128.txt
03962d5e9ebb458051bc111d3fd614f8
366c5f8f319dbc98067587a4a9339d8e28d72bf8
1543 F20101119_AACSYU betts_d_Page_105.txt
f0e715f0652bb88520c3c0f9c77d0b9a
e2019b725b0c7aa01686b78e9572a6396c49f7ee
3429 F20101119_AACSCB betts_d_Page_018thm.jpg
dc8e918c65cecec87d40468eba86827e
d3babb429bb95bd0a7e36f7fe7458b5d66799381
3383 F20101119_AACSBN betts_d_Page_002.QC.jpg
9e6ec3bccdbd64c0e2c6a9ff2989e351
4438e6550f7f83597fc88d2aa15aaf276379bf63
24601 F20101119_AACTFE betts_d_Page_120.QC.jpg
c54ded212a2d3c8e0a49d0e428a74c18
16c062759db7ae01c7a01b3914e0a7ed53486ad6
2736 F20101119_AACTEQ betts_d_Page_101thm.jpg
f571cad9a7dd368de5b4ff1360e292da
527776dcbaafb94b8f45b6e01f51dfd679c31df2
2638 F20101119_AACSZK betts_d_Page_001thm.jpg
04f66714d77bc54fe5284c81734f2b58
e9abb76fdbc98349203f7ff2bf4dee04d8484afd
1599 F20101119_AACSYV betts_d_Page_107.txt
e4e603796e0b545a73c00c93a2a82f02
5f1278d967b4ce285fdcf55335040511f81c4a05
871 F20101119_AACSCC betts_d_Page_101.txt
59dfb4a14a9c816b841b1c493840b894
be256d8db4a7f59c58008ed5a61a2add0fe5fd5e
1591 F20101119_AACSBO betts_d_Page_108.txt
b8ce6037c0c296960e277e2be0c13587
883c82a5c769c95e4fa0279c41d98cd97e95e669
6720 F20101119_AACTFF betts_d_Page_120thm.jpg
bf33b42daefe929dee827ea2a498ece7
c00081012c6ffb1586f11863512484fb89c42a81
5343 F20101119_AACTER betts_d_Page_103thm.jpg
707a8b46cce642d0ff3911759b25da6f
64820473ca009bc91277d938466a5ddd8f4865be
3495672 F20101119_AACSZL betts_d.pdf
00eb9b27bca5105e5d5b1a7131caf7d4
3fe86a6da575a1753d898d1fe9ff6a401805bcab
1640 F20101119_AACSYW betts_d_Page_109.txt
c30f3f0d69ee63f36440a1ddc95e66a1
4d402f619ae17962d93e1421b0c8f3ed294fea9d
26686 F20101119_AACSCD betts_d_Page_097.pro
611ec400984bc1ac3d124edbbc37366d
2fdc3d242db6d32beedea2548bb9f22296d1966e
F20101119_AACSBP betts_d_Page_019.tif
6d472b051a1e86a0c320a21dbd337d3b
79fb5d9e2a757b3c814dedb88e37f169c6e6ab93
22830 F20101119_AACTFG betts_d_Page_122.QC.jpg
4ec3276ebf99a4b04108bd64c0f94beb
3181b6ad215a11b33154e3d78e43d78d536b251c
15432 F20101119_AACTES betts_d_Page_104.QC.jpg
e781614f5b6407a0fff6a814a8f67416
35cda449ad810fb7daae132404cf1caa476185b8
17116 F20101119_AACSZM betts_d_Page_092.QC.jpg
8cac7416b2fc6742f0eb2ba3a2dcb0d9
14f114930dae7e894d92ff5e360c5b63ba294c38
498 F20101119_AACSYX betts_d_Page_110.txt
75ad401410da1bf143bd5bc92f63c524
8ba6d7ad30db33d5b533313fc4c7145d6eba409c
105344 F20101119_AACSCE betts_d_Page_014.jp2
7c1aae23c5a5021b8c5113c3a80c4362
6597809bd3a0c3b800c4f73a9911079d55f5bfa1
281 F20101119_AACSBQ betts_d_Page_125.txt
030b0eb9d8d58bf6311769d6e032df71
b56dbea624ecb223c7e71568e63ae8dec7a2b7ed
15322 F20101119_AACTFH betts_d_Page_123.QC.jpg
062a57252e7025d9d69688e8e1bf6fe5
9984bbb62157d8541d7f9001ea251af8afabddf6
5270 F20101119_AACTET betts_d_Page_105thm.jpg
be248bba1f51814501905242a28747aa
112551828b346b981f2d85d0de9164fbfcc7bb93
5070 F20101119_AACSZN betts_d_Page_092thm.jpg
bbc23b853053b928c3f263eaf53802b6
a1933c97087dfd8568cd23dfc373d68224d2c759
1110 F20101119_AACSYY betts_d_Page_111.txt
4086cc192f6235bd97f4dad7d4247ef0
ed9bcb557418f6c1336032c825aca8818229747a
5039 F20101119_AACSCF betts_d_Page_040.QC.jpg
107ae5088e959f93eee22e0d9d4954b7
713385ba2a0ae29b49843ad63c26004d64239237
46354 F20101119_AACSBR betts_d_Page_016.pro
650cc1b03929cba617fe3e26dd8e48e2
8a340e1b35c4e95e194efb9d33926358fd1e06ee
5303 F20101119_AACTFI betts_d_Page_124thm.jpg
d08fe834c84297dc938af9f15d3ee072
5f8556c49b3e6a87fabe55b1af053a94833fd7a1
16115 F20101119_AACTEU betts_d_Page_106.QC.jpg
6ff791b99e94605d35908e313a56b479
20c6973b884544519ab05139066df474a7fa5b8d
4950 F20101119_AACSZO betts_d_Page_114thm.jpg
084e30a686f0ae22a1f7f7721d7fbf74
f8ba2459addc8d7d4968813ba2f74d20a00c1c26
1352 F20101119_AACSYZ betts_d_Page_112.txt
39e2ae78e9de06f81afcaabcc7d95e4d
261d1c2e5812240b27a24065840d6277e99107f5
37410 F20101119_AACSCG betts_d_Page_056.pro
fd719922b282e04356490c0cc9469eb7
6c05259764e494df346af618e8593a5f5aa31b71
29465 F20101119_AACRXA betts_d_Page_115.jpg
579447c22c300c8f3ce481c8ddca4e64
a96bcb4e002b82387c5abff5ea0ec1e60c9f6a36
61731 F20101119_AACSBS betts_d_Page_041.jpg
896ed44886a06b6678bb79ce01bba859
384b230ca165cb20ad53c89b8420e02d5404257e
5487 F20101119_AACTFJ betts_d_Page_125.QC.jpg
c8c80328d6c1636b88e14927e604a568
a40198bf5a18bb623e0736ca50c5d7b64106bf10
19563 F20101119_AACTEV betts_d_Page_113.QC.jpg
1aecd83775c2cf305c46ec44c038eef3
fbab2b16beded61abd980f508a39701ee1be9690
6235 F20101119_AACSZP betts_d_Page_057thm.jpg
1f91009941fa9a3606cfb12089e954ec
1a0abc9f53ea08a78cb07715b55f7e4b2721f856
4210 F20101119_AACSCH betts_d_Page_017thm.jpg
b0dfdf70404f029ab8cedd708173d5a7
4b1b35d76d99ce30edffec2279f41cb24e3a3591
16434 F20101119_AACRXB betts_d_Page_006.QC.jpg
7bb3cb5c75a4ef12dfb9e828cb5c457b
9a435281ab18d31cb24fd361e669d584f8434f7d
85195 F20101119_AACSBT betts_d_Page_113.jp2
22fe1cd976e492593ca9c9590b09787b
fccd27958a4076e1bb5e51ad4415d6f5ca615302
1919 F20101119_AACTFK betts_d_Page_125thm.jpg
6fe0f189499b9560bf479df7aec2f0e3
906ccfb693534f2b9673278778950c1d4b78fb8b
16487 F20101119_AACTEW betts_d_Page_114.QC.jpg
3ff19920723a5f8bad87616563a11cf4
88b4fa6a0b381e9594ea50294ec7ed30f432b496
17844 F20101119_AACSZQ betts_d_Page_103.QC.jpg
577242cbfe0c3b74e8a3f90b2e197671
59cc8c419683f5325dce2a4a061b9f7dc1912888
73471 F20101119_AACSCI betts_d_Page_024.jpg
0a9cbd06b3d0c05c554103aa0ecd8bc0
f5cfb7096934a09ecec8a5e1c2938fd92941b125
1534 F20101119_AACRXC betts_d_Page_045.txt
ff3b43c0bd0513de04360780ff80be6e
0f3ce58677f9a508a3610f33c74ab85f2b343547
38926 F20101119_AACSBU betts_d_Page_072.pro
a29648941e97d0c0607f59eb3b4da7a4
1607567b93daa36d4d867bcb35e64e95ce5e5fc4
4344 F20101119_AACTFL betts_d_Page_127thm.jpg
3c482d74101d36d527ac0aafb99ad36d
d46324773dad509c385de5650677d133712b9893
9194 F20101119_AACTEX betts_d_Page_115.QC.jpg
8321e15125b78826a4f2b0c23113e23b
10a38e024ee6b67f1eaeb77f4e04d92145f2e478
F20101119_AACSZR betts_d_Page_031thm.jpg
07133f2ebee134300760c85fc3c06f37
52547f9a20eca60198366e67ec2e70de559bcf8c
1747 F20101119_AACSCJ betts_d_Page_078.txt
955d126f53d614425e688f036614a0fa
8da3a1c2b93b9fb5a47bf07bfff4f3a61fa12007
1522 F20101119_AACRXD betts_d_Page_106.txt
282987a9e2a9a368ea9f624737f5d6c6
11d86e31a97e9ada59fb5d1a21065e5370653e9f
15474 F20101119_AACSBV betts_d_Page_067.QC.jpg
09816c81a0fdf23161919da45729c24a
47c95ef84eaa37bfbf35111bf7cf903b37018eec
9207 F20101119_AACTFM betts_d_Page_128.QC.jpg
740f2da6711d412f182592e3d4c25bc7
8a4b944397ce4e2f6fe7f180be03b4804d91dd65
3001 F20101119_AACTEY betts_d_Page_115thm.jpg
869ee85d46309ed4e697c680384a0f97
f4cda5cef2f60713368142e3a9c95e154882522f
5822 F20101119_AACSZS betts_d_Page_113thm.jpg
98dbb16bc2c920b48206c88cd7e57a8b
988ea12eca8f0b2f18662f6dd3f42917c278f256
59162 F20101119_AACSCK betts_d_Page_043.jpg
da30478e61732375652981d2b343feb1
1d02b944e90e4913606e1fd78b71110743fb1ce4
5616 F20101119_AACRXE betts_d_Page_015thm.jpg
78b86331546776abb0350b92bd6d6505
e47e1893b61939609d205fa92fb73dd59bf1cc2b
22430 F20101119_AACSBW betts_d_Page_057.QC.jpg
13b212b4b54746caa8eaa2d6035e68bd
9fd42e348503d724f688cb97992c015b1595e207
17298 F20101119_AACTEZ betts_d_Page_116.QC.jpg
1a12d5ca4e0576b3aee37a738256f7df
af9a7c50f3d49d3bc37682264608a7cb318a24b2
3525 F20101119_AACSZT betts_d_Page_097thm.jpg
e0cd0cf66543461bca3499e3e9380d9a
95937ef5ea825afd27f9fcbb0166b77b54687c72
1651 F20101119_AACSCL betts_d_Page_077.txt
f059e491935c21ee4e9237a33323e0e4
f72b1a5f101e89f560af0be0ed15eba17eedd1f1
45122 F20101119_AACRXF betts_d_Page_047.jpg
95596eac270ac07fa28f92382b4eaeec
97ac114187d085dc299f932d527343b5a4d7e1e9
F20101119_AACSBX betts_d_Page_084.tif
c241773ef6081b8c395a45c129742f20
b62acf72c5989a1b723da55efd02c73792d3adc5
21769 F20101119_AACSZU betts_d_Page_038.QC.jpg
f658a67db958c65c727b56c305200d44
790a623593d0d830c4c05d53f1cee18e95546651
6484 F20101119_AACSDA betts_d_Page_010thm.jpg
11e260667ad097a0380e9e2ca2d5b5b6
18e97ab6c49e392af7880276ce63364896071aac
1851 F20101119_AACSCM betts_d_Page_058.txt
e87745c4d820c3123112e773a8fa62e7
e8b3ba9ff969ee5c389b35cacca305372b868b5a
6664 F20101119_AACRXG betts_d_Page_075thm.jpg
d95636f293961ff7ecddecf3978ba8f6
41a51281b82396decddd392db5dce48ab827492c
F20101119_AACSBY betts_d_Page_074.tif
6b0908ce0e29cf1c7996517e4655be63
dedf606854187a5f81fa1997415a7e1940ce97cd
F20101119_AACSDB betts_d_Page_017.tif
3a7c3837bc208cda7ae66ed80731ccee
6f36367aac94ab2a11f818507a9bf0918a34b538
F20101119_AACSCN betts_d_Page_096.tif
80e781fb9d908c8a6c831127b267ac44
c052079c4ff154ff0491fa5fac8bae3fc89e604c
67486 F20101119_AACRXH betts_d_Page_058.jpg
607736245da504c55e797c10abe8182a
22d0bd0bfa4b153ef7c215a41e710a63bbf89f7d
47013 F20101119_AACSBZ betts_d_Page_036.pro
77bb5f850c0dd72fca44a41ec4c7445c
8b67f7485b222e066d8597fc71895134e28274e0
20229 F20101119_AACSZV betts_d_Page_074.QC.jpg
5e9c107532c2986a9343c36ff01be2d4
c02d909d40f4f611cfb362ddacb5be0ec095ad51
49527 F20101119_AACSDC betts_d_Page_127.jpg
537a87e436def42eca8b998c13aa8d4b
22723943891d5a0eafefcb3b2ccef722898a9996
17468 F20101119_AACSCO betts_d_Page_094.pro
389296216850013e8abd3ea73b95f09e
dcd2b0690df4e0127f0a82b9ec7126e21fd30f14
F20101119_AACRXI betts_d_Page_063.tif
5ea5bf3d8ce30db8e30f511cdc8e11c4
eb78f6dd5adf46bdb7f6aff8589cdf4cadcfa032
12776 F20101119_AACSZW betts_d_Page_095.QC.jpg
87cca1746ff1f24f5a169724878a2607
2463e8baf86dffcd4cffddfbb0e09f8fa3a16d72
83121 F20101119_AACSDD betts_d_Page_056.jp2
90938dce98facf1f2a29879f0b6e5992
25c0e039fc92ff250b3187bdf400bf527d276e59
100196 F20101119_AACSCP betts_d_Page_107.jp2
b3326769178622a2ef50d2dbcac2424d
56c8bb19a567c2d48eef2a9bb01dfaff232e2f8c
6672 F20101119_AACRXJ betts_d_Page_121thm.jpg
cb5079c7cae6f9519cd65949c6c75dc1
cac4d79e13c581ff0ec14adab47b904bf2428600
5376 F20101119_AACSZX betts_d_Page_011thm.jpg
fd13b62715df54cb24fe7489301d30c1
f259e01019f5c841ffcf8d318e7e5fbd114e9e9b
6063 F20101119_AACSDE betts_d_Page_021thm.jpg
7e0b083ddad34a46740b466908a8838c
5c8e4dd2b07b168bee30a00d2346a90c57cc918c
F20101119_AACSCQ betts_d_Page_055.tif
9facaf5c4ccc470dd73ba036395edc01
ba5aa5744921b0184671636d44d4e4c712b6c654
F20101119_AACRXK betts_d_Page_009.tif
995039e565f2faa7ebf49f2b1cd38cc3
aecddb32a06912192e5857aa547507b786bf7789
3627 F20101119_AACSZY betts_d_Page_007thm.jpg
00ba7e6065604cad8b6d95bcd95d1b1f
9f23331946c1364340fbcda822809c26074f9c5e
22314 F20101119_AACSDF betts_d_Page_084.QC.jpg
3f55e49bd19381c34651fd639ab6dcc0
24b708c23d413a39650ffb0336391c253a617592
F20101119_AACSCR betts_d_Page_011.tif
f596d634f0c488dee9e80e1abe503fae
2b2c4314d9c62b541657d7634d40cddef282d276
73608 F20101119_AACRXL betts_d_Page_119.jp2
c1a5eb6b9a92d299c4fc2fd6a1905ef2
8e1f9bdae3ffd76fec0bb057e74a8e87d30374c4
2403 F20101119_AACRWW betts_d_Page_053.txt
20ded8b72f142b2ae9a73fde25ae2cb5
4fecc523015d955917d9829e4157a39a257a2563
8111 F20101119_AACSZZ betts_d_Page_101.QC.jpg
18f3712f27c6e168748e79e36c311d9f
7d646cda90e7101cc7fb9e504eea2f0043771664
6269 F20101119_AACSDG betts_d_Page_032thm.jpg
9d554fecfbce2d66eb47f2a0e67622c9
b6d9f6a8826c2f4e5748bbf18d66485e2b3a75ab
1878 F20101119_AACRYA betts_d_Page_084.txt
f762d0c062e727e1ff43ffe089a87a34
70079988148eb668ea5b91b9eff20649cefed631
F20101119_AACSCS betts_d_Page_002.tif
8e7c7dbb53ba7455c82f4eb8db69dea6
3e348ef77ad66579cef91cf25b949578540c2c4b
2147 F20101119_AACRXM betts_d_Page_076thm.jpg
943a1d8481f64d274f0a583fc2999efb
e1b7e35d2db1be7ade61379a385b21818fa2dbde
17559 F20101119_AACRWX betts_d_Page_065.QC.jpg
4fd8ff46679b1f75f4ee02d341e43eab
800f103065e822931bf2c0b932a980a6d7bd3a39
1402 F20101119_AACSDH betts_d_Page_002thm.jpg
3ec42acebf155a07aa3e2d17001e2083
1fe2146c330c2371e008dd2cef761981afeca622
F20101119_AACRYB betts_d_Page_109.tif
111dc848195953787cb61c822f859cba
28d27bfb0b0a22925b08d7dbefe99a27c92b23ea
F20101119_AACSCT betts_d_Page_075.tif
435d739bf92a3c629bb55480dd0b1357
5ac675e178772a0432eb82dab13641172b1f9bc0
F20101119_AACRXN betts_d_Page_036.txt
f105f5d23f9101bd97701d0a811a6164
ad6686a1950a8a84da9633bcf039716adb293f89
803794 F20101119_AACRWY betts_d_Page_078.jp2
7a7bf88b53de518980aadadc5faff00e
3570d15c17325290a6eb5072c65795ea2f7b7827
3315 F20101119_AACSDI betts_d_Page_003thm.jpg
506e1ddef2714a845de4b50e760e2638
d96def703c594bfbfda0ca0cab73945fb49f1370
108646 F20101119_AACRYC betts_d_Page_106.jp2
6f284d0c34e06c2e6df6f4314f821797
baca82e434b17e386138023dabf199bb186ea702
42574 F20101119_AACSCU betts_d_Page_031.jpg
b426893f4fa2d739d511be63a46e74d8
f02a9d47e66dcb60ed0718209275e482100353b6
56379 F20101119_AACRXO betts_d_Page_056.jpg
df00e90c6249c2df2f9657bfdaa067f8
04fb4b1075b0cba81784c10419aa00e1a9209af6
F20101119_AACRWZ betts_d_Page_072.tif
7c62ac42c85f75a22973eb9c33f1feda
a08c168c073a299a999a297c4c2e5114346efe2f
6001 F20101119_AACSDJ betts_d_Page_069thm.jpg
0e72b8aa0d4403c20aef2ba05e083101
7eb01f4557a3eb17e025a5396c9682085124d788
1817 F20101119_AACRYD betts_d_Page_013.txt
818c7be338c5429d4e17d9e9a0140365
b0854bb9d4f60e643bc1bebea47fcf242e31c5d3
34324 F20101119_AACSCV betts_d_Page_004.pro
8179f38e0da83e30365fa82b32c11686
3527017fc32bb43a9e6089787ad2f18bcfb096f1
F20101119_AACRXP betts_d_Page_066.txt
ce79bbadd07c72941e55dfb2b6adc07b
860a829b8798dca36588cd0a997d622a76c464b7
F20101119_AACSDK betts_d_Page_126.tif
8ca3414c54feb86c0ccf01c24484ff0f
9c0e81d2b0705d1983786dd39ff2426dcfd90aa8
F20101119_AACRYE betts_d_Page_092.tif
3b7631caa51679487001d6b380a7f13b
4755b8efa28f496cb909ba59bdca8313bb606958
F20101119_AACSCW betts_d_Page_052.tif
70d7fa80dcfb341c3cb0eb8cba46fdb1
a0c0633a58aaed47e77f785d1c188927ee7b7e75
4012 F20101119_AACRXQ betts_d_Page_093thm.jpg
5b2bd64f577ad1c201da63aedb3c502a
15b9e055f107e082a76652ebe39e4420490a9cd5
F20101119_AACSDL betts_d_Page_123.tif
7ada2ce327512ac9fdf96f6ec0fcd865
9cec5ad4c8126721a9b2835dbf210439c6709756
61585 F20101119_AACRYF betts_d_Page_083.jpg
45a2733f7fd5484e7b83e51c41f4544d
cb8e4a24d1297b918dc230340b8b4988ce5773db
51143 F20101119_AACSCX betts_d_Page_088.jpg
ae459e012a0821cfd2a939b13b9bac80
25eab50451c7b037ad382f7deaec43ad84371201
21042 F20101119_AACRXR betts_d_Page_013.QC.jpg
47325e4d7699a8e4112526b9666beb4d
b062dc4680fe1e1a42f7801d4ca799e65c74b2d7
96717 F20101119_AACSDM betts_d_Page_022.jp2
02203679febb6e19f790f369dc2f9b09
9d61c60c214302a5e429d9988276b747c2640eca
19868 F20101119_AACRYG betts_d_Page_019.QC.jpg
b3b96fd025604567f3132950db692f8b
716df5fea64edca646f82e0c705f1f85a92b45be
88483 F20101119_AACSCY betts_d_Page_011.jp2
2b6de20fe33d38033af2d613621f5933
55581e7fd6a5d2d5ba8bfcd85a32c99ce9ab73c0
1778 F20101119_AACRXS betts_d_Page_099.txt
ad0581bd94e8ccf92fa4329f9afb28a8
fbab1bd4257fc8912b68ac8add2d33999e4d4d7d
110218 F20101119_AACSEA betts_d_Page_044.jp2
f6a9ca242e4d7be559270a58cea0df80
b6effe1a243aecae10b532ef5c8bed2ae77cf741
60778 F20101119_AACSDN betts_d_Page_046.jp2
74808a5c2ffac0b18803bfba1892e3bd
9cad42df3e5e0b0aa0b34b49542159f8268c8692
4731 F20101119_AACRYH betts_d_Page_067thm.jpg
9b92dc27c7c6f55730b372ab2b607455
9cc728d925c293a9bcd4e92e2facbb385ddf9b45
52245 F20101119_AACSCZ betts_d_Page_068.jpg
0af38bbb767dd9b0b97967674026fecd
37f5f1347f501ae83ead0324f51e79b5024c0865
34319 F20101119_AACRXT betts_d_Page_104.pro
8693ff815f5f9d42c9623e7b10c42424
e71e6e1801c2040acd940edd1fd1ff8594f442ff
15137 F20101119_AACSEB betts_d_Page_026.pro
6024bbbfc8539c87a60994e08cd2f0a7
c3d8d92b5b441daa56cba74fbfce947278faad0c
16574 F20101119_AACSDO betts_d_Page_028.QC.jpg
1b1f040e9bdb5ce4650a526fa96aa51a
542fa84c9de0592dc448e20c9375796d43c72e49
57448 F20101119_AACRYI betts_d_Page_035.jpg
48d5c4fd8923e02f35dc793d7e4767fb
cd50b1b05c1caffdb009e1f653f7554199c8f45e
67371 F20101119_AACSEC betts_d_Page_038.jpg
a118a6248c9a5dc65c74c57be8d29570
1f2e56811ad67d6be6ca78336c06ab609dbf6614
5666 F20101119_AACSDP betts_d_Page_043thm.jpg
1695498b0984b4a7deed1d2a69fb5802
118cb67b18033985468193d37042929e8e2d962d
14327 F20101119_AACRYJ betts_d_Page_031.QC.jpg
43dc6d8d1faabff1ef3e17f51ca5444d
e4c89f82534463ddecc62b4fb80106ac742c4c3a
6509 F20101119_AACRXU betts_d_Page_122thm.jpg
8065343877a4c6218b628f45819478bb
efc54047c286ece797d44e373bd843045100907c
86369 F20101119_AACSED betts_d_Page_035.jp2
993bb48ef3457db3ef8d21e3df8fc448
a49a6dda2abff2f50315c27164d29d1a1576e94e
F20101119_AACSDQ betts_d_Page_040.tif
37b39910a9447ecfdadcfedf1d288247
b4e36d28449c3ca2bfdfb875e7110c9340c8e7df
4687 F20101119_AACRYK betts_d_Page_104thm.jpg
b0244cb5af982557e6b833c14b53678b
2053676a43bc663484c0bd0e85d3d23e30c4c12a
1945 F20101119_AACRXV betts_d_Page_122.txt
9e5c4da637c82cc53ecd7b7bca3f862d
0027a3cdb890201cab5eafe64db50f08522f327a
F20101119_AACSEE betts_d_Page_047.tif
7d9085c618fc877ab1fab153548b701b
14ee7bb372f2ced391a582572cfc74a4c5fadcd7
91947 F20101119_AACSDR betts_d_Page_043.jp2
f3c9807d23ae005dfa16ad5fa10cb7ec
97477c0bb67354de1828f3103311bb8c76af6789
102549 F20101119_AACRYL betts_d_Page_057.jp2
5226ced05f1dc15941c74343e5c7ad5a
4db0e25df93c44667e8c8a85b3df6d46a0b6c497
1830 F20101119_AACRXW betts_d_Page_022.txt
b7170b423a03111ca049e90c572610cd
17358a8e9e7ceac0059dfe2bbd21b24de7f0a36d
14388 F20101119_AACSEF betts_d_Page_040.jpg
8c24ad8fedef02865d0dce56fbdef146
98d77a2b2bf6885495d4e0527ed41985a46cf042
F20101119_AACRYM betts_d_Page_037.tif
25af0e3b65021b38418d60936241f1dc
afc592519aad633ebbb87ed0f0726f8de7fd16aa
1760 F20101119_AACRXX betts_d_Page_071.txt
5067f2de2c728f8cb2aaa01cc5f9e753
fdcffb50f32bfe9c22c01b1446f6ff9922bf1c5c
F20101119_AACSEG betts_d_Page_050.tif
4edf426d32dc615491e7421acce43866
e8248a69564e73938139589513db51c3967a8e97
38148 F20101119_AACRZA betts_d_Page_034.pro
118a981c8921d6bed4ea1c159762edb1
d7aa00dc2f42ea9c90ee8e7c9fa239157cb91781
42334 F20101119_AACSDS betts_d_Page_041.pro
cf716f2838ff9a3bfaa8ee1e396779da
767a7bcdab8e7974eb8787ab894b291da6df39bf
F20101119_AACRYN betts_d_Page_065.tif
ef4e200c68617f2a7b9a052cc6b8ca60
ca35639a0a9861df58925798e3ca350be3c54ec8
1691 F20101119_AACRXY betts_d_Page_056.txt
984f9c51ea4659ca303a59c9413db658
2d528ed60d8b5bfcd55502a1812b24e48c4eddbe
1190 F20101119_AACSEH betts_d_Page_123.txt
fd7c5c4e9284007aee12e64abe33f2d4
a5135f7eac37a3e02801b2aad3693ca9b27d7f71
1051973 F20101119_AACRZB betts_d_Page_126.jp2
40ddf196888dfba2cb73d2c95c37839a
f1f74870152293965ab0ef9e79bb9f03f820fc67
F20101119_AACSDT betts_d_Page_028.tif
9fdc16e3277d41e3823a21ab570560de
8877a521cb60042974765f5ddd0e075bb0bdb46b
119 F20101119_AACRYO betts_d_Page_002.txt
82888eeae093f9c863938f37036bbd03
346985b5d6bda33de1b7bfb798b0fef2f9e37556
16165 F20101119_AACRXZ betts_d_Page_127.QC.jpg
c142ce6cddc13f8235baf7b174b5f85e
5f1a2a01689c50d9d0e82a42877f8a2a59e51f62
19099 F20101119_AACSEI betts_d_Page_043.QC.jpg
f2ac341676273c41ddd461abdfc5419d
dcd4aa22de5a40382d18823e40d52655dff4fb59
73934 F20101119_AACRZC betts_d_Page_121.jpg
f4ac68fb6c72ade311c60fb52fb165bf
5c9ef785c5498085102c817f80ce431b0326527f
1311 F20101119_AACSDU betts_d_Page_098.txt
999aad6e25657eb4c9831ac757e0f769
0dd836bcdce7f8820bc827d46ee46038f8b58f59
50428 F20101119_AACRYP betts_d_Page_025.pro
85c362c8dc3739591d4645f18a5603b2
6ff19198bd827dc416de34f821c690799a77302e
21339 F20101119_AACSEJ betts_d_Page_025.QC.jpg
907c820b533a301b7bf4798839e898b7
51144417fec7cdc1d63e32246f4b260e5046f660
16674 F20101119_AACRZD betts_d_Page_068.QC.jpg
4110756ab0a0b58fb31076b431c10dec
a223257f6847297fb750e6b2093ef575d1ea5232
19824 F20101119_AACSDV betts_d_Page_041.QC.jpg
4ac9f65fbeb09b7d8dfa73dcfbebd990
5cf365bef704447d026fa641b64ff1bb5bd9ce46
1784 F20101119_AACRYQ betts_d_Page_043.txt
eb179772d1bf517e9f02d178d4f39d85
39f6b72ce65d4333874a0cfbdf7b85fc03c6a41c
4690 F20101119_AACSEK betts_d_Page_028thm.jpg
3915fd8bb9c7125f072482a6bc6d1dfd
451bcaff7b59dee6bbe90b817cd5a0470e51668a
14040 F20101119_AACRZE betts_d_Page_098.QC.jpg
5fb5e00246b85fd0f974a17fab135a18
6b8a4e21a2fad84b4c525721cc5ba10294b666b5
2085 F20101119_AACSDW betts_d_Page_120.txt
21100ec1a3c6dd63822dd13e3e082a57
e960bc55f0360bdac70ce7917fcc38242638a5bd
14193 F20101119_AACRYR betts_d_Page_055.QC.jpg
dd1142121370ee1c0fffaefa92da51d4
7ee5f1417408641fd80761b0d311950987e41486
39007 F20101119_AACSFA betts_d_Page_035.pro
c960d185197e085e881997324362b580
b7f9c76860133bd4b3b0285a03c8272caa29555c
17474 F20101119_AACSEL betts_d_Page_072.QC.jpg
50fb4ea43f5012bdb9ccfdd26fa809d1
0ffcf6dd6724ea3be02c5768113d0b1513a13072
5078 F20101119_AACRZF betts_d_Page_026thm.jpg
25ad926464b9fc0545b54960f1b0dd98
77dfa3ce80cd867ceb39ab41bf0a691c522d5c6a
F20101119_AACSDX betts_d_Page_078.tif
e2f799db23a1c53aae8b2ccdd0960d07
e748ea39979cb63a033fbd12dd540bf00a7fd080
27805 F20101119_AACRYS betts_d_Page_062.pro
bb5cc55c2d254aaa11ff18bafb328026
9ba85f701c3c2a71679498e71802c7b9b4b30d3c
5321 F20101119_AACSEM betts_d_Page_085thm.jpg
f9b737b507957c96a02076893e935c30
7697411275bb7a7da42b3368e5554bc0b6bef052
F20101119_AACRZG betts_d_Page_061.tif
300ed5e5e1fff41d3fd4752cfdb7e2d0
c1013c6362f202cc1f90982f41ea578f7ee74851
96690 F20101119_AACSDY betts_d_Page_069.jp2
132d4b7728439a39088d0bbffbf67130
faf233661b6cbdb73ec03835e6f574cc3f0a3c66
27313 F20101119_AACRYT betts_d_Page_068.pro
7050a08e5ef20629e7492ce0e2a4ff24
3b00e6b3da85126f22e81ade27827b154efc066e
36431 F20101119_AACSFB betts_d_Page_094.jpg
4f93b071041b3c0af0a4103fb83836f4
648d6ede67abe41e8d32f3d39eb015de958f0ace
636542 F20101119_AACSEN betts_d_Page_067.jp2
de58377f97cdad0793945028b0d7aac3
226de0b93107392746d7476f78d6e29aa72aa544
4779 F20101119_AACRZH betts_d_Page_123thm.jpg
70a42212f74fbdc320bd27c8159f7f89
49c64d9b98735c05203edb8210277d1812e7fd51
16027 F20101119_AACSDZ betts_d_Page_117.QC.jpg
fd6e721d29a31c492e17c724f1ecbde7
332f933dce63bb71d2b6d01cb3d8c93b39abbc2b
90817 F20101119_AACRYU betts_d_Page_118.jp2
35eea2c43d08aac247c63036eb42f185
18f7d0ab7191fc9bfd6ee8edd92fad729834b1d3
899593 F20101119_AACSFC betts_d_Page_063.jp2
aaea563238efd35f97dd47499f14220b
10822a3e03295fba92fe848badfd574a48121352
6002 F20101119_AACSEO betts_d_Page_002.jp2
b3ebb28eed6f810660a621f398f0c9fe
363e77e9d17c0a061626a31ef8232d89dbae47bf
4663 F20101119_AACRZI betts_d_Page_099thm.jpg
4749f2cc64afcbae7439e2ce60e933ff
54c39584c63aa0978b3b032d60a490e7d2bbb3bd
6519 F20101119_AACSFD betts_d_Page_044thm.jpg
f46b7f90973df1ae5642d80e94917de9
c02df1f755efd4f6ed2f71e0415035466c5bfe75
11685 F20101119_AACSEP betts_d_Page_018.pro
b440acd27aee323c958e03249d7c9f6a
7068858239130844f14edc53df2a2f8c34d66af4
12715 F20101119_AACRZJ betts_d_Page_089.pro
d708b25f0a37f8164d0f957a82844123
0fbe9ed1c5601a80dbed43150a632fd07e37432c
1923 F20101119_AACRYV betts_d_Page_060.txt
40d8c01ad539fec809c4a5a98a3b5269
b877ece2f8cb1fc6f3c6f83f76f89206c0833455
4626 F20101119_AACSFE betts_d_Page_070thm.jpg
3c1bef59e41b9ddd0cb30d11af6bf0a4
ada950bb39146847260ba65ee279efed730cac9d
F20101119_AACSEQ betts_d_Page_010.tif
43107d29fbcbf14316336d2e263558f3
f810b77b071e22f60a9ca5612465c506e77a4ec2
5974 F20101119_AACRZK betts_d_Page_012thm.jpg
38ce14a697267f23499d9f40cb997c5e
c7c8beb31c4c3c5764943e7ece462186e61ede0a
47521 F20101119_AACRYW betts_d_Page_084.pro
9c2e6e800eee68c50085f650947db3dd
f0090927b164dcc411c85136d66554762e17f228
F20101119_AACSFF betts_d_Page_099.tif
77e5e5677a375ebf20eddd00e4a94440
17bdcfb0acf5bd84bac60f2dd8ad46b8c7f9e84f
F20101119_AACSER betts_d_Page_120.tif
44c78f48a424680e8a244475b3d88e43
4a96dc30af17ceba982ef5fc23491e243bd13261
5135 F20101119_AACRZL betts_d_Page_109thm.jpg
1485a2eef6ba6714b6744d39013cb2c1
cb3e740e82ca30641bed66a9bfbdbc2ffffff2d0
87262 F20101119_AACRYX betts_d_Page_010.jpg
4e83493491db128b29605ef182be22a8
ea0c93b66aae0eb867e5c2eec763ea1dbb884c98
65452 F20101119_AACSFG betts_d_Page_097.jp2
aef600935f79e06cb9f7e4544ab35e5d
24556b6f78d93c7de3bdd033790276b93c6037d6
F20101119_AACSES betts_d_Page_083.tif
309f8d12b56423aba15f4ab279997562
e85806589690ad766f98e0349aca0afc458e8369
1764 F20101119_AACRZM betts_d_Page_012.txt
cc691fcd907d319d83f10d3deab4e3c3
a48c240dc8c28b5ec7b36e75d15b62fa0e2bd38d
42060 F20101119_AACRYY betts_d_Page_046.jpg
4ac2f0321dd5e2557e483482b6f8d65b
77605ee9cc7d3a7e89aefb7299e46445787a0e15
5738 F20101119_AACSFH betts_d_Page_041thm.jpg
95a9e1b1968e8067c3adf4476485794e
38d94994fab14f76c89d2a2afedf504761e5e75c
F20101119_AACSET betts_d_Page_012.tif
1e39f943da057cc94a7b45b9f5526455
cc0a7d42acfc0da3d8f1e46e5b102e09f6d449c0
16701 F20101119_AACRZN betts_d_Page_045.QC.jpg
fe1174b9f9a4137b6abc988126e06224
3920638d6158ebe51ccf7f2998241ce9742c5609
F20101119_AACRYZ betts_d_Page_069.tif
bf4fe2ef7f268f80ba2dfbb9b222c926
6fe23a0ab746af0a3137671f5278548cfe680a4e
20834 F20101119_AACSFI betts_d_Page_012.QC.jpg
b43c57e2720754d994a698a872244450
b0b4a6ccd6dd3ef749c73cc9efad4eff759840b5
28745 F20101119_AACSEU betts_d_Page_112.pro
cda1ed15a6f88350dc4c516edf40af93
6eb051eff54f54ced9bb72ee654b36b920e0b298
4403 F20101119_AACRZO betts_d_Page_062thm.jpg
582e1c4256ad35aed8d13f39ed77595a
b11a5f84a92ad673dad149d3220267df832dcab5
878874 F20101119_AACSFJ betts_d_Page_053.jp2
cec9be452f614477654454db1414b6e0
e0f61a65c93b899ec204c9017b0b6fc20f9d4110
6097 F20101119_AACSEV betts_d_Page_016thm.jpg
301dbac5ebcae8b7af5f231f33bc1f43
de70b1df552afcf1927d937134992feb99c221fe
6179 F20101119_AACRZP betts_d_Page_058thm.jpg
cb1eeb7132da61251009f3253b7f0b38
cd3672fe797e90f543d27af2c2603e7cd634c242
93909 F20101119_AACSFK betts_d_Page_061.jp2
a0a5056b3792ed851470a39cc79e8d94
777d1c372e0cc17b3412de6aab176296339a13da
F20101119_AACSEW betts_d_Page_042.tif
dcb4d383dddd840d1ee3cca7ed1ecf70
487820ce567c1904f3d8e97b9e755b92d0f40fc7
1226 F20101119_AACRZQ betts_d_Page_067.txt
766965631692b2316d60c7e72fc1ab6c
d49b54dec21bca6b54c1cbee2d06443014920475
38068 F20101119_AACSFL betts_d_Page_090.pro
a52322ae769586ecc27f394eaf415c60
4b221c0ecdaf62266456ffa0bae24d1a2f3d1687
804803 F20101119_AACSEX betts_d_Page_112.jp2
cc96c0b719a417ca8b36c62f243c46df
2cfe1a5108b9b270540d90a9a3e1b39cd6b1f94f
18161 F20101119_AACRZR betts_d_Page_050.QC.jpg
18e3d357b62b8fb7d439b16dab49779a
fcd5474c4efe5c911cea40f0b2ba7087b158858e
21738 F20101119_AACSGA betts_d_Page_014.QC.jpg
e894498e886f44ed18a427fad475d6b2
5bc66dcecebbbefc120a65e8125a4201eec672f6
10492 F20101119_AACSFM betts_d_Page_002.jpg
b8ec6ef761c1e100417c06295974064a
81426edffdb39068ee5d873641a77e61e28ab0e3
6562 F20101119_AACSEY betts_d_Page_059thm.jpg
88b2a1c4e9dc3ee1c8fc530c59b17be1
98eca1478181d6df9ca7c76851245bad5cc94b99
F20101119_AACRZS betts_d_Page_104.tif
4f37124e231600792d8260e63b8e587b
d5ea6745e8dd7ccd6b21dca996739c28f6eeaa36
F20101119_AACSGB betts_d_Page_103.tif
269fb49a28f8c150c3a006c9734ef5f6
57834fce06b495653f4486d43bc0edec680366f0
19144 F20101119_AACSFN betts_d_Page_035.QC.jpg
803e5b0840ed40ff4684e725ba94ad01
c3ddf56e3ca844039ae77374fffb587bd7751405
40895 F20101119_AACSEZ betts_d_Page_007.jpg
45c4e49cd3d850bfc9b7069f72d29169
a15607bcc67ec116565ea4e46a6e64efc193801f
585101 F20101119_AACRZT betts_d_Page_027.jp2
d4cfbd89b3e25a230bcdd63209ed22e1
030e98a3bc4e7271c904f093ee7cb22c3113733d
56539 F20101119_AACSFO betts_d_Page_090.jpg
4ee78b2a180f4225bfe5f34177492861
4b14a59808f3ab754dff9173cf0280cad5408f97
561 F20101119_AACRZU betts_d_Page_089.txt
5ee46595a0ee7161f608c7f872d93c61
77480da7757f150ed2c3526ef888405eeca76ca8
1993 F20101119_AACSGC betts_d_Page_029.txt
1eb108f7c56f629bfac62fea9a2cd071
708485696b1a6a96e04b5dbb913647293b0e6745
3931 F20101119_AACSFP betts_d_Page_095thm.jpg
518b25486670b158c1f67533c16eca65
4bc6211cedf9006f5832987190fa6596c12faecd
13919 F20101119_AACRZV betts_d_Page_052.QC.jpg
2dcdf8f457952bd3ebdf7d3438b4da9a
82bfb90c71b73779885048278be40e9a4dcc4a98
F20101119_AACSGD betts_d_Page_048.txt
e10c7ae04d394d6e6bc5fc784eedbbd0
c7359d1a5729a6b25105964ce87a577a37d11ccc
95435 F20101119_AACSFQ betts_d_Page_009.jpg
d073e946bb48fbc89521e7b2fe0990a8
18feb4a027721a87291435bc1dfa33d8b3f4d4a9
F20101119_AACSGE betts_d_Page_087.tif
dd2b24c3a91aaff005b6a9a18266ac30
ef174550a951cac6c41a5bbe93e3d049cac6c806
F20101119_AACSFR betts_d_Page_023.tif
6cbb67e36fac836e956bda6eb86c934b
aa73f73ba9a6571591446518141516a264692f6f
24639 F20101119_AACRZW betts_d_Page_123.pro
794bd51e809d9e4f794a41ed3f710760
ccdf3a27e9360f3e23da8dca6b736b573a5c69c0
15691 F20101119_AACSGF betts_d_Page_042.QC.jpg
151709cd795c40647a5c86b86feebb66
9d7d31e86fc5f73e7f11be42cb85a2b1558681e1
4837 F20101119_AACSFS betts_d_Page_073thm.jpg
5f667520e2e74fea1cbbf9992fd0baa9
692fa33bdb1bc615929b5f0888994f2e8989e1c0
F20101119_AACRZX betts_d_Page_124.tif
2bae29a12d025caed9e650367472ba97
e5895ac27b8c53f817521a017824e73675abeadc
14697 F20101119_AACSGG betts_d_Page_108.QC.jpg
e7fb0c5cd1b1b9b12a9add368f88dd2c
737a33f3f9e95e19787571edea3fdfecdb8617ff
F20101119_AACSFT betts_d_Page_041.tif
f32fe042fa4c33715dc530c236f91a91
c66b05c2e7d79000e4ba6a7a20b0adcd8ffdfbb3
97251 F20101119_AACRZY betts_d_Page_038.jp2
9e5ac5452ca75d95b576755799713daf
73012d3316a5329cce2ca1416d37bc8beb5fa137
104092 F20101119_AACSGH betts_d_Page_060.jp2
e4fb2a68cca0602626a04f93ea76451d
329e289040b80e548bfd5dd1a8fabc3004ef18fc
63950 F20101119_AACSFU betts_d_Page_022.jpg
df7873a827dc345df4b74e255b5b0bac
4d21d43a6be0b5c421d4df12202b7a9594936adf
46880 F20101119_AACRZZ betts_d_Page_107.jpg
510cd154b6c7b8e9338e62e3fc032602
6f65b4fe16294cd86cffe02a23f7bce928686a3f
19937 F20101119_AACSGI betts_d_Page_061.QC.jpg
91926e520d668b66bbf022926100124b
d2b660f1f5ecf11a596354a53c333379b95174df
1189 F20101119_AACSFV betts_d_Page_063.txt
e811f3306c13c1f56e6cc902df9f39a1
2aece92cbed56076c46a970be57b4da99bba5446
48288 F20101119_AACSGJ betts_d_Page_081.jpg
5c07791490de4cdaa16198a1517f366c
67b0ff1fa5bd27310e9178c87c2a3dea6a6460d3
33738 F20101119_AACSFW betts_d_Page_108.pro
131da2201c4456a478c1f99a0b3aa18f
01710d06b26bc216723f4ac31961df29640d471d
37542 F20101119_AACSGK betts_d_Page_077.pro
34f645ae99efab487e4d0e79939a47be
8b5a888ae35fa3b5ae2bf5d660e7ee561fecbb07
5075 F20101119_AACSFX betts_d_Page_049thm.jpg
2d284a2849b56171f06e2dffd0ab5d87
1ed3ef0c1e522b9324d4755f8b97d67a7b415bcc
44660 F20101119_AACSHA betts_d_Page_038.pro
4be4442dc25e2222575c32a671a08eca
6be7f391b475c3eaf0da05a310d5c63577bf09a0
62393 F20101119_AACSGL betts_d_Page_118.jpg
546db27234d791909df86dbd5ce45c33
d1b11897f4431aa9b1427afd4bd49453591dc1d8
24558 F20101119_AACSFY betts_d_Page_010.QC.jpg
5a0a01acea82bd64c128118be0018b27
b76ad3c1a5ee14eea00c35411b49607a14ffe173
95828 F20101119_AACSHB betts_d_Page_020.jp2
5605df58dae32293268668323324b84e
76ba80c71824ca75490e73b0d35fc0fcdb001873
76242 F20101119_AACSGM betts_d_Page_004.jp2
9b2bf0f95750eb8dbc4580c65226c6b9
a1709a3d62d6cb532396538ec1f32fdcf42a4d6c
5878 F20101119_AACSFZ betts_d_Page_019thm.jpg
3a239cf7d98bdcd4f29c111835236e4d
783ca6bb5e179c65366b2dcafaece707c9b31b30
4742 F20101119_AACSHC betts_d_Page_106thm.jpg
cd93f7db448490d5ea1182b265b09dd6
08afe46194d8b9c18489495e2eb798bbd78bccc1
F20101119_AACSGN betts_d_Page_077.tif
c6af9b483acf982178d905d72427ac1a
e011d00ca6ced993c99a7b565252215bafa6b261
18975 F20101119_AACSGO betts_d_Page_112.QC.jpg
51b2a93fe221ddc5516768c1ab96d47b
5c1ea1a3e8222b9ed6364c586287d1a4c84cc81f
2073 F20101119_AACSHD betts_d_Page_083.txt
a7057ccbfe6b9e43914b715390006f07
92abad100648c28f875a883d5bc5adfb11bec50a
800268 F20101119_AACSGP betts_d_Page_065.jp2
10a03f66b595c32c96093f7a095e3a3e
1ebbaeb52522dbe5ba6cc4902b47271dfb1c0f4e
47476 F20101119_AACSHE betts_d_Page_067.jpg
9bb996a538d74a7a28cfee156365f122
149992a963527a123805ca3db53282860a53a993
22944 F20101119_AACSGQ betts_d_Page_117.pro
ea1c25a633ac27731f738efd46456d6c
5e241222f6f2c392d458972114e431bbf546b22b
24227 F20101119_AACSHF betts_d_Page_121.QC.jpg
197d6b07b54e7e833e5ec2a24c21e939
e2e3a4a31fe36fb247758e0a3dceb5bfe811f5ef
830 F20101119_AACSGR betts_d_Page_124.txt
065e0cec810027780462973a16794b63
656b813a2e3f2ce86432a7876090989c5d60c12b
F20101119_AACSHG betts_d_Page_081.tif
192810cefa066f8f04a82c26dcac71c6
8d8d2cfcd5a02f599dd05603289f99a202949128
36184 F20101119_AACSGS betts_d_Page_107.pro
721cfd73ff3ccb6692ce9088631ed670
b5f17e86ebf6611d2e14637c018057866932f209
41967 F20101119_AACSHH betts_d_Page_099.pro
54cc5b93c59cdb2abd64e161ef4cc3f3
97f477843c06a0aa22c4fb296801173a77fad8d2
14096 F20101119_AACSGT betts_d_Page_096.pro
3c91bbc10f8f07c98c108c004a659164
1923d5b8b80346b3665d8a359adcb2fa5259b43d
69394 F20101119_AACSHI betts_d_Page_060.jpg
4ace39ce18bed16604c8245fd89f1ef8
289053ced2a9a9ccd4b5739f8b0ed31d02078080
16038 F20101119_AACSGU betts_d_Page_119.QC.jpg
7932edb6d7e1a2b0d86a34cc57fda462
472be4ea701729c2ad94ef0a22bed6859eadfefc
3806 F20101119_AACSHJ betts_d_Page_054thm.jpg
4655565475fb7be9a36436194219957c
62d2145aad701b01b8350671a72aca5ed60099bf
449032 F20101119_AACSGV betts_d_Page_054.jp2
2e23d541deb53f595840f58625bffbc7
ebb0e7f58c75e3bda3af61cd2d914b669dec529d
18839 F20101119_AACSHK betts_d_Page_124.QC.jpg
d4c843fb0c32d8521cb59d0147a558e0
78cf00d703309753511561e1dc8e1bc08e6462c4
1119 F20101119_AACSGW betts_d_Page_049.txt
41a889d31adbd81ae676857385be38bf
95d9f23e465dae2864f032d185de0346f4ae4414
4207 F20101119_AACSIA betts_d_Page_081thm.jpg
2122cbef1cf344c518003c4b36fb63ab
0115e77a52af59b72fd7944b769a9c5eb1c477bf
18683 F20101119_AACSHL betts_d_Page_003.pro
d31b08a06ae732536b9a7d603602a3f4
657f6cc50c4059d40a9f07c388150a206b5574da
5863 F20101119_AACSGX betts_d_Page_025thm.jpg
44aa7d659ec4c07a5d6401d6a5b3a366
f3a3bb7f5f01c2931361a405ef341981ca850b78
F20101119_AACSIB betts_d_Page_004.tif
0eb4961a407789189726cb022076c671
1871038e167dedc554b46608cce8b904f860cba3
6070 F20101119_AACSHM betts_d_Page_008thm.jpg
efb561720c1f77c46c33bfca385c7e8a
73712e14299772a0525b35d74ed8132ccc08d3c7
1443 F20101119_AACSGY betts_d_Page_114.txt
bbb6f24521ff831bb1d9c31f35894f4f
d8bdf8c67ac4306fab6c090472d3f1a235beac93
5446 F20101119_AACSIC betts_d_Page_079thm.jpg
b0a94a60fa9f00926165f743120847ff
bf9597cfee32bdb10958f4fe663bb058bfe0072c
5930 F20101119_AACSHN betts_d_Page_013thm.jpg
570c3971a909def317fece9e97dac259
70a3661b9f89c93aa68a4ab462d167842e78da20
17480 F20101119_AACSGZ betts_d_Page_090.QC.jpg
46ac66730a638a9637cc98b84016658b
6e7f70368bb2e75eb3c8f1817f2bcb7f4903a226
53890 F20101119_AACSID betts_d_Page_116.jpg
3c842b44e5f1ff90e129e00575564b3a
6de75c8c23d5c3a6dfba105d575ab75650846a38
52644 F20101119_AACSHO betts_d_Page_072.jpg
6642187eed951332d6d5f9dc0521c92c
f6695f7df43036550ed21ccea3b9500a488482da
F20101119_AACSHP betts_d_Page_036.tif
25e997235ce4d2dced3988f3bf28bacb
54c3be75aaa61f58fa44b2f8414c57a9f38880a1
46319 F20101119_AACSIE betts_d_Page_123.jpg
613914405c23c4ecdee6b64cbcce02ae
bc27c60bc05260060b520d09eb2d79802ebe0205
4853 F20101119_AACSHQ betts_d_Page_117thm.jpg
61390deb2dfcb676ca5afd77884374ba
3c12158fd9093efcbeb5f522e2ac8ca3f629a668
5498 F20101119_AACSIF betts_d_Page_050thm.jpg
a99d016355d372bdd2449ce00fdf8f2f
9edd06be4c4be6fe86cb55c8ab1204e959b6a6ac
16097 F20101119_AACSHR betts_d_Page_026.QC.jpg
3f9ca45df15cda7185a11454cec481ff
260f27bfbf44be35deca9422a0d0e5dc3c5ee422
5153 F20101119_AACSIG betts_d_Page_111thm.jpg
4804aca6ca52eba6bded0d3da98bafe6
38f883074c8bb3254fa89f5377b9dd8c8a7edc77
17297 F20101119_AACSHS betts_d_Page_004.QC.jpg
ed42519855a010b3048ba6d31e356107
a475d32e95ef49adc2befdbe8b4a596a9acb4e01
11978 F20101119_AACSIH betts_d_Page_102.QC.jpg
4168b9414629b0587773475431209863
6cab6778eae10da420bce10d522f2b99caf5e422
35990 F20101119_AACSHT betts_d_Page_102.jpg
255b91314fabd4ee73d812d3b4df72b2
44318f04b40911bfd762157bd3eb4fb275980bc4
1407 F20101119_AACSII betts_d_Page_081.txt
3d2790e95aa3c1e1e893b8f2eb0c7227
0d5d6e5623eea0e7aa389b9725cf6bd6da637bb8
F20101119_AACSHU betts_d_Page_125.pro
967b85955ba8a748501f0159b0f3be72
81a74997f83a7cd39860e323ef078d663f2ec7f8
27088 F20101119_AACSIJ betts_d_Page_081.pro
bdb5216e2a2cc6216975fe195fed6360
905585044ed3235fc1c9a3e0d53a812390f38136
49667 F20101119_AACSHV betts_d_Page_027.jpg
e2bb606078f866067f4152773012039e
a3ed211de89456d5f038e3d1edb90b8e79e87560
16461 F20101119_AACSIK betts_d_Page_109.QC.jpg
e2a8fb6c4497d09f0a011a84f54c2c9d
ec9474db2e00671401563c9eb2133962a15093fe
10335 F20101119_AACSHW betts_d_Page_097.QC.jpg
05e9adee59dedd9c85eb229fd073236a
2903bcd360b67c6a0187de520830f58ae8558d4b
F20101119_AACSIL betts_d_Page_127.tif
fc8dc8c0bee0d0dff05b8d3028b7bf02
f2081651530aaa5d3be41251ea84e8cd5115940d
84686 F20101119_AACSHX betts_d_Page_071.jp2
b3726c2c110cd4f27ef85b13b3d21877
28b48f3cf846f3cdba2020d12b98bb281af5412d
66163 F20101119_AACSJA betts_d_Page_016.jpg
8007f8de399cb925a3172acc17fa4c85
b88ee813bacff731de082231efab5151207cd9d4
147561 F20101119_AACSIM UFE0012167_00001.mets
b2934abf0ed1f14186f09f98fc25c126
4cf47fc4f89d1957cd9839c7446a735b47704ff8
42662 F20101119_AACSHY betts_d_Page_103.pro
f83e96a048e1d72d3ca1bd9e2e589cea
8df858cefb8ce8c09b4ff00182ee98fc39b4e68b
45128 F20101119_AACSJB betts_d_Page_017.jpg
c32e5c3cf31a700153a2f89516f21fca
182343b28341dd777fea930beb46b95dbd844d01
1703 F20101119_AACSHZ betts_d_Page_091.txt
cbb2ecc2dd911b17167602eb6f63aaa9
abf4c6491a1b1358e956a156b334f280e94fa69c
29829 F20101119_AACSJC betts_d_Page_018.jpg
469962bf0a5bf5a2c89bce939e9354bf
7f83a1a05c9e58c0beb1ca95a7f34887c7ba0d76
64734 F20101119_AACSJD betts_d_Page_019.jpg
2bfdc39c1cf3f7d4e6df2515258bd112
1a86b6c51800c4f20fd9998fe65f88da847a2397
63223 F20101119_AACSJE betts_d_Page_020.jpg
ab81675280204dea3c7c2727158df133
4de99e0a542a1cea2817f55e2f547346fc219005
24419 F20101119_AACSIP betts_d_Page_001.jpg
021a3b001ac22ac890e4f04fa42802a1
0d80ce26525cfaa1c391caf7be746b782c9eb4d4
32756 F20101119_AACSIQ betts_d_Page_003.jpg
610212d233767bc2e483040e453bc60e
ba13df6e901dc5319e403d70626585c4dcd8d700
64144 F20101119_AACSJF betts_d_Page_021.jpg
836ef5beb7b51ff8c14f592b0d0dbf6c
3a5a10f9e6d5b9e16ecb7ac763d1ceacdb839fcf
52980 F20101119_AACSIR betts_d_Page_004.jpg
3cb1ccb78faa33270656115a66492b61
ca01ffafc3c647e7a8a44fe60b28f8cd1409b2a0
60238 F20101119_AACSJG betts_d_Page_023.jpg
5267b88aa8c7353dc135388077db15dc
d38deaa0956975928a880bb9f60daa15159c41d5
65177 F20101119_AACSIS betts_d_Page_005.jpg
7bfa4e2abc7de7bdb414358e69a7b2ce
a62dc70dcfddb47459b655a19fb8c88c1926256a
67262 F20101119_AACSJH betts_d_Page_025.jpg
0fa2ff30d5c5c88b0bdba934c6c6717b
795907a5a25c0e9c1626b3abc673926e4329ed65
64013 F20101119_AACSIT betts_d_Page_006.jpg
7fe33df59474967055498f4803b239c5
170a449f37ebe4deb6f515dd537b96a0e2d288a3
49213 F20101119_AACSJI betts_d_Page_026.jpg
0a3062684d1fcb7cd1e8988171afb694
1ae32532ba3f1a7bd381034e2bf5542f1a74d02f
79551 F20101119_AACSIU betts_d_Page_008.jpg
d00683345d3b1ace6bdae347c8965eab
76322611477bbe5d8bea73e70f742407c534e707
49846 F20101119_AACSJJ betts_d_Page_028.jpg
50d05dcf7f21c591c08172f84aeedd75
2e5b9b29d890305406e6418a7828d54adf7c3f19
61663 F20101119_AACSIV betts_d_Page_011.jpg
32f81df9ba79bedbb6fbe4e696836166
80d17f4943fffc0099e97a0e8c93a68d3c32f39b
71905 F20101119_AACSJK betts_d_Page_029.jpg
36f7aad153e5c474acec190f13ef473d
47c5f6580c6b533aee3c25088ea53b2385c112b3
64102 F20101119_AACSIW betts_d_Page_012.jpg
48e82e45b68d96cdbaa1ae5d49659e3a
302aa764a89c7143b30bbaa0e2a17935a72320e2
44238 F20101119_AACSKA betts_d_Page_055.jpg
b3c3999e93d81aee7bcddcafc2151a93
34522d37d57f1ea0cffe4896c1d39f954cd99a94
58423 F20101119_AACSJL betts_d_Page_030.jpg
80f02071f48a6ea4f68b9ce09da0f911
738ab6b47100a3250947db413de93bda71b413d4
64545 F20101119_AACSIX betts_d_Page_013.jpg
45c50e05a649f890e5e1e6df2de26b7a
9df770757890a02317ce1e6ec75b00cb0f131f5d
68686 F20101119_AACSKB betts_d_Page_057.jpg
169eb13c817988011f2ecb7c003c58d1
4f80243124b18104f431bc215d97ae986a033fac
69210 F20101119_AACSJM betts_d_Page_032.jpg
71d0d142ee5f552aad38b0d0682cacc0
66b4b311e8b579ef8fb94e3d6ea037d8f65ae459
67890 F20101119_AACSIY betts_d_Page_014.jpg
213492e0e4397cd53e8225caa126cf5f
cc464eaa879327baf27b8167532ead35aa80b65f
73274 F20101119_AACSKC betts_d_Page_059.jpg
513fdb269e9332e7ac695c599415137a
2b8d3513ddd01d342c6d895312f4a3df69dcc3f9
65421 F20101119_AACSJN betts_d_Page_033.jpg
fd91073005fa930dafe24a44347e7adb
db84b464a5289ed7a2d65836ffccee36ea154a1d
65497 F20101119_AACSIZ betts_d_Page_015.jpg
de2fd5e86addfa0978cbc622dea4b1c7
364abe553744b6cc9581d05203ec6e08ba8012fb
64050 F20101119_AACSKD betts_d_Page_061.jpg
c30b806a1b1890026ad74426b0c4c973
c1956e9065ef361afbbe0596099ae48a0e3555cc
68646 F20101119_AACSJO betts_d_Page_036.jpg
1802d03a0e037ad1c1c0d5144d19a9e5
8f42a5b7772304d6cbc0a0e6eabc5d2c4e27f1cf
48493 F20101119_AACSKE betts_d_Page_062.jpg
a2f1496bdb7ab0612967db8e10b337c7
36434768b9b0308275bf0be658ff851ff6ff195d
71883 F20101119_AACSJP betts_d_Page_037.jpg
4c5e7cbcaec2bbc70e138a746de0cf67
1070a4601d6ddd7da930b5d8eb3a3a46270d14d8
81188 F20101119_AACSKF betts_d_Page_064.jpg
f469f0884b71ffdb4b2d9c37caa8ad6f
1ca0f8c5541e292e7c798721562dab607caa3b5d
69385 F20101119_AACSJQ betts_d_Page_039.jpg
440732d4c28b532c0505734da201eb03
08385fa1d94a669c0fad310295d366d109be3ef5
51500 F20101119_AACSJR betts_d_Page_042.jpg
a9ced9f3e07e89ed1e4d3ae7279afa07
c93ef84803423f437ac1dd85a1f64d3e3514dccf
53745 F20101119_AACSKG betts_d_Page_065.jpg
ce7b8f87315a72f0e740111be17dc6be
d0ea20ad0a1e888b8f5279a761a1181423ce0be3
72931 F20101119_AACSJS betts_d_Page_044.jpg
1e726659ec850c6cf4cabf65b202237a
cd6d527715af0927ae2af82535775224622d290f
37111 F20101119_AACSKH betts_d_Page_066.jpg
2855cf8c913463b423f8ff65004012ed
f2611da7cd90ab3ae53134d555da0574851d0375
53280 F20101119_AACSJT betts_d_Page_045.jpg
06b29eaecc737d9ac16ebbc7bf5b8914
c6af2d5310e54474529feff506f7ef8420b22d0b
63484 F20101119_AACSKI betts_d_Page_069.jpg
07dce6c1cf9d3a07633b9cac0ab1a7d3
98f42d0149492226bfbe60b82e68db086246af08
59245 F20101119_AACSJU betts_d_Page_048.jpg
939a25feecba2b9589a1d6fc0cb9b6d9
9718c8439493ac9f8ab35c3d459cf29a6808624c
43413 F20101119_AACSKJ betts_d_Page_070.jpg
b62cff199d0b5492914f3f5bd06722f8
369b0fbdf4ecf72ce32a96c120d4712be72fdccd
52394 F20101119_AACSJV betts_d_Page_049.jpg
40541d266550628509b97f916d45dc8b
75ecdd1d6ce14838f71d0ac933f1f1912ad5dc73
56220 F20101119_AACSKK betts_d_Page_071.jpg
ffaa5ce96a55b42cd5a87a509d642017
f13739fdcf52714dda1bace653f048d921e0e546
57433 F20101119_AACSJW betts_d_Page_050.jpg
2344335e6f72edcfcfa43ceb86fb2afe
9ab3ec042ea4cabc81bcadbf8fe2dbb4713d08ae
45556 F20101119_AACSKL betts_d_Page_073.jpg
27ef6cc652eed7d45250e05317594300
26738991b475bf5e2d9a4c6edc2dee124f93f282
43487 F20101119_AACSJX betts_d_Page_051.jpg
229c07a343feb239981956f672c3ef61
db2b6c43631caedaefd361ff56f208ff9336d016
52277 F20101119_AACSLA betts_d_Page_092.jpg
cf6c182dbc1d46a11b3939a46759d40d
934a05c32ea01da98db14a7dcd3cbfc391f9d994
64157 F20101119_AACSKM betts_d_Page_074.jpg
5f08fe1f7584ec29fc8091560f17fa01
229dd8f7539ba1ef7aa997bfda48e430e4fe8475
63482 F20101119_AACSJY betts_d_Page_053.jpg
56f60a4686d101cb2e4502e813a22b1a
15dfa52445bbf5d3486f8a5ab7e0726eac1ed73d
39359 F20101119_AACSLB betts_d_Page_093.jpg
5999dc967e551dcf328f9556730a2a83
bcd921df46847d99eef746d9f3d8d9aaba61b143
74971 F20101119_AACSKN betts_d_Page_075.jpg
4f6ca43f27352c4020527d726f6bd2ec
7bd11fd2832a77e131af880385116c7b705cf2a3
38282 F20101119_AACSJZ betts_d_Page_054.jpg
1c9177ccc027515b00ff80c834765b4e
8e6a8290db0dab48a94dc93bf18a34b875a8dd0a
41180 F20101119_AACSLC betts_d_Page_095.jpg
cdee315cdeaca053225c618c79fa9f6b
8699af3d6f4f44520c790cb22f1642778a66985d
17956 F20101119_AACSKO betts_d_Page_076.jpg
943509694123198020372ac0f30cd437
8e3eb24393ded1ba2a712cb63352421ccb3ed690
29861 F20101119_AACSLD betts_d_Page_096.jpg
0dc7f1acbbbf8f6ac10a91356a174138
dcc9cca2c4cdf3b0ec39731681ff8446cfe49832
54874 F20101119_AACSKP betts_d_Page_077.jpg
9e448454e8aa3aba657cb51dd24da4ee
cbde9b006b07faabdc38ad5ffb699465173bf34f
32249 F20101119_AACSLE betts_d_Page_097.jpg
d3a204e9a80985107d68a32dd88f9bd0
430446a6755d35b0d0a8fae19884a9ecf556efca
56626 F20101119_AACSKQ betts_d_Page_078.jpg
3da21fd615349624dbbf1389367159bb
caaf789cb9cb30c366d5990e7ef03bbc1a2698b7
50847 F20101119_AACSLF betts_d_Page_099.jpg
e1ac2e075aba74d8ea46dbae30032cb7
ee2ce9fa6f3101823a68b448d20688fa9b75efcf
59296 F20101119_AACSKR betts_d_Page_079.jpg
31cd8ac99d7a1036a366cefabf9550bd
184554ad21cf36abee8747ba10e5efd3ad069595
32087 F20101119_AACSLG betts_d_Page_100.jpg
e93a66e8dbbb0ff3fee7d81e2389cee7
f0cb0c702e9db57b599638a4b922f3f3ebfc25f2
58840 F20101119_AACSKS betts_d_Page_080.jpg
1354102e1fb268ec8cdbd1bd4d454f84
fe230d4133be6cd08d3486724953157d30d65d01
64917 F20101119_AACSKT betts_d_Page_082.jpg
8cddb065aa10cd26a31262c0f5044766
7538d53f925e96d495001ce57089edf6f40a488f
25098 F20101119_AACSLH betts_d_Page_101.jpg
24802af0b77a98470701e51acfb7dc09
4b7e7eb1550554ed4ce9060a926be47df52d714c
69539 F20101119_AACSKU betts_d_Page_084.jpg
224fefdc0d183e9cd9ece7b5d847866a
7ffc080480f5a9b2433eedb07f9073c7ec9ccc71
56314 F20101119_AACSLI betts_d_Page_103.jpg
aa5a5583eac46c9490ed839b64a42e44
6801efa95cd8cf901ac8ee40b851a6dc1f701d0c
52938 F20101119_AACSKV betts_d_Page_085.jpg
e4ea1b3c1d66364e463abc52ea77e02f
15c8e7af94dd9b083a108564e502d6623da9705d
48088 F20101119_AACSLJ betts_d_Page_104.jpg
ed9faf2c13854a50c75d210623f567c4
5303d1b2f9bde9065980f9c56c931296aebff005
54200 F20101119_AACSKW betts_d_Page_086.jpg
150da2b028eef4aa1089561940ce36be
78f902ad9ec8e01ea7a9ac2cbf152b69e6d123e6
52722 F20101119_AACSLK betts_d_Page_105.jpg
b9dc219c3f58966ba59e8dd031735277
2d80c5665f8fb7299e0cc3e1f84d47f38c55ae4b
50298 F20101119_AACSKX betts_d_Page_087.jpg
f020c30f3df803f87df31dbfca56f534
8a80fa318c8f7d4a0a0358e80b0b44b24e9aaf47
25988 F20101119_AACSMA betts_d_Page_001.jp2
3f04d9c224a44b538495a246a9c68373
aa827255b6dee766d240e4847518686573f4639a
50464 F20101119_AACSLL betts_d_Page_106.jpg
a4e658b294075a8bd36fc072146a543c
6b1efa813f5a720fa019a523a4644c86d55ba23d
29841 F20101119_AACSKY betts_d_Page_089.jpg
948dd0c0c5e6d2c5563325733b8e4f00
c5eca3520037f28b3fc6142dba0d356101923e02
43367 F20101119_AACSMB betts_d_Page_003.jp2
653093143906acc3da3224b64be44cdb
35e4004cdeb48698916324be355ef38e7bf4de5e
45354 F20101119_AACSLM betts_d_Page_108.jpg
260b1b45d1bd5a063adeeb4fd11df15f
74eaa947447ebf1009894f481ed793d6e15ca208
61642 F20101119_AACSKZ betts_d_Page_091.jpg
faa3ed6179d9c5672ed5734ce89a4db2
524da16c90d90a1963faa829d263709fb5df7ae6
1051984 F20101119_AACSMC betts_d_Page_005.jp2
9fc8dd8912ddbe6d90d4ef845a682103
0a13c3c626ca0031b6eb26bbb142d0bcde4b277a
49625 F20101119_AACSLN betts_d_Page_109.jpg
2c564ef1c56d77daec9d8a8b453021be
7fbcafbf0f13d5cd1b01c24419637b503525f07a
F20101119_AACSMD betts_d_Page_006.jp2
f12388a98d317b2f28fd697dd2a1d512
bed27b248a78fb91ce25a5285b4188153e4437aa
22342 F20101119_AACSLO betts_d_Page_110.jpg
6a403a79f4ad218a08c781cbb8ba414f
50ecfa6b749353380bc84a10bbb29b9c96a8fe80
1051951 F20101119_AACSME betts_d_Page_007.jp2
f4e3a89b051db3807cd64ff432cb967a
dcb0adf881bd1825bb9c415878c55c4331796038
55051 F20101119_AACSLP betts_d_Page_111.jpg
db6401d5bedbb8a3774a3b62937c9262
8ccf996b120f2016b5d6f67d22b526f7cba35c80
1051985 F20101119_AACSMF betts_d_Page_008.jp2
b77fe576de34ca7c948c934521c7e210
91a67de5beb6fbe7c519ed87d27396dc68319ad1
59079 F20101119_AACSLQ betts_d_Page_112.jpg
7cb4c1a9a9d6b5bcf50cedb6e1936e18
b1e015601e379de1f965216bceb1494655d605d0
1051969 F20101119_AACSMG betts_d_Page_009.jp2
935c34e4eb1a592b163f121a13c0ccd8
4087c0bce26715b508c8fe4d82a6137d6f474347
58890 F20101119_AACSLR betts_d_Page_113.jpg
a9709a5b06fabf3f2845fbf1293cd16a
efeac6af4aa3b5e25be2a31f07d75d45c89b2606
1051979 F20101119_AACSMH betts_d_Page_010.jp2
49e6352d28d7f9b677f12c42b9f6f98b
8f2a71f306f34392bc364db17720f8ff4ba5610e
49559 F20101119_AACSLS betts_d_Page_114.jpg
f6dc2cab7a7cc678d0999381d351509e
3cb22a648f5df49d3df7d72ad11e382430bbd2b3
49120 F20101119_AACSLT betts_d_Page_119.jpg
7def540e0e924d08a0fbb3b5fff6ca92
dcd1ccc75aa53dacddbbd767fa16273a42447783
96776 F20101119_AACSMI betts_d_Page_012.jp2
db59b746d378a17bcca2a667ad24e900
39e682a5c807191b867402af0fd337196ac4b842
75091 F20101119_AACSLU betts_d_Page_120.jpg
19fda4339d4d1aadaa6e6ee80aeeb707
362a34f51530c972c1b9b9b45bd9d4c28624a01b
96624 F20101119_AACSMJ betts_d_Page_013.jp2
2765264961aaf87685c9ed2611a2233f
10f08ff7535166fdd0ebc5561b45c6ce183c1641
70660 F20101119_AACSLV betts_d_Page_122.jpg
d8e7e7c3b58dbc3c1b308e7cb15c49e7
e2c2bb1d415073950448ed0c3c29ec05a2624eb7
97808 F20101119_AACSMK betts_d_Page_016.jp2
05f7fa41bebc3f148a40e38fdf7708e9
d1f833179e89151361dfb353dbc480d5f64c8093
59023 F20101119_AACSLW betts_d_Page_124.jpg
df741e7d37ffc015a69281550b6bd71d
25af3bae497b41bd102852ce62369b1460e17688
103601 F20101119_AACSNA betts_d_Page_039.jp2
5108b48d7f95328f6c49c63b3f292854
2ba0785a04950b5d504d532597ba97e10d07a3ec
66380 F20101119_AACSML betts_d_Page_017.jp2
f698aeb93245d355fb91d7364e4facf2
6fd628ccabb617856e638ca1676cc1b4c691b84a
15869 F20101119_AACSLX betts_d_Page_125.jpg
b4801f88b578f01d57c772be5911042b
bbeb6cfab3f41a143875bfe43a6d522e43a8ca3b
13163 F20101119_AACSNB betts_d_Page_040.jp2
1d08aa96b601a765b3b0f5d3a2970e07
8d80b6577621a1e4a06bc55fb9f4ce80f35b0cfd
467718 F20101119_AACSMM betts_d_Page_018.jp2
4c7ed7163764972936f9e9334bdc15db
f75a264c9342cbd681faa10b54986837fac906ae
74842 F20101119_AACSLY betts_d_Page_126.jpg
6803cc27b52229b6c06693dde20d0e91
85a61174927ab9d5ed9a04653344497c975dac5a
92223 F20101119_AACSNC betts_d_Page_041.jp2
e6fffdd6ed27f6c9d8ddc847c80d09e0
a9717f1b72dac43b6def6aa10eb151fe906c712e
97519 F20101119_AACSMN betts_d_Page_019.jp2
78b4b7dd2a26ac7a2bf1601849632bed
a34617413d2561d26afd04c4c18b9035b17b61ce
28506 F20101119_AACSLZ betts_d_Page_128.jpg
097c9d810e90e8b2b9b7703009fefe31
f63361d5acc625add2ed131ad34239991fb86dac
963134 F20101119_AACSND betts_d_Page_042.jp2
a4ae0f0ebba9591e03675df9cc66e05b
bc22cff24c65228d651d15951b6e828452e4b54f
98767 F20101119_AACSMO betts_d_Page_021.jp2
f47f29de5890eb5c68ea6b8533a6449a
edff0fe8ad855ba5ea4e6e02edec9bd337b2873b
76576 F20101119_AACSNE betts_d_Page_045.jp2
5b6b6671249b79cdb763d38443d25359
a0318acfe56ab2e08feb38ab103fac9748297d73
89449 F20101119_AACSMP betts_d_Page_023.jp2
b0d0bda1e7070895c0386f6e963e101c
8a8bc5ce7929cfed04e83ea7ec7a605e45337c78
64938 F20101119_AACSNF betts_d_Page_047.jp2
ebf51490867e9b25092f5a4f948d2c0d
f5907da79b6ca9a170c5a607cb80ac8870bcf8ee
102925 F20101119_AACSMQ betts_d_Page_025.jp2
726f44d36ac34ab3a0edaa0fe4a159f1
e41653f6dff073719ccafcd7f77712f0fdb96174
89402 F20101119_AACSNG betts_d_Page_048.jp2
ab85d76a4340a79c0af3dd76a890fdd5
9766867d4a91418148cb97cd659ace32dce523f2
738095 F20101119_AACSMR betts_d_Page_026.jp2
0db9979f3577632ce2193204635b2ff1
5b57046de986308c3609940258d770de19b3856e
1051938 F20101119_AACSNH betts_d_Page_049.jp2
20f67bec2be10f56405c7ab069814e2a
bdc4bf80b939ea3758f5314506366964b7e9cb11
110198 F20101119_AACSMS betts_d_Page_029.jp2
ba6e99ba2d372c985d12b2e6ef8bf510
76b67294a83d34e9ca983aef85b5de957601b7d4
83510 F20101119_AACSNI betts_d_Page_050.jp2
6407ef993aba8d4d687c113cbc5b43be
a852f49b87c30108a3787780040afb53ba0a915f
96164 F20101119_AACSMT betts_d_Page_030.jp2
c0a3c77b0b004e098c9de0cd446d27df
d872f70912e89ff4f6598e888787860197a283db
74800 F20101119_AACSMU betts_d_Page_031.jp2
bad26dcb845cbe36ec9aca5021c3fa2e
7e092bb051bf892c561e18a458c76a67bb9f5c96
61773 F20101119_AACSNJ betts_d_Page_051.jp2
0ab21c32bfdec26426852f572e954e64
b5f7133b1e14f61bc0d0e5c27105ea160353aa48
101965 F20101119_AACSMV betts_d_Page_032.jp2
6db23c6bffe91f6b0ccfbbce107f2ff7
b428e992c1a1b2cf8fc3ffa3b8d1737893695ce9
613810 F20101119_AACSNK betts_d_Page_052.jp2
b6d532caee8b36957c1fbe7b3dd2ee21
dd6cb0c87b7f2d8ad9ac4c2eb6d57e4e310086a1
98770 F20101119_AACSMW betts_d_Page_033.jp2
8547e63b33897346b50e74d2bc2abec1
0105de2584e0291ccc4d25762635c2b408c5f592
62013 F20101119_AACSNL betts_d_Page_055.jp2
4beeb05819fc8e6c478bc5f3c768bb0e
771f3c5c603dc59e8d875cacb5e4e55e8b23d58f
81436 F20101119_AACSMX betts_d_Page_034.jp2
8fe1105a4a7467a06bb61b9906682a7e
51b73acc4a2b24e7ebbd91d010520a6e763e6a54
883521 F20101119_AACSOA betts_d_Page_080.jp2
ad780a46fad48d3a932d095c005dcbef
ab46a519efa928ce872365658362916133b09033
101780 F20101119_AACSMY betts_d_Page_036.jp2
5eba5809db7dfbbe716079778952cb44
d60506b2b9187d178ad1ca822988d338c5517b14
713352 F20101119_AACSOB betts_d_Page_081.jp2
fa06d0de70a9153ed5266b350b524bbe
b28b6498a531555f2e8d0f5384e636f85e46723d
102585 F20101119_AACSNM betts_d_Page_058.jp2
1f7544935a8f4575f7c83e81a5a2ac05
56fdb5e12a177d0a074250f6f3123deae4d412f6
110413 F20101119_AACSMZ betts_d_Page_037.jp2
6a764781f79c0db58e79d32804cf8e6a
6a98f0f54ce985ec0debc7a48ebb011427c2c22b
945318 F20101119_AACSOC betts_d_Page_082.jp2
417c0a0ab03e3ad5c9a834330f54c1a1
21d9837f139dea566fc6bb9d84e2c8ce63f3e1f0
111290 F20101119_AACSNN betts_d_Page_059.jp2
4248d06ece21a4f9c7de3982d145634b
391b0404b823d79327045fb5d352f824c38b2a46
919106 F20101119_AACSOD betts_d_Page_083.jp2
763166030b3cff92bfd97fa3fddde9c3
34c7d5a4a2eef1954d4692728973b68240e6cad4
700077 F20101119_AACSNO betts_d_Page_062.jp2
93464da531ae6d17215ddee2015a5dbe
4c5e47182d864a7399d99b01a2945e4b24fc207e
104602 F20101119_AACSOE betts_d_Page_084.jp2
a02b9f91a9e6ebf999332047c9aa0b98
8f5eb3fae42dbaaf6d7f7feedb5290758fb98191
1051971 F20101119_AACSNP betts_d_Page_064.jp2
b7e0a20e754ab0467652d9c679ed8007
563ea9432be138f6b73620c4abd69a7aaa102cfa
878394 F20101119_AACSOF betts_d_Page_085.jp2
4f8c84e45df8b721b67ded36326b7a98
02bc4834d3cf155ae065125a60414e6fc5b241d4
469224 F20101119_AACSNQ betts_d_Page_066.jp2
583ebf62c18b4d9bb0fa8c55e0dffef3
a926a1f7f56cfbd7ca544dbd39955829913019a6
936219 F20101119_AACSOG betts_d_Page_086.jp2
6a4b22e6b7711187b2908f3d5eb75ef8
9f58c172b98f9154710874ed218b6abea9889dfc
1051964 F20101119_AACSNR betts_d_Page_068.jp2
e6cf681d6c3911796aac0af3a8cba7a8
57d55522e9955ab859adcd4aea50fc9688a73980
833272 F20101119_AACSOH betts_d_Page_087.jp2
85a9e3f0485a44092c1dc793ff5f99bc
6124b43fb8b0481e6b82b6513ae810b531c30fa2
63666 F20101119_AACSNS betts_d_Page_070.jp2
b68f87e2a4c79a61b4f04f9b3dc7a74b
ec2408e9bbe135a3e35603eb9def13a594b0c623
728259 F20101119_AACSOI betts_d_Page_088.jp2
2e6c92a3b95fadf1ad8ebd0a464661b1
5adc025f77e1e55ea1da74693bcbfd2a883b4e38
79490 F20101119_AACSNT betts_d_Page_072.jp2
b11e2329d72c395ee4d69ab3196bebcc
c2275ffa95d7daa0cf76424c4a4b94903ae09439
397652 F20101119_AACSOJ betts_d_Page_089.jp2
c47dfb6ed4fd5406456eb71b1f95e4e2
b49557ef9ef71b0f04a85d2883a2cde0ef5a3bc2
68660 F20101119_AACSNU betts_d_Page_073.jp2
b261238f218a3de611b508e82d09ed16
fee6dab67fbdd4705092c4d455062313bb85c5d0
954760 F20101119_AACSNV betts_d_Page_074.jp2
3c9fdd69dc5dfa5e806bb283d918e6d6
591ff8dde8451ad768ad37944c813273602b783f
84354 F20101119_AACSOK betts_d_Page_090.jp2
9ac824d8507170652703618de3c5fa91
2e7b7a14f67669562d2207778911fba8bb824961
109557 F20101119_AACSNW betts_d_Page_075.jp2
ea51c80c9866efc5e16108518980438a
c2ec4b774a1887310c16757c2bf89a9d149bb584
796661 F20101119_AACSPA betts_d_Page_111.jp2
32236c9a92be866bbcf68d8988d5a14c
567e1e8b417014bef8b18cf28b624ab27293c0a4
92597 F20101119_AACSOL betts_d_Page_091.jp2
f5f3d8cecf6d981c7c6669b4b16464e0
1912c6b82cef7efea3620f5dd0ecee4e96cccae1
20627 F20101119_AACSNX betts_d_Page_076.jp2
edc7697a4b8db700e9900e95e800d061
c3df65c450bc1c6c8e445d66fa99b83b66ba8bb6
71413 F20101119_AACSPB betts_d_Page_114.jp2
6da399c8f7b935dd8066562ae7a28181
da45909d7a1f02dd24a37503e88abe4e869086df
71153 F20101119_AACSOM betts_d_Page_094.jp2
c28025f1c6644e52cb59ecbb21549735
4fc1058835879ae20b371319da8b817b056a16f7
80310 F20101119_AACSNY betts_d_Page_077.jp2
72347334f7a5e07922f9a45baccf30dc
2d13a111a8ff8c43453773d55d032da43d4fb3b1
490164 F20101119_AACSPC betts_d_Page_115.jp2
d1ce42f741b4904ab90eaa1096cb9e04
22a15ef09355048551c7b2c72b2a419b21036f43
85521 F20101119_AACSON betts_d_Page_095.jp2
7455c24d422186935eff90cc51c079e3
fca98e65e16f837726c229cae3535631b6b8f777
867421 F20101119_AACSNZ betts_d_Page_079.jp2
cc59ba5666a16c4ee8ae4edeab909d91
e1fc4c525d5c0a969c889f75e268a0faa092a3c4
74292 F20101119_AACSPD betts_d_Page_116.jp2
075df7e9893077d17bbf2c996197c6bb
6f384c2b9700b7b550e76e8b248fe703de962b79
376614 F20101119_AACSOO betts_d_Page_096.jp2
d5600eac4fda1d6f102bf88e83a0936e
41220eaea46558d9148ee1f53562f78916e9af80


Permanent Link: http://ufdc.ufl.edu/UFE0012167/00001

Material Information

Title: Transient performance of steam reformers in the context of automotive fuel cell system integration
Physical Description: Mixed Material
Language: English
Creator: Betts, Daniel Augusto 1975- ( Dissertant )
Lear, William E. ( Thesis advisor )
Roan Jr, Vernon P. ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2005
Copyright Date: 2005

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering thesis, Ph.D
Dissertations, Academic -- UF -- Mechanical and Aerospace Engineering

Notes

Abstract: Fuel cells are electrochemical power generators capable of converting a fuel's chemical energy into electrical energy. The most advanced and commonly investigated fuel cells for automotive applications consume hydrogen. Since there are no substantial hydrogen gas reserves in our planet, hydrogen is generally produced from some feedstock. Due to methanol's relatively high energy density, high hydrogen to carbon ratio, and ease to process into a hydrogen-rich gas, it is one of the mostly widely studied fuels for fuel cell engines. The most efficient process by which methanol is converted into a hydrogen-rich gas for fuel cell use is called steam reformation. In this dissertation this process is studied through the use of analytical models with the purpose of understanding the transient outlet gas composition of methanol steam reformers. Attention was given to the outlet carbon monoxide concentration of these reformers, since they have a particularly negative effect on the performance of fuel cells.
Subject: cells, CO, fuel, hydrogen, methanol, reformer
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 128 pages.
General Note: Includes vita.
Thesis: Thesis (Ph.D.)--University of Florida, 2005.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0012167:00001

Permanent Link: http://ufdc.ufl.edu/UFE0012167/00001

Material Information

Title: Transient performance of steam reformers in the context of automotive fuel cell system integration
Physical Description: Mixed Material
Language: English
Creator: Betts, Daniel Augusto 1975- ( Dissertant )
Lear, William E. ( Thesis advisor )
Roan Jr, Vernon P. ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2005
Copyright Date: 2005

Subjects

Subjects / Keywords: Mechanical and Aerospace Engineering thesis, Ph.D
Dissertations, Academic -- UF -- Mechanical and Aerospace Engineering

Notes

Abstract: Fuel cells are electrochemical power generators capable of converting a fuel's chemical energy into electrical energy. The most advanced and commonly investigated fuel cells for automotive applications consume hydrogen. Since there are no substantial hydrogen gas reserves in our planet, hydrogen is generally produced from some feedstock. Due to methanol's relatively high energy density, high hydrogen to carbon ratio, and ease to process into a hydrogen-rich gas, it is one of the mostly widely studied fuels for fuel cell engines. The most efficient process by which methanol is converted into a hydrogen-rich gas for fuel cell use is called steam reformation. In this dissertation this process is studied through the use of analytical models with the purpose of understanding the transient outlet gas composition of methanol steam reformers. Attention was given to the outlet carbon monoxide concentration of these reformers, since they have a particularly negative effect on the performance of fuel cells.
Subject: cells, CO, fuel, hydrogen, methanol, reformer
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 128 pages.
General Note: Includes vita.
Thesis: Thesis (Ph.D.)--University of Florida, 2005.
Bibliography: Includes bibliographical references.
General Note: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0012167:00001


This item has the following downloads:


Full Text












TRANSIENT PERFORMANCE OF STEAM REFORMERS IN THE CONTEXT OF
AUTOMOTIVE FUEL CELL SYSTEM INTEGRATION















By

DANIEL AUGUSTO BETTS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Daniel Augusto Betts


























During my PhD studies a lot of happiness and tragedy occurred in my life. For this reason

I would like to dedicate this work, which represents the culmination of all these things, to

the following individuals: Erica E. Carr-Betts, my wife (thank you for always supporting

me. I love you very much); Carmen A. Carrington Betts, my mother (thank you for being

strong and fighting the good fight. You have immense strength and courage); Claude D.

Betts, my father (thank you for your reason and courage); Lydia C. Carrington, my

grandmother (I am so sorry that we could not save you. Thank you for teaching me that I

can); and to Matilda Eva Carr-Betts, my daughter (thank you for bringing me so much

joy).















ACKNOWLEDGMENTS

I would like to acknowledge the following individuals who directly or indirectly

contributed to this dissertation. First, my wife Erica Eva Carr-Betts was instrumental in

the elaboration of this study. She funded this endeavor, and maintained my sanity during

difficult times. Her support and love had no bounds.

My advisors (Dr. William Lear and Dr. Vernon Roan) were very patient with me.

Their instruction and encouragement fueled the hope that an end was near, even when it

was not. I am eternally grateful to them.

Those who worked at the Ford Fuel Cell Laboratory are also responsible for this

work coming to fruition. Special thanks are given to Dr. Timothy Simons and Ryotaro

Honjo, who were always available to give me a hand.

Dr. Paul Erickson, an alumni of UF and an old friend, was gracious enough to

allow me to conduct experiments in his reformers at UC-Davis. His collaboration made

this work possible.

The US Department of Education and the UF Mechanical and Aerospace

Engineering Department supported my research with a fellowship. Without this

fellowship I would not have been able to obtain this education. I am grateful. I promise to

repay this favor by dedicating part of my efforts to providing educational and

professional opportunities to others.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES .............. .......................... ........... ............. ......... .. vii

LIST OF FIGURES ..................................................... .......... ................ viii

ABSTRACT ........ .............. ............. ...... .......... .......... xi

CHAPTER

1 IN TRODU CTION ................................................. ...... .................

2 B A C K G R O U N D .................. .................................... .......... ........ ......... .. ....

F u el P ro cessin g ...................................... ............................... ................ 7
Steam R reform ing ....................................................... ............... 8
Partial O xidation R eform ing ........................................ ........................... 9
A utotherm al R eform ing................................................ ............ ............... 10
T transportation F uel C ells ............................................................... ............. 11
The Transportation Load Environment ........................... ....................12
CO and Hydrogen Concentration Effects on PEMFC performance.........................16
M eth an o l a s F u el ............. .. ............... ................................................................ 19

3 L ITE R A TU R E R E V IE W ........................................ ............................................23

Kinetics of the Catalytic Methanol Steam Reformer Reactor ................................24
Reform er Perform ance................................................................ ............... 26

4 A N A L Y SIS ............................................... 29

Steam Reform er M odel................................................................ ............... 29
H eat Transfer ............................. .......... ........... .. ......... ...........31
Continuity ............. .................................. .................. 35
C conservation of Specie...................... .. ............. ..................... ............... 38
C hem ical M odel ........................ ................ .... ....... .... ....... 39
Finite D difference Solution ................................................. ............................. 41
FEM LAB Solution.................. ...................................... .......... .. ...... ...... 44
Steam R eform er M odel V alidation........................................ ........... ............45
G general D description ......... ..................................................... .. .... .... .....46



v









UC-Davis Reformer Data............................ .................................. 49
Reformer Design ............. ........ .............. ...............57
D design G en eralization ........................................................................ .................. 57

5 R E S U L T S .............................................................................6 5

Transient Efficiency ......... .. ................ .. .............. ........ .. ...... 65
Case 1 Infinite Reformer ............... ............... ................... 65
Case 2 The Medium Sized Reformer .......................................................67
Case 3 The Short Reformer...................... ... ........... ... .. ............... 68
Carbon M monoxide Concentration ........................................ .......... ............... 70
H hydrogen C oncentration.................................................. ............................... 76

6 C O N C L U SIO N S ..................... .... .......................... ........ ........ ...... ........... 78

Steam R eform er M odeling .............................................................. .....................78
Steady-state R eform er O operation ........................................ .......... ............... 78
U nsteady Reform er Operation........................................................ ............... 78

7 SUGGESTED FUTURE W ORK ........................................ .......................... 80

APPENDIX

A COPY OF FEMLAB SOLUTION REPORT.......................................................81

B PRELIMINARY LOAD FORECASTING STUDY...............................................99

C PHYSICAL DESCRIPTION OF THE UC-DAVIS REFORMER..........................104

G general D description ...... .. .............. .......... .. ........................ .................... 104
Pum ping Subassem bly ........................ .. ............ ... .................... ............... 106
Catalyst Bed Housing Subassemblies...................... ... ......................... 109
Condensing U nit Subassem bly ............. ......................................... ... ........... 111

L IST O F R E FE R E N C E S ........................ .. ......................................... ............. ....... 114

BIOGRAPHICAL SKETCH ............. ............................................................116
















LIST OF TABLES

Table p

3-1. Test param eters for N akagaki tests......................................................... ... ........... 26

4-1. Estimated maximum times for vaporizer volume flow change ...............................49

4-1. Values constants used to develop virtual UC-Davis reformer .................................52

4-2. Steady-state design of 60 kW reformer with an 11.9 mm radius and a catalyst bed
density of 1983 kg/m 3 ....................................... ................... ......... 63

5-1. Case 1: Summary of the transient response of the infinite reformer to changes in
fu el flow rate .........................................................................66

5-2. Case 2: Summary of the transient response of the medium size reformer to
changes in fuel flow rate ................................................ .............................. 68

5-3. Case 3: Summary of the transient efficiency response of the short reformer to
changes in fuel flow rate ................................................ .............................. 69

C-1. U C-D avis reform er subassemblies ............................................... ............... 104
















LIST OF FIGURES


Figure page

1-1. Types of fuel cells and fuel choice used for fuel cell vehicle prototypes.................2

1-2. Energy density of different fuels on a lower heating value (LHV) basis compared
to hydrogen ............... ......... ........................................... .......... ....... 3

2-1. Typical polarization curves of various types of fuel cells.................................14

2-2. Driving load of a 30 ft phosphoric acid fuel cell bus operated at the University of
Florida. The bus weighs is approximately 30,000 lbs.........................................14

2-3. Indirect methanol automotive fuel cell system power delivery schematic.................15

2-4. TBB-2 fuel cell bus methanol steam reformer and steam reformer burner
temperatures during transient bus operation .................................... ............... 15

2-5. University of Florida bus indirect methanol fuel cell system..................................16

2-6. Effect of hydrogen concentration and utilization on PEMFC performance ..............18

2-7. Effect of CO poisoning on PEMFC performance. ......................... ............... 19

2-8. Use of air bleed to recover PEMFC performance. ........................................ 19

4-1. General schematic of modeled steam reformer reactor .................... ..................30

4-2. C control volume e diagram ......... ................. ................. ..................... ............... 30

4-3. Convection-conduction ratio in the UC-Davis reformer fuel entry regions...............33

4-4. Calculated values of dT/dt for the UC-Davis reformer at the entrance region...........37

4-5. Measured temperature distributions of the UC-Davis steam reforming test rig
operating at varying space velocities and varying conversion efficiencies..............37

4-6. Schematic representation chemical reaction model in the control volume ...............40

4-7. Equilibrium products of the steam reforming reaction (CH3OH + H20 aH20 +
bCO2 + cCO + dH2) with respect to temperature......... ...................... .............. 41









4-8. Separation of the reformer volume into annular elements with constant gas
properties in space for the finite difference solution method used.........................42

4-9. Experimental steam reformer developed at UC Davis ............................................50

4-10. Example of CO spike in the reformate of the UC-Davis reformer after an
increase in premix fuel flow (transient condition). Also shown is a decrease of
the steady-state CO concentration in the reformate with increased premix fuel
flo w rate ........................................................................ 5 0

4-11. Reformate CO composition relationship to changes in temperature in the
reform er catalyst bed ........... ..................................... .. ...... ........ .. ... 51

4-12. CO and H2 concentrations in the reformate gas in the UC-Davis reformer.............51

4-13. Estimation of catalyst bed thermal conductivity using UC-Davis reformer data.....53

4-14. Virtual UC-Davis reformer steady-state CO composition with varying premix
fuel flow rates ........... ............................................... ... .................. 54

4-15. Hydrocarbon signal output obtained during UC-Davis reformer run....................54

4-16. Virtual UC-Davis Reformer CO spike from 10 to 15ml/min premix flow rate
ch an g e .............. ............ ......... .................... ............................ 5 5

4-17. Virtual UC-Davis reformer temperature distribution (color plot) and methanol
concentration (contour plot) for premix flow rate equal to 5 ml/min ....................56

4-18. Virtual UC-Davis reformer temperature distribution (color plot) and methanol
concentration (contour plot) for premix flow rate equal to 30 ml/min ..................56

4-19. Reformer Efficiency with varying catalyst effectiveness at varying fuel flow
rate s ................................................................................6 2

5-1. Case 1: Infinite reformer operation when undergoing step variations in fuel flow
rate ...... ......... ................ ....... ........................................................6 6

5-2. Case 1: Reformer efficiency changes due to change of fuel flow rate for the
infinite reform er ................... .... .......................... .. .. ...... .. ............. 67

5-3. Case 2: Transient efficiency of the medium size reformer under transient load
changes ................ .............. .. ........... ......................... 68

5-4. Case 2: Reformer efficiency changes due to change of fuel flow rate for the
m medium size reform er....... ....... ...................................................... .. .... ........69

5-5. Case 3: Transient efficiency of the short reformer under transient load changes ......70









5-6. Case 3: Reformer efficiency changes due to change of fuel flow rate for the short
reform er ............................................................... .... ..... ......... 71

5-7. Infinite reformer transient CO concentrations for various for changes in power
o u tp u t ..............................................................................................7 3

5-8. Infinite reformer transient off-steady-state CO concentrations for various changes
in reform er pow er output............................................................... .....................73

5-9. Medium reformer transient CO concentrations for various changes in power
o u tp u t ..............................................................................................7 4

5-10. Medium reformer off-steady-state CO concentrations for various changes in
reform er pow er output...................................................................... .................. 74

5-11. Short reformer transient CO concentrations for various changes in power output..75

5-12. Short reformer off-steady-state CO concentration for various changes in
reform er pow er output...................................................................... .................. 75

5-13. Transient CO comparison of various size reformers operating between 40 and 50
kW hydrogen pow er output......................................................... .............. 76

5-14. Comparison of transient changes in H2 concentration induced by step changes in
fuel flow corresponding to hydrogen yields from 10 to 60 kW (lower heating
value based) for various size reform ers......................................... ............... 77

B-1. Transient load profile for TBB-2 under the University of Florida Fuel Cell Lab
driving cycle data obtained on 4-2-2004 ....................................... ............... 99

B-2. Power Change Index (PCI) for TBB-2 under UFFL driving cycle data obtained
on 4-2-2004. .................................................................100

B -3. M ulti-layer Perceptron. .................................................. ............................... 101

B-4. The MLP performance as a 10 second PCI predictor....................... ...............103

C-1. General process diagram of the UC-Davis reformer ........................................105

C-2. Premix reservoir and gear pump .................................................................. 107

C -3 V aporizer design ....................................................................... .... .... ......... 107

C -4. Superheater design ........................................... ................................ 11

C-5. The UC-Davis reactor design................................ ......... .. ........... .... 112















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

TRANSIENT PERFORMANCE OF STEAM REFORMERS IN THE CONTEXT OF
AUTOMOTIVE FUEL CELL SYSTEM INTEGRATION

By

Daniel Augusto Betts

December 2005

Chair: William Lear
Cochair: Vernon Roan
Major Department: Mechanical and Aerospace Engineering

Proton exchange membrane fuel cells (PEMFCs) and, to a lesser degree,

phosphoric acid fuel cells (PAFCs) have been widely studied as possible replacements for

transportation internal combustion engines (ICE). These fuel cells consume hydrogen as

fuel, which is electrochemically oxidized through an acid electrolyte. Because of the low

energy density and scarcity of elemental hydrogen, alternatives such as methanol and

natural gas have been investigated as primary fuels for PEMFC and PAFC fuel cell

systems. Of these fuels, methanol is the most easily reformed into a hydrogen-rich gas.

The most efficient way of doing this is through catalytic steam reforming. Therefore a

clear understanding of the performance of steam reformers may lead to better integration

of these devices into fuel cell engines.

In this dissertation the results of studies regarding the transient and steady-state

performance of methanol steam reformers for automotive fuel cell system integration are

provided. To power an automobile, a fuel cell system needs to be capable of changing









power output rapidly, to adapt to changing driving loads. Although most fuel cell

systems currently use batteries to reduce the required response time associated with the

fuel cells and other balance of plant components (including the reformer), the ultimate

goal may be to reduce the reliance of the system on batteries (which add cost, weight, and

complexity to the system). This means that the fuel cell stack and the reformer must each

be capable of changing power outputs quickly and efficiently. Fuel cell stack efficiency is

intimately related to reformate gas composition (especially CO concentration). Since the

reformer is upstream of the fuel cell stack, it has great influence on the overall power

output and efficiency response of the fuel cell system.

The results of this study are based on data obtained from an experimental reformer

and numerical models. Non-dimensionalization of the governing equations derived for

reformer model development resulted in identification of potential reformer similarity

variables. Based on these reformer sizing and scaling theories were developed. Of

particular importance was the further demonstration of the insufficiency of space velocity

and aspect ratio as sole reformer similarity variables.

Step changes in fuel flow into the experimental reformer produced transient CO

concentration spikes. These spikes have also been identified in reformer literature.

Through the use of the transient reformer model, potential physical mechanisms that

cause these CO concentration spikes were identified.

Studies in conversion efficiency and transient and steady-state reformate hydrogen

concentration were also carried out.














CHAPTER 1
INTRODUCTION

Interest in fuel cell technology for transportation has led to an ever-increasing

number of prototype fuel cell vehicle demonstrations in recent years. For example, 210

new prototype fuel cell vehicles were built in 2004. This is more than the cumulative

number of fuel cell prototypes developed between 1959 and 2002 [1]. Enthusiasm for

fuel cells centers on their high energy-conversion efficiencies, the cleanliness of its

primary fuel (hydrogen), and the cleanliness of its operational by-products (water,

electricity and heat). Yet commercialization of fuel cell vehicles faces major hurdles

including design of low-cost and robust systems, proper water-management design for

transient operation at various power levels and conditions, improvement of stack life,

improvement of hydrogen storage technology, and improvement of hydrogen-production

technology.

Generally, hydrogen supplied to the fuel cell stack comes from one of two sources:

on-board hydrogen storage, or a fuel processor. Currently, most fuel cell vehicle

manufacturers are developing vehicles that use hydrogen as their primary fuel

(Figure 1-1). On-board hydrogen storage is intended to simplify fuel cell engines and

reduce costs. However, even with pure hydrogen, fuel cell systems are far from achieving

operational goals, especially when the power output of the fuel cell system is cycled

through time. In addition, hydrogen storage technology is far from achieving the energy

density goals needed for widespread vehicle commercialization. Figure 1-2 compares the

energy densities of common hydrogen-rich fuels. Even with the most advanced hydrogen










energy-storage mechanisms, the energy density of hydrogen is far less than that of other

fuels. Currently, hydrogen-fueled fuel cell vehicles have achieved a maximum driving

range of approximately 300 km [1]. Major research efforts are now aiming to improve

hydrogen storage capacity. These efforts include using carbon nanotubes as a hydrogen

storage medium, and developing safe hydrogen tanks capable of operating at pressures of

over 700 bars.

NetaE
rPAC A HFH C. A F,11-1







PEMFC \
CH2
Figure 1-1. Types of fuel cells and fuel choice used for fuel cell vehicle prototypes. This
figure was published with the permission of its original authors [1].

Introduction of reformation technology into fuel cell vehicles has been proposed as

a possible inexpensive way to solve the energy density problems associated with

hydrogen [2]. Industry analysts have pointed out that alternative liquid fuels could make

use of parts of the current global fuel delivery and storage infrastructure, potentially

reducing the cost of the fuels. In addition, alternative liquid- fuels could still continue to

be used in internal combustion engine vehicles without major modifications. Of the fuels

typically considered, methanol seems the most attractive because it is easily reformed at

relatively low temperatures and can be produced from a wide range of feedstock. In terms

of energy density, methanol has approximately half the energy density of gasoline, and

methanol is approximately 1.45 times more energy dense than hydrogen stored in

advanced hydride beds (Figure 1-2). It is important to point out that the process of steam









and autothermal reformation includes the dissociation of water to produce hydrogen,

therefore yielding greater hydrogen per carbon concentration than the simple methanol

dissociation reaction.


40
j35 -R2 Carbon
0] Hydrogen
i 30
5 25

S20

15
( 10 i











Figure 1-2. Energy density of different fuels on a lower heating value (LHV) basis
compared to hydrogen [3].

Using methanol as primary fuel for PEMFC vehicles is technically challenging.

Methanol-derived reformate contains hydrogen diluted mostly in carbon dioxide, water,

carbon monoxide, and methanol. Among the problems is that the dilution of hydrogen

produces a drop in cell efficiency. This drop in cell performance can be measured as a

drop in cell voltage (from V, to V2) that corresponds to a drop in cell hydrogen partial

pressure from (PI to P2) as shown in Equation 1-1.










V2 -V = lnKC (1-1)


The value of the constant C has been measured to range from 0.03 to 0.06 for

PEMFCs. This drop in voltage leads to a drop in fuel cell efficiency and operational

power density range. A positive aspect of reformation is that water vapor is among the

diluents in the anode fuel stream. This reduces the need for anode gas humidification,

which in hydrogen-fueled fuel cell systems can be very complicated.

The reformate gas also contains carbon monoxide, which has a poisoning effect on

low-temperature fuel cells1. The PEMFCs have been poisoned with CO concentrations

from 30 ppm. The equilibrium concentration of CO for the steam reforming reaction of

methanol at 2250C and 300C (the typical range of temperature for catalytic steam

reforming) are 0.8 and 2.3 mole percent, which is beyond the limits of most current

PEMFC technology tolerance. Various reformate gas "clean-up" schemes aim to reduce

the concentration of CO in the anode fuel stream. Notable among these are preferential

oxidizers and specialized membranes (generally palladium) capable of filtering CO and

other gases from the reformate stream.

Coupled with these technical issues is the need for controlling the fuel cell and the

reformer when changes in load are required, in such a way that system efficiency drops

are minimized and response speed is maximized. Figure 1-3 shows the recorded

efficiencies of one of the 30ft Georgetown fuel cell buses (TBB-2) during a typical drive





1 CO poisoning in fuel cells is a reversible phenomenon. Small concentrations of oxygen can be used to
reverse the CO poisoning effects. Other techniques include increasing cell current to promote the
electrochemical water-gas shift reaction.









cycle.1 These buses use methanol as their fuel and use a steam reformer to produce

hydrogen that is consumed by a phosphoric acid fuel cell (PAFC). The transient changes

in system efficiency shown are caused by variations in anode hydrogen stoichiometric

flow, changes in reformate hydrogen gas concentration, and changes in reformate carbon

monoxide concentration. Temperature variations in the fuel cell stack may also contribute

to these efficiency changes.

The goal of this study is to provide a broader understanding of the role of the

reformer in the dynamic performance of automotive load-following fuel cell systems, and

provide a basis for reformer design and modeling. Analyses in this study are derived from

a mixture of physical modeling results and experimental observations from which general

trends in reformer steady-state and transient conversion, hydrogen concentration, and

carbon monoxide concentration were derived.
























1 The TBB-2 bus is a 30 ft phosphoric acid fuel cell bus housed at the University of Florida FORD fuel cell
laboratory. This bus uses on-board methanol steam reformation for hydrogen production.












0.7


0.6


0.5


0.4


0.3-- M -


0.2
Max PoMer = 50kW
Rate of Chemical Energy in the Premix flow was calc lated using the lower heating value of MeOH in liquid form.
0.1 -
35 45 55 65 75 85 95 105 115
Run Time (min)

FC System Efficiency Ratio (FC Power/Max Power) Stack Efficiency

Figure 1-3. Measured transient efficiency of the 30 ft Georgetown Fuel Cell Bus (TBB-2)















CHAPTER 2
BACKGROUND

In this chapter, several basic concepts essential to the general understanding of the

analysis and results of this dissertation are provided. The chapter describes key fuel

processing techniques and compares these techniques to each other. In addition, a short

description of low temperature automotive fuel cells is provided. Also, general systems

integration issues of indirect methanol fuel cell systems1 are presented. In this section,

data obtained from the Georgetown fuel cell test bed buses is used to show transient

interactions between the fuel processor and the fuel cell in an actual system. Finally, a

general discussion regarding methanol as a potential automotive fuel cell fuel is given.

Fuel Processing

Because of the difficulties of direct electrochemical oxidation of hydrocarbons,

hydrocarbon fuels are normally processed (or reformed) to give a hydrogen-rich fuel

mixture. Reforming plays an especially important role in low-temperature fuel cells,

namely proton exchange membrane fuel cells (PEMFCs) and phosphoric acid fuel cells

(PAFCs), since they are unable to oxidize any other fuel but hydrogen (H2)2. There are

three basic reforming methods: steam reforming, partial oxidation reforming, and



1 Indirect methanol fuel cell systems refers to the use of methanol as the system fuel. The methanol in this
systems use a reformer to produce hydrogen, which ultimately is the fuel cell fuel. The term indirect is used
to differentiate these systems from direct methanol fuel cell systems, which are capable of oxidizing
methanol in the fuel cell without the use of a reformer.

2 Direct methanol fuel cells (DMFCs) commonly use similar electrolyte material, namely Nafion, as
PEMFCs. In this text the term PEMFC refers to hydrogen proton exchange membrane fuel cells and
excludes DMFCs.









autothermal reforming. This study concentrates on steam reforming since, of the

aforementioned methods, it is the most efficient form of hydrogen production from

hydrocarbon fuels.

Steam Reforming

Steam reforming is the reacting of a hydrocarbon with steam to produce a

hydrogen-rich gas (reformate). Using methanol as an example feedstock, steam reforming

can theoretically produce a maximum three moles of diatomic hydrogen gas for every

mole of carbon involved in the reaction (3:1). Steam reforming processes can occur with

or without a catalyst.1 However, most low-temperature fuel cell systems that use steam

reformation make use of catalysts in the reformer. Whereas non-catalytic steam

reforming requires temperatures in excess of 1100C for simple hydrocarbons, catalytic

steam reforming can occur at temperatures ranging from 200C to 700C for simple

hydrocarbons. Nonetheless, non-catalytic steam reforming is still under investigation as a

possible method of reducing the thermal mass and volume of reformers.

Equation 2-1 shows the general, ideal steam reforming reaction for alcohols.


CHmO, + aH2O m +a H2 + (n + a 1)C02 + (21 n -a)CO
\2 ) (2-1)

1, m, and n represent the number of moles of carbon, hydrogen and oxygen,

respectively, in the hydrocarbon being reformed. The variable a represents the number of

moles of water per mole of hydrocarbon reacted.

Greater concentrations of water (a) in the reactants reduce the concentration of CO,

and increase the quantity of CO2 in the products. Moreover, the amount of hydrogen is


1 Non-catalytic steam reformation requires very high temperatures.









increased with greater steam. The maximum concentration of water that can be reacted in

the general steam reforming reaction is

amax = 21 n (2-2)

In practice, the catalytic steam reforming reaction is dependent on three-phase

chemical kinetics with mechanisms that are complex and currently not well understood.

Modem steam reforming catalysts can be designed to suppress the formation of CO and

promote the formation of CO2. The most commonly used methanol steam reformer

catalysts in published literature are composed of copper-oxide and zinc-oxide

(CuO/ZnO). As a general rule, the steam reformation of methanol across a CuO/ZnO

catalyst is executed with 1 to 1.5 moles of water per mole of methanol. Increasing the

H20 concentration to 1.5 has been found to reduce the CO content in the products while

increasing the hydrogen yield when the reformation process is maintained in the range of

250oC-3000C. Although higher water concentrations in the fuel increases hydrogen yield

and decreases reformate CO concentrations, these benefits may be counterbalanced by

the energy penalty associated with water vaporization and superheating.

Partial Oxidation Reforming

Partial oxidation reforming is a process in which the fuel undergoes combustion in

a fuel-rich environment. The elevated temperatures generated from this combustion

breaks down the hydrocarbon fuel into a hydrogen rich gas. Assuming, ideally, that the

reaction products are only hydrogen, carbon monoxide and carbon dioxide, the general

partial oxidation reaction for an alcohol is given in Equation 2-3.


CHO, + aO -e H, +(n+2a-e)CO +(21- n 2a+e)CO +eHO

(2-3)









To obtain the maximum possible amount of hydrogen in a partial oxidation

reaction, the amount of water (e) formed must be negligible (e 0). Taking methanol as

an example, the maximum, ideal, diatomic hydrogen to carbon ratio for the partial

oxidation reaction is 2:1. This is 2/3 of the maximum possible obtainable hydrogen to

carbon ratio from steam reforming of methanol. It is important to note that for most

applications the use of pure oxygen for partial oxidation is impractical. Thus nitrogen will

be present in the products of the reaction, which further dilutes the reformate hydrogen

concentration.

Some advantages of partial oxidation over steam reforming are that water is not

required for the process, the equipment is more compact, it allows rapid start-up, the

same reactor can reform various types of fuel, and the high-temperature reactor can

tolerate many impurities (including sulfur). On the other hand, care must be taken to

prevent carbon deposition. Also, since the process is exothermic, relatively high local

temperatures are produced, causing possible materials problems. For integration with

lower temperature fuel cells, a gas clean-up step is necessary to reduce the inherently

large CO content.

Autothermal Reforming

Autothermal reforming combines aspects of steam reforming and partial oxidation

reforming. A catalyst bed, typically nickel, and steam promote the reforming reaction.

The necessary heat is generated through combustion of a portion of the reformer fuel.

The general, idealized autothermal reforming reaction is given by Equation 2-4.


CHO, + aO + bH0 -> + b H2 +(n+2a+b -)C0+(21-n-2a-b)CO

(2-4)









In autothermal reforming, the maximum amount of water (b) that can be reacted is

given by Equation 2-5.

bma = 21- n -2a (2-5)

Compared to the ideal steam reforming reaction, the maximum amount of water

that can be reformed is reduced by the presence of oxygen, thus reducing the overall

hydrogen yield. On the other hand, the addition of water to the reaction allows for higher

hydrogen yields and lower carbon monoxide concentrations in the products.

Autothermal reformers have had a large amount of interest because of their

capacity to reconcile between the benefits and detriments of using either catalytic steam

reforming or partial oxidation reforming.

Transportation Fuel Cells

For transportation applications and small-scale residential power production,

PEMFCs and PAFCs have been most widely considered and explored. These fuel cells

operate similarly, oxidizing hydrogen through an electrolyte and a current collector as

shown:

Anode: H2 4e-+2H+

Cathode: /2 02 + 4e-+2H+ H20

However, PEMFCs operate at a lower temperature than PAFCs (25C-80C as

opposed to 1500C to 2000C). In addition, PEMFCs, due to their high efficiency over a

wide range of current densities, typically have higher energy density than PAFCs. Mainly

for this reason, PEMFCs are widely considered the preferred automotive fuel cell. This

distinction does not mean that PEMFCs are trouble-free. Major engineering challenges

are yet to be resolved in order to commercialize automotive PEMFC stacks. Foremost,









these fuel cells must dramatically improve their operational life; they must achieve

consistency in performance under mass production, which has not yet been demonstrated;

they must be designed to have effective humidity control and positive water balance

during all operating regimes; and their cost must be dramatically reduced. In terms of life

and humidity control, PAFCs have been demonstrated to far surpass PEMFCs. However,

PAFCs are much more expensive to produce than PEMFCs and their cost is harder to

reduce to target levels, even through mass production.

Figure 2-1 shows typical polarization curves of various types of fuel cells including

molten carbonate fuel cells (MFCs), solid oxide fuel cells (SOFCs), and alkaline fuel cell

(AFCs). The reader might find it curious that AFCs are not considered in this study even

though they exhibit much higher voltages than all other fuel cells. The reason for this is

that AFCs do not tolerate air (due to the presence of C02). This limits their use to space

applications and as electrolyzers.

The Transportation Load Environment

The power load associated with automobiles and buses is high and extremely

transient. Vehicles are driven with a rapid succession of speed adjustments, especially in

city traffic. Any fuel cell engine that is going to be incorporated into a vehicle must be

able to change power delivery over a wide range of conditions fast and efficiently. Figure

2-2 shows driving loads obtained from a 30 ft fuel cell transit bus operated at the

University of Florida (UF) on a low speed circuit around the city of Gainesville, Florida.

The transient response of fuel cell systems is limited by their capacity to deliver fuel

(hydrogen) and oxidant (oxygen) to the fuel cell stack. In a reformer based fuel cell

system, as shown in Figure 2-3, the fuel processing subsystem must be capable of

changing hydrogen production levels at the same rate as current draw changes occur in









the fuel cell stack. In turn, fuel cell voltage at a certain current draw depends greatly on

H2 and CO partial pressures in the reformate gas.

The process of steam reforming is complicated by the endothermicity associated

with hydrogen production. The composition of the reformate gas is largely dependent on

the temperature at which the reforming reaction takes place. Yet, in steam reformers the

heat required to maintain the reforming reaction is provided by an external source and

thus is limited by heat transfer considerations. This results in varying reforming reaction

temperatures during steady-state and transient operation. As an example, the University

of Florida fuel cell bus (TBB-2) reformer temperatures during transient conditions are

shown in Figure 2-4.1 In TBB-2 the reformer catalyst bed is heated via an external burner

that consumes stack anode flue gas (Figure 2-5). The anode flue gas contains varying

quantities of hydrogen, depending on stack current draw and anode stoichiometry. Anode

stoichiometry is dependent upon fuel flow rate and reformate quality. Current-draw

depends on the load environment. For TBB-2 the stack power draw and the reformer

hydrogen production are coupled. University of Florida researchers have found that this

interrelation between the stack current draw and reforming heating has led to instances of

reformer overheating and under-heating, both of which have led to drops in overall

system performance [2].








1 Reformer top, center, and bottom refers to the position of thermocouples in different regions of the
catalyst bed. The catalyst bed entrance is located close to the reformer bottom thermocouple. The center
thermocouple is located close to the midpoint of the reformer and the top thermocouple is located close to
the exit of the reformer. The reformer burner thermocouple is located close to the reformer burner flame.



























0.6+


0 200
0 200


400
400


1000


Current Density (mA/cm2)
Figure 2-1. Typical polarization curves of various types of fuel cells. This figure was
published with the permission of its original authors [4].


10


0


-10


600


Run Time (s)


Figure 2-2. Driving load of a 30 ft phosphoric acid fuel cell bus operated at the
University of Florida. The bus weighs is approximately 30,000 lbs.


AFC


MCFC




PEMFC

SPAFC
SO FC


4-2-2004 test 1


Transient Load Profile for TBB-2 under the UFFL driving cycle


0.9--



0.8--





























Figure 2-3. Indirect methanol automotive fuel cell system power delivery schematic


650
Idle
600


550 -


500 -


450


400
E
1- 350


300


250


200
00:00.0


Reformer


07:12.0 14:24.0 21:36.0 28:48.0 36:00.0 43
Time (minutes)
Top Reformer Center Reformer Bottom


:12.0 50:24.0 57:36.0


- Reformer Burner


Parasitic and
Aux Loads


if


MetanolWater Steam Reformer Phoph Acd
Premix Tank Fuel Cell Stack


Figure 2-4. TBB-2 fuel cell bus methanol steam reformer and steam reformer burner
temperatures during transient bus operation













Combustion
Products




Air
Reformer
Burner


Anode Flue Gas
A
Reformate By ass

Excess Air
Steam and Water
Reformer


Vaporizer

S_______ Heat
Neat Water Exchanger
Methanol Methanol
Premix Air
xAir Start-up
S_ Burner


ombustion
Products


Figure 2-5. University of Florida bus indirect methanol fuel cell system

CO and Hydrogen Concentration Effects on PEMFC performance

Small increases in reformate CO concentration lead to drastic reductions in fuel cell


performance in low-temperature fuel cells that use platinum catalyst to promote the cell


anodic reaction. In these cells, CO competes directly with hydrogen for active sites in the


anode due to its high affinity for platinum at low-temperatures. Springer et al. propose


that the reactions depicted in Equation 2-6, Equation 2-7, Equation 2-8 and Equation 2-9


represent well the competition for active sites at a PEMFC anode [5].


CO + M bfc >(M- CO) (2-6)


H2 +M < bh >(M CO)


(M H)-hH+ + e + M


(2-7)


(2-8)


(2-9)


H20 + (M CO) k M + CO2 + 2H+ + 2e


Fuel Cell Stack


Heat Exchange
Plate


1 )









Equations 2-6 and 2-7 represent the competition between H2 and CO for active

sites, where kfc and kjh represent the forward reaction-rate constants. The electro-

oxidation of hydrogen and CO are represented in Equations 2-8 and 2-9, with their

respective rate constants (keh and kec). PEMFC models and experimental results have

demonstrated that drops in anode hydrogen concentration produce very mild decreases in

cell performance (Figure 2-6). However, high CO concentrations have been shown to

drastically drop cell performance, especially when diluted in reformate bearing low

hydrogen concentrations (Figure 2-7).

In order to further reduce the CO concentrations produced from steam reformers,

many fuel processing systems use preferential oxidizers and anode air bleeds (the

introduction of small quantities of air into the fuel cell anode chamber). Preferential

oxidizers require the addition of air into a catalytic bed that promotes the oxidation of

CO. Typically a small percentage of the hydrogen in the reformate is also oxidized.

Currently, control of air delivery systems for preferential oxidizers is complicated due to

the high sensitivity of the chemical rate equations to gas and catalyst temperature and CO

concentration. In addition, the amount of air associated with preferential oxidation

operation is very low and difficult to meter. In the overall reforming process high

upstream variations in CO concentrations may lead to increased concentrations of CO

leaving preferential oxidizers, thus increasing the probability of anode catalyst poisoning.

This is especially problematic since CO spikes occur during periods of increasing power

demand. This phenomenon will be further explained in subsequent chapters.

Various solutions have been proposed for solving the problem of CO poisoning of

the fuel cell anode catalyst. Of these the most commonly reported has been the









introduction of small quantities of air into the cell anode during periods of decreased cell

performance due to CO poisoning. This produces catalytic oxidation of the CO in the

presence of anode platinum. Nonetheless, the air bleed also has a tendency to oxidize a

portion of the hydrogen introduced into the fuel cell. This results in a further reduction in

cell efficiency. Figure 2-8 shows how the effect of air bleed introduction in the anode

reduces the poisoning effects of CO. Also, it is difficult to distinguish between cell

performance deterioration due to CO poisoning and other operating conditions (humidity

variations, temperature variations, pressure variations, cell degradation, etc).

Based on this, a reformer that produces minimal increases in CO level during

increased fuel flow events is more desirable than one that produces high hydrogen

concentrations (within typical reformate hydrogen and carbon monoxide concentrations).



"\ 1 atm
$, H ,..J5 N, .25 CO)


Ncal. H
0-.7 tca' H


!>



S.IJ
Curr ti L'zJ-riin.'






Figure 2-6. Effect of hydrogen concentration and utilization on PEMFC performance.

This figure was published with the permission of its original authors [5].
II "..- .

U -3 .




Curry'mr L'c'ciily P.' i;"
Figure 2-6. Effect of hydrogen concentration and utilization on PEMFC performance.
This figure was published with the permission of its original authors [5].







19



12(10 0
HI
I1000 A H 25 ppmrnCO
m H .1* 5J0 ppmI C
: K : H? *+ 1,0 ppm' CO
m800 HI + 25':1 ppmv Cu

I L m ,a ,
S60 *


400 w m a
O > m
(,) L- 4 ,
200

0 1
0 200 40I0 COO 800 1000
CURRENT DENSITY (rnA'cmn2
Figure 2-7. Effect of CO poisoning on PEMFC performance. This figure was published
with the permission of its original authors [5].






..0










Ai,: .* *P>** :.. slaich at
: 0,6 V j *,
-j I


SMethaol ase n C
o22 .-.- tfrmabwit CO% Air


0 0+2 IA 0. 14
I mi -r-r I -11 ,1,1i. N n \


Figure 2-8. Use of air bleed to recover PEMFC performance. This figure was published
with the permission of its original authors [5].

Methanol as Fuel

Since most of the results of this research are based on the study of methanol


reformation, a brief discussion of methanol as a potential fuel cell fuel is relevant.









Of the alternative fuels which are being considered for vehicles, many view

hydrogen as the ultimate long-term alternative fuel. However, while hydrogen has some

very desirable attributes, there are many extremely difficult issues to resolve, which

might make the time scale for a "hydrogen economy" far longer than the time scale of

readily available and affordable petroleum. Fuels which can be stored as liquids offer

major advantages for transportation applications. Of the alternative liquid fuels, in many

ways the most promising is methanol. This is because methanol reforming technology is

well developed as compared to reforming technology for other common automotive fuels.

Also methanol can be produced from natural gas, biomass, electricity, coal, and any other

hydrocarbon. These result in a fuel that can potentially be relatively inexpensive in most

regions of the globe. In addition, methanol can be used in internal combustion engines,

gas turbine engines and in fuel cell engines.

Most of the methanol produced in the U.S. is derived from natural gas. Twelve

major methanol production plants exist in the U.S., which in 2001 produced over 1.5

billion gallons of methanol (94.2 trillion BTU or equivalent to 753.6 million gallons of

gasoline). The overall U.S. methanol consumption in 2001 was, in energy terms,

equivalent to -1,507 million gallons of gasoline, half of which was imported [6]. The

price of delivered methanol has averaged -$0.50 per gallon (-$8/million BTU) since

1991. If the price of natural gas did not figure into the cost of methanol, the methanol

cost would be -$1.50/million BTU.

Currently, most of the U.S. produced hydrogen is derived from natural gas steam

reforming. The U.S. consumes around 9 million tons of hydrogen (925.8 trillion BTU or









equivalent to -7,407 million gallons of gasoline) per year, of which approximately 17%

is sold to chemical plants and refineries [7].

Current hydrogen costs at large scale chemical plants, where hydrogen is consumed

on site, are generally around $5.3 I/million BTU. If transported, the price of hydrogen is

much higher. The typical price of delivered liquid hydrogen oscillates between $1.00 and

$1.40 per pound ($20/million BTU to $27/million BTU). Therefore hydrogen prices are

comparable to methanol prices only if hydrogen is not pressurized, liquefied, or

transported.

The cost of producing hydrogen using steam reforming of natural gas minus the

cost of the natural gas, for a large plant, would be -$4.15/million BTU. This price does

not include transportation, storage, compression or liquefaction of the gas.

While for methanol, storage and transportation costs are not essential components

of the overall fuel cost, this is not the case with hydrogen. It is expected that, due to its

lower energy density, methanol transportation costs per unit energy basis will be twice

the cost for transporting gasoline, in energy terms (-$6.92 per million BTU) [7].

Hydrogen transportation cost, in contrast, may vary drastically depending on the distance

transported and the method of transportation. Taking current gasoline transportation costs

as a baseline, it could be assumed, as an initial estimate, that because of its lower energy

density, liquid-hydrogen transportation would cost at least 5 times more than gasoline

transportation if similar transportation forms were used (mainly pipeline and truck

transport). Thus, hydrogen transportation costs could be expected to be around $18.94

per million BTU [7].









Hydrogen distribution costs could also be estimated by assuming that the eventual

hydrogen distribution infrastructure will be similar to that of natural gas. This assumption

is legitimate since both fuels are gaseous. The average distribution cost for natural gas,

from wellhead to residential consumer1 in the U.S., from 1991 to 2003, was $4.80/million

BTU. Thus, by applying an energy density penalty it could be estimated that the average

hydrogen distribution price in the U.S. will be -$18.5 I/million BTU (similar to the

previous estimation). Note however that these estimates do not include additional cost

penalties associated with hydrogen transport such as boil-off, gas leakage, and special

materials needs.

Based on this basic analysis, it is fair to assume that methanol could be an

important fuel for transportation and fuel cells since it has the potential to be much

cheaper and more easily distributed than hydrogen. A study by the University of Florida

found that based on future fuel cell vehicle fleet estimates for 2020, and according to

various economic scenarios, methanol derived hydrogen from coal would be the least

expensive automotive fuel for fuel cell vehicles in the US if environmental concerns are

not taken into account [7].















1 Residential natural gas prices were used to estimate hydrogen distribution prices because it most closely
resembles the distribution requirements of the automotive sector.














CHAPTER 3
LITERATURE REVIEW

This literature survey has the purpose of framing the work and results of this

dissertation as part of the latest research and scientific debates regarding the design and

integration of methanol steam reformers into fuel cell systems.

Generally, research conducted on reformers can be catalogued into two different

approaches, the chemical and the physical. Chemical research has concentrated in trying

to understand the chemical mechanisms and rate of reaction relations of the steam

reforming reaction. A large volume of papers and presentations have been published in

the last couple of years regarding this subject. Common among these is a general concern

among researchers that almost every research case (even when identical catalysts are

used) has lead to different rate of reaction equations, although in the same mathematical

form.

The physical approach to reformer research has dealt with understanding the

complete reformer assembly and its behavior. The volume of work in this area is limited.

In addition, most of the work has concentrated on the steady-state case. Dynamic models

and studies on fuel cell systems in published literature typically rely on user-defined

linear exponential decay functions for each fuel cell system component, including the

reformer. These investigations concentrate on the question of fuel cell systems control.

Typical of this approach is the study conducted by El-Sharkh et al. 2004 [8].









Kinetics of the Catalytic Methanol Steam Reformer Reactor

Although the actual kinetics of the methanol steam reforming reaction were outside

the scope of this study, a general review of the work in this field is herein presented. Of

note is the diversity of rate equations that have been presented in literature. This

demonstrates the complexity of the reaction and the variability that can be obtained based

on catalyst selection, catalyst preparation, and reaction conditions.

Choi et al., 2005, proposed that the methanol steam reformer products can be

obtained by with the rate equations for 3 main reactions in a Cu/ZnO/A1203 catalyst

manufactured by Sud-Chemie. These were [9],

CH30H CO + 2H2
CH30H + H20 CO2 + 3H2
CO + H20 < CO2 +H,

The study conducted by Choi greatly simplified the widely cited 12 reaction rate

model previously given by Peppley et al. [10]. Peppley and Choi point out in their studies

the large degree of disagreement in literature regarding reaction mechanisms and the

debate on identifying the physical location of the steam reforming catalyst active site.

The results of their experiments were a series of rate equation for each of the proposed

methanol steam reformer reactions. Based on these reactions a methanol steam reformer,

a water gas shifter and a PROX reactor model were developed. The results of their

models show an increase in conversion with greater temperature and catalyst weight in

the reactor, however this also increased CO. The study conducted by Choi did not model

heat transfer in the reactor bed or the transient behavior of the reformer.

Lee et al., 2004, [11] recognized the importance of steam reforming kinetics in

sizing methanol steam reformers. They also recognized the large diversity of rate









equations found in literature. Lee, in his paper includes a list of 9 different catalytic

methanol steam reforming rate of reaction equations found in recent literature. Many of

these reaction rate equations are contradicting even when the same catalyst is being used.

Lee observes that higher water concentration in the steam reforming reaction promotes

effluent CO concentrations closer to equilibrium. The discrepancy between experimental

CO concentrations and equilibrium concentrations were between -lmol% to <0. mol%

for a water feed concentration between 15mol% to 30mol%. Finally Lee, in his study,

also proposed a new rate of reaction of methanol for the catalyst used.

The correlation proposed by Nakagaki et al. [12] was used in the models produced

in this study. Nakagaki obtained his correlations through a series of experiments that

closely resembled the test conditions in which data was obtained for this study. The study

conducted by Nakagaki used a temperature controlled packed-bed reformer operated at

various fuel flow rates. In obtaining his correlation, Nakagaki conducted constant catalyst

temperature studies and constant wall temperature studies (the latter allowed for

temperature gradients to exist in the reformer catalyst bed). Finally his correlation was

validated through actual reformer operation and modeling. This type of experimentation

varies from others in literature in that the effects of fuel flow (transport phenomena) and

varying reactant concentrations were considered. Also considered are the effects of

temperature gradients in the catalyst. The following is a general description of the study

conducted by Nakagaki, et al.

Nakagaki et al. obtained their correlation through experiments conducted in a

packed-bed experimental reformer. The catalyst used was Cu-Zn in alumina support

commercially available from Nissan (G66-B). The catalysts were 1/8in by 1/8 in









cylindrical tablets. Various concentrations of methanol diluted in simulated reformate gas

assuming conversions equivalent to 0, 10, 20, 50, and 75 percent were flowed through the

packed bed reformer. The reformer temperature was maintained constant and controlled

through the use of sheath heaters. This first test was used to determine the methanol rate

of reaction in a constant temperature reaction.

A second test controlled the temperature of the reforming reaction by maintaining a

constant wall temperature. In the second test the temperature distribution and the

conversion were recorded. The exiting reformate gas composition was recorded using a

TCD gas chromatograph. Based on second test results, the Nakagaki correlation was

refined.

Table 3-1 shows a summary of the tests conducted by Nakagaki. Nakagaki also

conducted modeling studies through a combined energy, momentum (Darcian equation),

and conservation of species equations to further validate his correlation. However, since

the application of the study was for employment in gas turbine engines, the study of CO

concentration in the reformate outlet stream was neglected.

Table 3-1. Test parameters for Nakagaki tests
Parameter Test 1 Test 2
Pressure (atm) 1, 5, 10, 15 1, 10
Catalyst Temperature (K) 473, 498, 513, 538 --
Wall Temperature (K) -- 518, 548
Mass Flux (kg/(sm2)) 0.25 0.25, 0.36
Simulated conversion 0, 10, 20, 50, 75 0, 80
percentage

Reformer Performance

Various studies have been published on reformer performance and design but the

majority of these have concentrated on reformer steady-state performance and catalyst









dynamic behavior (rates of reaction correlations). The work of Betts et al. [13] provided a

mechanistic model for observed fluctuations in reformer CO output. The identification of

CO spikes as being caused primarily by heat transfer effects in the reformer was

discussed in that study. A study by Horng [14] showed for an autothermal reformer the

existence of CO spikes once fuel was introduced into the reformer. However, the study

concentrated in the start-up process of the reformer and did not place much emphasis on

the CO phenomenon.

The importance of heat transfer and temperature on the performance of methanol

steam reformers was studied by Hohlein et al. [15]. Hohlein concludes that reformer

heating is a very strong determinant of reformate CO concentration.

Perhaps the most comprehensive experimental study on reformer performance was

conducted by Davieau [16]. In this study, reformer performance was measured in terms

of methanol conversion for two reformers of differing diameters. The initial interest of

Davieau was to study the degree in which space velocity and aspect ratio determines

reformer similarity. The reformers were operated at various space velocities. The results

demonstrate that space velocity and aspect ratio were not a good determinant of reformer

performance. That is, two reformers operating with the same space velocity, same

catalyst loading, same wall temperature and same aspect ratio had drastically different

performances. Yet, Davieau points out that space velocity is the most cited reformer

similarity relation in literature. Additionally, Davieau notes that temperature distribution

within the reformer is a critical performance parameter. Furthermore, it is noted that

diameter influences heat transfer and therefore has an effect on performance that is






28


beyond catalyst loading. The results obtained by Davieau were based on experimental

observations. His study did not focus on the transient behavior of the reformer.














CHAPTER 4
ANALYSIS

In accordance with the objectives of this research, a general methanol steam

reforming model was developed. The details of the derivations and major assumptions

applied to the development of this model are included in this chapter. In addition, this

chapter includes a general comparison of the steam reformer model results and

experiments conducted at UC-Davis using an experimental reformer test rig developed at

by Dr. Paul Erickson (Appendix C). Part of the analysis includes the development of a

generalized design and comparison criteria for steam reformers through similitude

parameters. Finally, in this chapter transient operation of the modeled steam reformer is

presented.

Steam Reformer Model

Packed bed catalytic steam reformers are generally composed of a reaction

chamber filled with catalyst pellets (Figure 4-1). The methanol water mix along with the

product gases flows around and through the catalyst pellets. The chemical reactions occur

at the surface and within the catalyst pellets. In the development of this model, it was

recognized that the main driver of reformate chemistry is temperature.

A control volume approach was used to obtain differential conservation of species

and conservation of energy relationships. In choosing the scale of the finite differential

control volume, it was assumed that the catalyst pellets and the gases exist in the control

volume as a homogenous mixture. That implies that the control volume used, although

small compared to the overall reformer, has length scales larger than that of the catalyst.









It was also assumed that the gas and the catalyst exist locally at the same temperature.

This comes as a result of the assumption that the catalyst and the gases are regarded as

part of a continuum. This also implies that the surface area of the catalyst inside the

control volume is very large, resulting in the gas being in thermal equilibrium with the

catalyst. In addition, the flow was assumed to occur in the axial direction (i.e. the radial

component of the velocity was assumed negligible). Figure 4-2 shows the control volume

with the major heat inputs and outputs at the boundaries.


Steam Reformer Walls


Methanol + Water .. ...n Chan .er Reformate



Catalyst Pellets


Figure 4-1. General schematic of modeled steam reformer reactor


qr+dr



C.V.

Sqx+dx

hg, ddr
"1,x I g,x+dn


Figure 4-2. Control volume diagram









Heat Transfer

Based on the chosen control volume and the aforementioned assumptions, the heat

transfer equation can be derived (Equation 4-1).


C T kff T 82T d 2T dE + (4)T
PeCf fft --r +k +k up(1-)Cg, (4-1)
r Or effa2 efaC2 c/V ga

The variables in Equation 4-1 are defined as follows: peffis the effective mass

density of the catalyst-gas mixture, Cgf is the effective specific heat of the catalyst-gas

mixture, Tis the temperature inside the control volume, keffis the catalyst-gas mixture


effective conductivity, r is the radial coordinate, x is the axial coordinate, dEg is the net

rate of heat absorbed by chemical reactions, Vis the reformer volume, pgas is the density

of the gas, e is the void factor of the reformer, and Cp, is the average specific heat of

the gas. The thermal capacitance of the mass inside the control volume is assumed to be

mostly because of the catalyst. Therefore pefjCff is approximately equal to the catalyst

density (pc) multiplied by the catalyst specific heat (Cc) and the void factor (e).

The use of effective properties for modeling of catalytic reactors is a fairly common

practice. Incoprera et al., in their widely used heat transfer text book, "Fundamentals of

Heat and Mass Transfer," describes heat transfer correlations for packed bed systems

using effective gas/packed-bed properties [17]. In published literature, various methods

for determining packed-bed effective thermal conductivity have been explored. Typical

of this work are the studies conducted by Munagavalasa et al. [18], and Brucker et al.

[19].









OT
The convection term in Equation 4-1, upg, Cp, for modeling purposes was


assumed to be small as compared to the radial conduction terms. Typically, temperature

gradients in the radial direction are typically higher than those in the axial direction

through most of the steam reformer. A possible exception to this could occur at the

entrance of the reformer if the entrance temperature is significantly different than the

reformer catalyst bed temperature. As an example of the disparity between the convection

and conduction terms, an order of magnitude analysis of these, based on data obtained

from the UC-Davis reformer (see Appendix C), is presented in Figure 4-3. The analysis is

executed for a run in which various changes in methanol flow were executed. The results

were obtained from temperature data recorded at the entrance regions of the reformer. A

finite difference approximation of the derivative and second derivative of temperature in

the axial and radial directions, respectively, were calculated using K-type thermocouple

readings. The radial thermocouples were placed at the wall, at half the radius and at the

midpoint, where the radius of the tube was 11.9 mm. The axial thermocouples were 67.6

mm apart. It is recognized that the significantly larger axial distances used to approximate

the finite difference axial temperature gradient leads to greater averaging. This in turn

may lead to an underestimation of the axial gradient in the entrance region. In spite of its

limitations, this analysis provides a general relative magnitude of the axial and radial

temperature gradients. During most of the run, the convection term did not exceed 0.4%

of the value of the conduction term. This, in order of magnitude terms, supports the

decision to disregard the convection term in the energy equation. The UC-Davis reformer

is discussed in greater detail at the latter sections of this chapter and in Appendix C.









0.01
0.008
0.006 -
0.004 -
11 0.002-






S- --0 ratio Zone 1 ratio one 2 ratio
I 2C -0 10 0 30 40] 50 60










cia ls -p.oum (- ', )e c it
-0.008 -
-0.01
Run Time (mins)
--Zone 0 ratio Zone 1 ratio -Zone 2 ratio |

Figure 4-3. Convection-conduction ratio in the UC-Davis reformer fuel entry regions

The heat dissipated in the catalyst bed is due to the reforming reaction.


Consequently, dEg is a function of the rate in which methanol is reacted in the steam


rethnolam reforming reactions3 This in turn is a function of the temperature (),


catalyst volume (dVc), and the molar concentration of methanol in the reformate

gas (xcHH ). The rate of methanol consumed per unit mass of catalyst (rm) and can be

obtained using Equation 4-2

l1 d[CHOH]_ 1 d[CH3,OH]
rM = f(T,P, XCH,OH )
mc dt rpcdrdx(2r dr) dt (4-2)

The calculation of the rate of methanol reacted in a CuO/ZnO catalyst for

methanol-steam reforming reactions was extensively studied by Nakagaki et al. at

Toshiba Power Systems [12]. One of the results of this study has provided a good

correlation for rm, shown here:










( T RT
rM P eXCHOH
r. = kP 51 e CH30H

ko = 1.35 x 106mol/(gca, s. atm)
m = -10 if T > 513K otherwise it's equal to 0 (4-3)
E=lxl05J/mol
= 0.13
n=1.3

The Nakagaki correlation is an empirical variation of the Arhenius equation and

was obtained experimentally. Further information regarding the Nakagaki study is found

in the literature review chapter of this text.

Given these relationships, dE, can be defined explicitly (Equation 4-4).


d[CH7130H]
dE, = E, = pi[C ] xdr(2r + dr)Err (4-4)
dt

where the expression for the heat of reaction per mole of methanol consumed is

expressed as Er.

Therefore the energy equation for the control volume can be rewritten as Equation

4-5.

OT a OT a2T '2T E r,
-= -- +a +a (4-5)
at r Or ar2 a2 CC

The quantity Er, is a function of temperature, thus Equation 4-5 is non-linear.

This non-linearity is the source of the complex and sometimes unexpected behavior

associated with packed bed steam-reformers. The boundary conditions for the heat

equation (Equation 4-5) are









OT
=0
ar r =0
T(r = R)= T
T(x = 0) = T (4-6)
aT
=0

T(t =0) =

Given these boundary conditions the energy equation can be solved by using a

finite difference scheme or using finite element analysis. Also, Equation 4-5 can be made

non-dimensional to yield:

OT* 1 T* 02T* R2 2T* R2 ME,
OFo r* Or* ar*2 La2 x2 (47)
ST r x at
T =-r = ,x = ,Fo=
T7 r L R

For purposes of this study, the heat transfer model was solved in dimensional terms

in order to relate results to real physical parameters which can be readily verified through

experimentation.

Continuity

The fluid flow within the steam reformer tube is transient, compressible and

viscous. If it is assumed that the mean flow travels in the axial direction (u, >> ur), then

the continuity equation can be written as Equation 4-8.

Pg a(p(4-8)
at t x

where ux is the axial component of the fluid velocity vector, and pg is the density of the

fluid. If the fluid is assumed to be an ideal gas, then the continuity equation can be

written in terms of fluid temperature. In addition, it is convenient to use mass flow rate,









m, instead of the velocity because cross-sectional averages of the velocity will be used.

This yields


8m AeP dT
a (4-9)
ax RT2 at

In essence Equation 4-9 states that variations in the axial mass flow are a function

of the transient change in temperature at any given location inside the reformer.


However, reformer data collected from the UC-Davis reformer suggests that for
at

methanol steam reformers is small even at the entrance, centerline region (Zone 1 and

Zone 2). This is because the thermal capacitance of reformer catalysts is relatively high,

and because the primary mode of heat transfer is through conduction, which is relatively


slow. Figure 4-4 shows recorded values of from the UC-Davis reformer. For
at

illustration purposes, even if the maximum value exhibited was held at -2 K/s (the

maximum recorded from the UC-Davis reformer during reformer warm-up) and

assuming that this level of heating were to be carried out through out the entire reformer,

the change in reformate mass flow rate that would be detectable at the reformer exit

would be 0.018% for a 1 meter long reformer, based on Equation 4-10.

L A P OT
hout = I dx (4-10)
o RT2 at

Therefore, the mass flow rate within the reformer will have very small transient

variations. Despite this, Equation 4-10 was used in the elaboration of the reformer model.

Temperature distribution plots for the UC-Davis reformer at varying flow rates are shown

in Figure 4-5. Note: for modeling purposes the pressure drop in the reformer was

neglected.























I3




-0.5





0.5
-0.5 -


-1


-1.5


-2

0 10 20 30 40

Run Time (min)


50 60 70 80


-Zonel dT/dt -Zone2 dT/dt




Figure 4-4. Calculated values of dT/dt for the UC-Davis reformer at the entrance region


1 0 LHSV-M, 99 8% 2 0 LHSV-M, 97.5% 3 0 LHSV-M, 93 1% 4.0 LHSV-M, 85 8%
L-- -.1- L .., L~u..


. EIL


1~ EL





l.rs :Ll





_~' Ll












~a~ir





Ilrr LEL


r-h..-r .r


fr- rn .


r-.. .*


.somc
-4 0.CO

-220.CO

-*am


*,-,,n C


Figure 4-5. Measured temperature distributions of the UC-Davis steam reforming test rig

operating at varying space velocities and varying conversion efficiencies


~r~ L~





I.~: LL


~.-' LEl I


-I -E


S-B Edl


;.R 'il

.sI~C ~I


J' r-









Conservation of Specie

The continuity equation for each of the component gases in the control volume can

be written as Equation 4-11.

rate of increase mass flow of j mass flow of j rate of production
of mass of j in = into the control out of the control + of j inside the control
the control volume volume volume volume
or

-- = m ,n ,ou jt + j,gen
Qt

(4-11)

The component gases in the steam reformer products are assumed to be methanol,

water vapor, carbon dioxide, carbon monoxide, and hydrogen. The mass generated of

components inside the control volume can be estimated using the Nakagaki correlation

(Equation 4-3) and chemical equilibrium. Thus, for a control volume at a certain

temperature and pressure, the rate of methanol reacted is known and, consequently, the

rate of generation of the product gases is also known. The rate of mass of species

entering the control volume is generally known from boundary conditions. However the

rate of mass escaping the control volume must be calculated.

For purposes of the model it was assumed that all species flow at the same axial

velocity as the bulk flow. The bulk flow velocity exiting the control volume can be

expressed as Equation 4-12.

.4Pfn, .2 at + n RT 1h
(J (-Tdx +m\RT
_t RT RT2 n) ) RT 1 aTx
OUt Pou,A PA,, PA,, PA,, oT2 ,out
(4-12)

Consequently, the mass flow rate of the exiting gasj is given by Equation 4-13.









P A
vis u = xjm (4-13)
m],out RT j,',ot Xmout (4-13)


In the above equation, P, is the partial pressure of gasj inside the control volume.

The term x, is the molar concentration of gasj, and is defined as Equation 4-14.



x (t = to) (4-14)

S1 to

Chemical Model

The chemical reaction for steam reforming of methanol being considered is shown

in Equation 4-15.

CH30H + xH20 aH20 + bCO2 + cCO + dH2 + eCH30H (4-15)

In a control volume a certain rate of methanol can be reacted (rMmc). The remaining

methanol and water that enters the control volume is not reacted (Figure 4-6). It was

assumed that the water-gas shift reaction is not the time limiting step in the reformer

reaction and thus proceeds to equilibrium rather quickly. For modeling purposes the

water gas shift reaction was assumed to reach equilibrium concentrations, while the

reforming of methanol was bound by the rate of reaction constraints of the control

volume. Similar approximations have been made in literature and the work of Rabou et

al. [20] is an example.










cv


Water
CI 1CO2
CO
Methanol (mol/s)
SNtcalcuareacted (mol/s)n t f
H, --------- H,
M a (mol/s) ..........................................













zero) leads to
bCd2 .......(4-17)










c=3-d (4-18)

The solution for ch can be obtained from the water-gas shift reaction
calculated in the following way.










(Equation 4-19)
a=N2+x-d (4-16)











HO + CO H2 + CO, (4-19)

The equilibrium constant for the water gas shift reaction can be approximated with

the Moe (1962) correlation [21] (Equation 4-20).

4577.8
InK = -4.33 + (4-20)
T


where Tis the temperature of the reaction in Kelvin. Given this,







41



d xKp 2xKp(Kp 4)+Kp +4Kp + 4 xKp 5Kp + 2
d = (4-21)
2(Kp 1)

With these equations the number of moles of each chemical component per mole of

methanol reacted can be obtained as a function of temperature. The number of moles of

methanol reacted can be obtained from the Nakagaki correlation (Equation 4-3). The

equilibrium heat of reaction and products as a function of temperature for the methanol

steam reforming reaction is given in Figure 4-7.


3.5 90000
Equilibrium Reformate Composition
3 // 80000
3- -
o 70000 -
2.5
60000 5

50000 4
E Magnitude Depicted in Secondary Axis*
E 0
1.5 40000
o
3 30000
"- 1 -= -- -- -- "-' b n .0
S20000 cU
o 5,
0.5
E 10000 4
0 0 -
I- 1 0 0 1 0 I 0 I
(C
298 398 498 598 698 798 898 998
Temperature (K)
-- CO2 H2 -o-CO Hr [kJ/kmol of MeOH]

Figure 4-7. Equilibrium products of the steam reforming reaction (CH30H + H20 -
aH20 + bCO2 + cCO + dH2) with respect to temperature

Finite Difference Solution

The control volume analysis easily leads to a finite difference solution of the heat

transfer and fluid flow problem presented in the previous sections. A grid was set up

where each element has a temperature (Tm,,), in which m represents the node index in the










x-direction and n represents the node index in the r-direction (as shown in Figure 4-8).

The derivatives of the catalyst bed temperature were approximated to be

02T Tm+i,n + m-,n 2T,n
2 (A2
02 T +TL,- 2T,
-- Tn +mn- (4-22)
r32 (Ar)2

1m,n+ 1 Tm,n-1
Or 2Ar

Similarly the time derivative of the temperature could be written as


BT T l" -T :
T mn- I (4-23)
at At

where / is the time index.


Rr







Z-x
mn+l
Ar -


Ar* Ar*
mn m+1,n




Figure 4-8. Separation of the reformer volume into annular elements with constant gas
properties in space for the finite difference solution method used

Given these equations, Equation 4-5 could be rewritten as









T /+1_ /

At
+T -2 + k T ^L + 7^ ^,n -2T7^ n
STmn +l ~nB-l .Tl n+ +,B-l 2Tnlmn J +1 -n ~-,n n
r 2Ar (Ar) 2pCC (Ax)2

(Er'M )m,n'
Cc
(4-24)

Similarly the boundary conditions can be rewritten (Equation 4-25).

T =T
M,0= 0 lT


Tn = Tn (4-25)
T 1=T 1
TM,n M 2,n

T, T

The term (E r)m,n, + is unknown and it is a function of the finite volume

temperature and methanol concentration. At first, the temperature equation can be solved

assuming a temperature and a methanol concentration. For the computer program

developed for this study, Tm,n,+ was initially assumed approximately equal to Tm,,'.

Similarly the concentration of methanol inside the volume, XCH30H,m,n+ was at first

approximated to be equal to XCH30H,m,n1. Once an approximate solution for the temperature

was obtained the conservation of specie equation was solved.

The finite difference expression for 4-10 is given by

1+1 1
mjn ,mn +1 -. +1 ;+1 (4-26)
At j,m l,n ,m,n +mm,n,gen


The expression for the mass flow of species leaving the finite volume can be

expressed as











1,n1 Xjm A 0.5 "1 T m (4-27)
At (m1,n (Tm,n 1

Note that as with the temperature solution, xjn,, 1+' is unknown. In the computer

program this value was initially assumed to be equal to Xj,,,,,. From the solution of the

content of all the species in the finite volume using Equation 4-24 and 4-27, for a given

temperature field, a new value for the molar concentration of each component can be

obtained (Equation 4-28).

M,m n
MW
Xjmn = +1 (4-28)
i(m n



This new, revised solution for xj,m,, 1 was used to solve for the finite volume

temperature and mass contents. This process was carried out until satisfactory

convergence was achieved.

FEMLAB Solution

Besides using a finite difference scheme, the system of differential equations was

solved using finite element analysis using FEMLAB Multiphysics modeler software. The

FEMLAB software is designed to solve systems of coupled partial differential equations

using proprietary solution schemes based on finite element analysis. The system of partial

differential equations for transient heat equation (including the convection term),

transient conservation of species, transient continuity, chemical equilibrium and the

Nakagaki correlation were programmed into the FEMLAB solver. A triangular element

grid was used to cover the two-dimensional reformer space. Appendix B specifies details









of the number of elements and number of degrees of freedom for each of the reformers

modeled.

The results of the FEMLAB finite element solution and the finite difference

solution were similar and compatible. However, it was found that the finite difference

solution resulted in lower reformate CO concentration. This was due to problems with the

inherent lower magnitude of CO concentration in the reformate gas as compared to water,

hydrogen and carbon dioxide concentrations. This difference created problems with

convergence and minimum error search. The FEMLAB solver uses sophisticated

minimum error search schemes which include parameter step annealing and magnitude

parity. Because of this computational error reduction feature, the FEMLAB solution was

assumed to have a higher degree of accuracy than that of the finite difference scheme.

Steam Reformer Model Validation

Validation of the steam reformer model results were done through comparison of

the performance of modeled steam reformers with the UC-Davis reformer. Experiments

were conducted using a small-scale laboratory steam reformer located at the University of

California Davis (See Appendix C). Data from this experimentation was used for model

validation. The UC-Davis reformer was built and instrumented by Dr. Paul Erickson and

his students. Experiments for this study were conducted by UC-Davis students under the

direction of the author of this dissertation. The experimental plan was also developed by

this author for the purposes of this study. Various peer-reviewed publications have been

based on research conducted using the UC-Davis reformer. These include Liao et al. [22],

Yoon et al. [23], and Betts et al. [24] (the latter regarding some of the results of this

study).









Data obtained from the UC-Davis reformer cannot be used as a complete model

validation tool. The reformer suffers from loose wall temperature control and gas

sampling limitations that inevitably lead to discrepancies between the magnitudes of the

model results and the data collected. Nonetheless, important physical phenomena, such as

transient CO spikes, CO concentration dependence on reformer temperature, and steady-

state and transient reformate composition trends are captured in the data obtained.

Agreement between the model and the UC-Davis reformer were expected to be

approximate.

General Description

The UC-Davis reformer is a catalyst filled cylinder with a radius of 0.47 inches

(0.012 m), and an axial length of 21 inches (0.533 m) (Figure 4-9). The reactor

temperatures are controlled using eight electrical band heaters located on the surface of

the reactor. Thermocouple readings of the internal wall temperature of the reactor and the

centerline temperature are used to gauge the level of chemical activity in a certain region

of the reformer and are also used as feedback controller inputs for the heaters. The

reformer heaters operate in on or off mode, as triggered by computer controlled relays. A

simple feedback control loop, programmed in Labview, was used to maintain a constant

wall temperature.

Before entering the reformer water and methanol are vaporized and superheated

using a series of pipes containing cartridge heaters. The temperature of the reactants

entering the reformer is recorded and is controlled via a control and data acquisition

computer, which is capable to turning on or off the vaporizer and super-heater heaters.

Water and methanol fuel flows are metered using positive displacement gear pumps. The









gear pump RPM was used to measure flow. A scale, on top of which the reactants sat,

was also used to measure methanol and water flow into the reformer.

Reformate gas analysis was done using a NOVA Analytical Systems, Inc. gas

analyzer (Model 7904CM). This analyzer was designed for simultaneous analysis of

hydrogen, carbon dioxide, methane, and carbon monoxide. The instrument recorded

carbon dioxide, carbon monoxide and methane using a Non-Dispersive Infrared (NDIR)

detector with temperature and pressure compensation. Hydrogen was detected using a

temperature controlled thermal conductivity (TC) cell.

Before being analyzed for composition, the reformate gases were cooled to an

almost freezing temperature (i.e. -0C) and the condensed liquid was extracted. The UC-

Davis reformer was designed with the capability to capture all condensed liquids before

gas sampling to perform conversion studies. Due to the condenser volume, a short time

response delay exists in the gas analysis data. Although the exact time response delay is a

complex computation problem that falls outside the scope of this study, a general

estimation can be performed. The length of the condenser tube is 18ft (5.5m). Reformate

enters the condenser at approximately 300C and leaves at approximately 0C. If the mass

flow rate of fuel entering the reformer is assumed to be the same as that entering the

condenser, then the reformate volume flow entering the condenser can be calculated. For

purposes of estimation the mass flow rate leaving the condenser can be said to be equal to

the reformate mass entering it. Based on these assumptions the time it would take the

time it would take to have a complete condenser volume gas change (i.e. retention time)

for a premix flows at STP ranging from 5 mL/min to 30 mL/min would range between

1.3 to 0.2 seconds. Additionally, for all premix fuel flow rates, the flow through the









condenser can be estimated to be mostly laminar due to low Reynolds number. This

means that a well organized concentration boundary layer is expected to exist within the

condenser pipe. Therefore, although the reformate gas composition may change, traces of

previous reformate gas could still be trapped in the boundary layer, close to the walls.

This may lead to a decrease in the sharpness of the data obtainable from the gas analyzer.

The NOVA gas analyzer manufacturer-quoted response time is 3 seconds to 90%

response. The gas analyzer takes samples at a rate of 1 L/min through a positive

displacement pump. At this flow rate the condenser unit would have a retention time in

the order of 1 s. Since the response time is in the order of seconds and the measurable

effects (that is reformate gas concentration changes) occur in the order of minutes, these

effects are recordable using the UC-Davis hardware configuration.

In terms of time response, passage through the vaporizers may be the limiting

factor. For the model, the fuel flow rate boundary condition is at the inlet of the reformer.

This is not the case for the UC-Davis reformer. If it were assumed that only liquid fuel

passes through the vaporizers, then the residence times of the vaporizers would be as

shown in Table 4-1. The estimated residence times presented in this table are maximum

values. For lower fuel flows the space velocity is over-estimated because vaporization

occurs earlier in the vaporizer tube, which increases the volume flow rate of the fuel.

Also, the degree in which vaporization occurs within the vaporizers is not immediate and

this will result in a ramped increase in fuel flow into the reformer instead of a step

increase as in the model. This should result in smoother, drawn-out, transient gas

composition data for the UC-Davis reformer.









Table 4-1. Estimated maximum times for vaporizer volume flow change.
Fuel Flow Rate (cm3/min) Vaporizer residence times
5 5 min
10 3 min
15 2 min
20 1.3 min
25 1 min
30 -50 s


Experimental Method

The reformer catalyst bed was maintained idle for storage completely sealed and

pressurized with carbon dioxide. This process ensured minimum catalyst deactivation due

to oxidation. For experimentation, the reformer was first heated using the reformer band

heaters under a CO2 gas purge. Once the appropriate reformer temperature was reached

(250C), vaporized and superheated water and methanol were introduced. The experiment

maintained the reformer wall temperature at a nominal 3000C with fluctuations due to

heater binary control. Water-methanol flows between 5 and 30ml/min were passed

through the reformer in 5ml/min intervals. Changes in methanol-water flows were

executed every 10 minutes. The data sampling rate was two seconds.

UC-Davis Reformer Data

The data obtained from the UC-Davis reformer shows the following trends:

* There is a temporary increase in the CO concentration in the reformate gas leaving
the reformer when there is an increase in fuel flow rate into the reformer (Figure 4-
10).

* The CO concentration in the reformate is very closely related to the catalyst bed
temperature (Figure 4-11). Catalyst bed temperature fluctuations in the UC-Davis
reformer were mostly because of band heaters binary operation and imposed
control scheme.

* The steady-state CO concentration in the reformate lowers with increased fuel flow
rate into the reformer (Figure 4-10 and Figure 4-12).







50



1.33"

2.66"


2.66"

2.66"

Overall Catalyst 2.66"
Reactor Packing
Length Length 2.66"
(24") (21")

2.66"

2.66"

2.66"

1.33"





Figure 4-9. Experimental steam reformer developed at UC Davis


16 3
Data obtained at U C Davis on 2/5/2004

14
2.5
S12 Increased levels of CO are registered
o when there is an increase in flow rate 2
LL 10 y
25


8 1.5c

S6

S4
Steady state CO levels are lowered 0.5
2 with increased flow rate
CO Spike Due to Change in Flow Rate
0 -0
22 23 24 25 26 27 28 29 30
Run Time (min)
Pump feedback (ml/min) CO concentration (%)


Figure 4-10. Example of CO spike in the reformate of the UC-Davis reformer after an
increase in premix fuel flow (transient condition). Also shown is a decrease of
the steady-state CO concentration in the reformate with increased premix fuel
flow rate in units of mL/min.






















o
250

240

E 230
0)
I-
220

210

200


4.5

4

3.5

3

2.5
C-
0
2

0)
1.5 1
C-)
1 0

0.5

0


5 15 25 35 45 55 65 75
Run Time
Reformer Exit Centerline Temperature -Average Centerline Temperature (oC)
CO concentration (%)

Figure 4-11. Reformate CO composition relationship to changes in temperature in the
reformer catalyst bed1


80

70

60 -

SE 50


t 40
M -

o 30

co .x 20


0
a 10

0

5 15
Feb. 5th, 2004 Test Run at UC Davis


3


2.5


2





1 0
0




0.5
0
0
0

0


25 35 45 55
Run Time (min)


-Pump feedback (ml/min) -H2 concentration (%)


-CO concentration (%)


Figure 4-12. CO and H2 concentrations in the reformate gas in the UC-Davis reformer







1 The drastic decrease in CO concentration at the end of the graph comes as a result of shutting down the
premix flow into the reformer.









It was a major goal of this investigation to be able to analytically reproduce the

general operating characteristics of actual reformers. For this purpose an analytic version

of the UC-Davis reformer was developed using the heat and chemical rate equations

previously outlined. The virtual version of the UC-Davis reformer exhibited all of the

major performance characteristics of the actual reformer to varying extents.

The constants used for the virtual reformer are shown in Table 4-1 and were

matched as best as possible, to measured or inferred UC-Davis reformer constants.

Table 4-1. Values constants used to develop virtual UC-Davis reformer
Name Value Used How it was
determined
Catalyst Bed Heat Capacity (C,) 900 J/kg-K Inferred
Catalyst Bed Thermal Conductivity (keff) 5.00 W/mK Estimated
from Data
Catalyst Bed Density (epc) 1983. kg/m3 Measured
Length (L) 0.53340 m Measured
Radius (R) 11.9 x 10-3 m Measured
# of Radial Elements (NR) 5
Number of Axial Elements (NL) 150
Time Step (At) 0.5 s
The catalyst heat capacity was assumed to be equal to an average of the heat
capacities of copper and zinc.
** The catalyst heat capacity was assumed to be equal to an average of the heat
capacities of copper and zinc.
The void factor used was 50% and the catalyst density was measured via a
water displacement method for a single pellet.


To determine the value for catalyst bed thermal conductivity (keff), an average was

taken of the measured thermal conductivity of the UC-Davis reformer catalyst bed at

different points during the warm up stage, when no fuel flow exists. Figure 4-13 shows

an example of the data used a band heater zone in the UC-Davis reformer.

The resulting steady-state carbon monoxide composition of the virtual UC-Davis

reformer are shown in Figure 4-14. In arrows are listed the average steady-state CO







53


compositions for the actual UC-Davis reformer. In this figure, note that greater flow rate

results in lower steady-state CO concentrations.

Error between the UC-Davis reformer steady-state CO concentration and that of the

virtual reformer is potentially due to the existence of un-reacted methanol in the sampled

gas. Although before sampling the reformate gas was cooled to extract condensate, the

difference between methanol and water freezing temperature only permitted a reduction

in temperature to almost 0C. The existence of unreformed methanol in the sampled gas

is detectable through the methane (or hydrocarbon) signature detected by the NOVA gas

analyzer (Figure 4-15).


Zone 2 Thermal Conductivity
340 6
Average Conductivity = 4.87 W/(K-m)
320

300 --
o 4 E
2 280
3 '
CL 260 __ _
E
2 "-2
240 -

220 1

200 0
0 1 2 3 4 5
Time (mins)
-Zone 2 Surface -Zone 2 Exit Wall -Zone 2 Exit Centerline Effective Conductivity


Figure 4-13. Estimation of catalyst bed thermal conductivity using UC-Davis reformer
data











0.0250


0.0200


S0.0150
X6-

E 0.0100
0

0.0050


0.0000


0 5 10 15 20 25 30 3'
Premix Flow (std.mL/min)

--x CO (dry) --x CO (dry w/ MeOH)


Figure 4-14. Virtual UC-Davis reformer steady-state CO composition with varying
premix fuel flow rates


0 10 20 30 40 50 60 70

Run Time (min)

Figure 4-15. Hydrocarbon signal output obtained during UC-Davis reformer run.


S* 1. 02 1.85E-02


02 o93- 02 1.36E-02
S1 093%
UC-Davis UC-Davis










Figure 4-16 shows a representative CO spike at the point in which the premix flow

rate changes from 10 to 15 ml/min. The change in flow rate was produced at time zero.

This figure is meant to be a parallel to Figure 4-10. Figures 4-16 and 4-17 show the

temperature distributions of the modeled reformer for 5 ml/min and 30 ml/min flow rate,

respectively. The temperature distribution is similar to those recorded for the UC-Davis

reformer (Figure 4-4), where the minimum temperature is recorded at the reformer

entrance region. These results support the observations made from data obtained from the

UC-Davis reformer and therefore demonstrate the validity of the assumptions made in the

model presented.


0.0165
Virtual Reformer CO Spike (10 to 15ml/min)
0.0164


0.0163
O
Z 0.0162
.2

0.0161


0.016


0.0159


0.0158
0 10 20 30 40 50
Time After Point of Flow Rate Change (s)

Figure 4-16. Virtual UC-Davis Reformer CO spike from 10 to 15ml/min premix flow rate
change


















qpmix(1J=6 Surface T Contour ineoh


0.2 0.3 0.4 0.5

Reformer Length (m)


Temperature (K) Methanol Concentration
Max: 575 7 MBa 0.390


570 0372
-0352
S0.332
5 0031
0.21



02T2
o0.211

JL 0171

0.151
in 0.131
0.111


0J.T1
0051
500 0031

S440 in 011
rin 44.20 Mm 00110


Figure 4-17. Virtual UC-Davis reformer temperature distribution (color plot) and

methanol concentration (contour plot) for premix flow rate equal to 5 ml/min


qpmi(6)=30 Burface I Conlour rieah


0o012


0.1 0.0..3 0.4 0.5
Reformer Lengthri (m)


Temperature (K) Methanol Concentraton

Ma: 675 7i Maw 0.80(

570 -03
S-036
S-0341
580

0.303
550

o0261
540 -0245
-0226
.L' 0207
-01B2
520 -0.156
-014A

510 0.13
011
oil

500
0072
0053

Mn 427 7 Mmn 0 0 f2


Figure 4-18. Virtual UC-Davis reformer temperature distribution (color plot) and

methanol concentration (contour plot) for premix flow rate equal to 30 ml/min


0.01



D goo

4.1

0'
0_ O.D0O




cr

O 001




0


0.1









Reformer Design

Currently, most reformer design is done based on a space velocity basis. This

method is not consistent with the complex nature of the reforming reaction and the

virtually infinite number of possible reformer designs. As part of this study, a set of non-

dimensional reformer performance variables are introduced in order to provide generality

to the results obtained from analytical modeling and experiments. This section generally

describes some parameters that affect reformer performance. In particular the question of

how reformer length, catalyst loading and fuel flow rate affects the reformate gas

composition is investigated using these suggested non-dimensional parameters.

Design Generalization

It has already been established that heat transfer and chemical rate characteristics

are major determinants of complex reformer performance. Equation 4-5 can be rewritten

as Equation 4-29.

p T ke ST f2T 2T T
PsC + f -k + k EpEr.
dt r Or dr2 x (4-29)

In order to generalize the design of reformers, it is useful to define non-dimensional

variables. For a reformer tube of a certain radius (R), the volume of catalyst in the

reformer (Vc) will determine the length of the reformer (L). According to Nakagaki and

other correlations it could be said that the mass of the reformer has a potential

reformation capacity, Q. The potential reformation capacity of the catalyst could be

defined as the theoretical hydrogen power (PH2) that a certain mass of catalyst could

produce if the reformer were maintained exactly at the wall temperature, and with the

inlet methanol concentration. This can be easily calculated by using the Nakagaki









correlation. For example, the potential reformation capacity of the UC-Davis reformer

can be calculated in the following way.

The catalyst mass in the UC-Davis reformer is


m = py = cpv (R2L)= 1983k T(0.0119m)2(0.5334m)]= 0.470565kg (4-30)


If the wall temperature were to remain constant at 300C, then rM,max and the

potential rate of methanol reacted (,H3H ,reacted max )can be calculated as shown in

Equation 4-31.


rMmax (Tw, XCH30H.nlet)= 1.039610-4 kol
sec.kg (4-31)
kmol
iCH30H,reactedmax = rMm = 4.8921 x 10 nol
sec

Since the theoretical chemical reaction for the reformation of methanol yields three

moles of hydrogen per mole of methanol, the potential reformation capacity (Q) can be

obtained as shown in Equation 4-32.


S= 3LHVHCH30H,reactedmax = 3241620 4.8921 x 105 ko = 35.4609kW
S kmol sec

(4-32)

Also a non-dimensional temperature, 0, can be defined as

T
0= (4-33)


and a non-dimensional length and radius can be defined as

r
r _
R (4-34)

x
x =
L (4-35)









Given these formulations, Equation 4-29 can be non-dimensionalized to yield

VpJscTJ 00o VTkef 1 o0 VTk, 20 4Tw 20 pcE rMV
Q 9t RQ r* r* R2 r*2 V x*2 Q

(4-36)

Equation 4-38 is made up of various important non-dimensional groups. A non-

dimensional conductivity, k*, can be expressed as

S- VT^k
k-
R2f (4-37)

The non-dimensional conductivity can be construed as a measure of the ratio

between the potential radial conduction heat transfer into the reformer and the potential

rate of energy absorbed by the reforming reaction. It is important to point out that the

non-dimensional conduction is not actually a function of catalyst length. The non-

dimensional conduction could be expressed as

Tk
k-
3LHVH2 p,R2 (4-38)

For the UC-Davis reformer the non-dimensional conduction is equal to 0.135.

The last term in Equation 4-36, is a measure of the ratio of the actual rate of energy

absorbed due to the reformation reaction and the potential rate of energy absorbed by the

same reaction in terms of hydrogen production. The potential reformation capacity, Q, is

actually a measure of the potential hydrogen production of a certain mass of catalyst at a

certain temperature. Thus it would be convenient to express the last term in

Equation 4-36 in terms of hydrogen production and not methanol consumption. In

practical terms the hydrogen production is easier to measure than the methanol converted.









With this formulation a catalyst effectiveness, k, can be defined as the ratio of the

actual hydrogen power output obtained from a certain amount of catalyst (PH2) and the

potential reformation capacity of the catalyst (Q) (Equation 4-39).

PH2
Q (4-39)

The catalyst effectiveness is a measure of the degree of reforming activity that a

certain mass of catalyst undergoes. High catalyst effectiveness is desirable because it

means that the reformer is operating closer to its idealized potential hydrogen yield. High

catalyst effectiveness occurs in short reformers. On the other hand, the larger the

reformer, the lower the catalyst effectiveness will be.

From Equation 4-40, it is also notable the time constant term, r,

VpESC TI ClTl
Q 3LHVH2rMmax (4-40)

The transient response in a reformer would be strongly determined by r. A low

value of r would lead to very fast temperature changes in the reformer tube leading to a

decrease in prominence of the CO spikes. For the purpose of fully expressing the heat

equation in non-dimensional terms a non-dimensional time (t*) is defined

(Equation 4-41).

tQ
t = (4-41)


The non-dimensional heat equation is

0O k* O0 .020 e24Tw d 20 E.r,
+ k- + (4-42)
at* r* Or* Or*2 VQ a*2 3LHVH2rMmax

It is also convenient to non-dimensionalize the fuel flow into the reformer as










LHV MU mfuel
flowND
C (4-43)


where LHVfuel represents the lower heating value of the fuel, and mfuel represents the

mass flow rate of the fuel. The non-dimensional flow is a ratio between the rate of fuel

energy introduced into the reactor and the reactors reformation capacity. The non-

dimensional flow can be expressed in terms of space velocity (Sv) as shown in Equation

4-43.


flowND SvPfn (4-44)
3rM max Pc

Also useful is the definition of reformer efficiency, r)ref,

PH,
7ref =
fuel (4-45)

The chosen reformer efficiency definition neglects the power required to promote

and sustain the reforming reaction. This definition only deals with issues of conversion,

since a detailed analysis of reformer heating design is outside the scope of this study. In

Equation 4-42, PH, represents the rate of hydrogen chemical energy obtained from the

reformer, and Pfuel represents the rate of fuel energy introduced into the reformer. The

maximum reformer efficiency for a methanol reformer is given by

3LHVH
377ex LHVH2 (4-46)
LreCH30H

The non-dimensional flow can be re-written as


flowND CH 30H,, (4-47)
17ref max n CH 3OH ,,reacted,max







62


Given these definitions, the reformer model was used to obtain characteristic

reformer efficiencies versus catalyst effectiveness curves for reformers with the same

non-dimensional conduction (0.315) as the UC-Davis reformer at various fuel flow rates

and different lengths. These curves are shown in Figure 4-19. Additionally Figure 4-19

shows the corresponding non-dimensional fuel flow rates as a function of catalyst

effectiveness.


7

100 6


80 5 "
LL
xo 4
60 -


LU 40- ;
W 40

z

M agnitude represented on secondary axis ...... ...... 1

0 0
0 0.2 0.4 0.6 0.8 1
Catalyst Effectiveness

Efficiency (5ml/min) Efficiency (10ml/min) A Efficiency (15ml/min)
x Efficiency (20ml/min) Flow ND (5ml/min) -Flow ND (10ml/min)
Flow ND (15ml/min) Flow ND (20ml/min)


Figure 4-19. Reformer Efficiency with varying catalyst effectiveness at varying fuel flow
rates

From Figure 4-19 certain observations can be drawn. First, for non-dimensional

fuel flow values less than 1, the variations in reformer efficiency versus catalyst

effectiveness and non-dimensional fuel flow versus catalyst effectiveness, collapse into a

single representative performance curve. Second, as mentioned previously, lowered

catalyst effectiveness results in higher reformer efficiency. In other words, longer









reformers will exhibit higher steady-state efficiencies at the cost of ever increasing

catalyst mass. Therefore, there are diminishing benefits to a long reformer, which will

have a similar hydrogen yield as a reformer designed with catalyst effectiveness equal to

0.1. As the fuel flow rate of a reformer approaches infinity, the catalyst effectiveness will

approach 1 and the reformer hydrogen yield will approach zero.

These results can be used to scale a reformer using non-dimensional variables. For

the purposes of this study, three reformer designs were considered: an infinite reformer

with catalyst effectiveness equal to 0.05 at maximum power; a mid-size reformer with

catalyst effectiveness equal to 0.1 at maximum power; and a small reformer with catalyst

effectiveness equal to 0.2 at maximum power. The terms small, medium and large are

used as visual clues to the general aspect ratio that these reformers would have in

comparison with each other if the diameter were to remain the same. All reformers were

designed to produce a maximum 60 kW of hydrogen (lower heating value based). If a

reformer tube with the same non-dimensional conduction as the UC-Davis reformer were

to be scaled to produce 60 kW of power, its length and steady-state operating

characteristics would be as shown in Table 4-2, based on steady-state catalyst

effectiveness and reformer efficiency analysis.

Table 4-2. Steady-state design of 60 kW reformer with an 11.9 mm radius and a catalyst
bed density of 1983 kg/m3
Catalyst Reformer Tube Steady Sate Premix Power to
Effectiveness (k) Length (m) Reformer Produce 60 kW of H2
Efficiency (%) (kW)
0.05 18.05 112.9 53.14
0.10 9.03 112.5 53.35
0.20 4.51 110.5 54.28

D. Davieau [16] studied the use of space velocity and aspect ratio as major

similitude parameters for reformer design. His conclusions are that these two parameters






64


are not sufficient to fully characterize reformer steady-state conversion and reformate

quality. These conclusions are supported by the analysis presented in this study. The

space velocity and aspect ratio, although important, do not capture the governing physics

of the reformation process.














CHAPTER 5
RESULTS

In this chapter, the primary chemical and heat transfer conditions that produce

variations in hydrogen and carbon monoxide levels in the reformate are presented and

discussed. Special attention is given towards the benefits and detriments associated with

reformer size and catalyst loading in the context of fuel cell system incorporation. For

this purpose three general reformer length with the same non-dimensional conduction

were compared based on their reformate quality.

Transient Efficiency

Case 1 Infinite Reformer

Figure 5-1 shows predicted transient changes in reformer efficiency for various

changes in non-dimensional fuel flow rates for the infinite reformer. This figure shows

how increasing fuel flow results in an immediate drop in reformer efficiency followed by

a gradual approach to the steady-state efficiency. Figure 5-2 shows the same effects on

the reformer efficiency in terms of the efficiency difference variable, re~, defined as


77ref = rref rref,SP (5-1)

The variable 7ref, .p. is the set point efficiency, which is the steady-state efficiency

at a certain flow rate. In other words, for any steady flow rate the difference efficiency is

equal to zero.

The magnitude of the initial efficiency overshoot is summarized in Table 5-1. It is

notable that the magnitude of the efficiency-overshoot increases with increased change in







66


non-dimensional flow. However, if the change in non-dimensional flow remains constant,

changes between smaller non-dimensional flow quantities results in less overshoot. The

magnitude of the drop in efficiency before reaching steady-state is also recorded. This is

called an efficiency undershoot. High overshoot magnitudes reduced the extent of the

efficiency undershoot.

Table 5-1. Case 1: Summary of the transient response of the infinite reformer to changes
in fuel flow rate
Flow ND Change in ND Flow Overshoot
Initial Final
1.10 2.20 1.10 0.019%
2.20 3.31 1.10 0.027%
3.31 4.42 1.10 0.046%
4.42 5.52 1.10 0.064%
5.52 6.63 1.10 0.072%
1.10 6.63 5.53 0.260%



123.3
Infinite Reformer Transient Efficiency
123.2 I


1 3 *1- ---- - - - "- -" "" "-
123.1 ...... ..........................

3 123.0-
I ^_----~----------- -^------ ----
.2 122.9

122.8 i i

122.7

122.6
0 20 40 60 80 100
Run Time (s)
10 to 20kW -20 to 30kW ----30 to 40kW ..... 40 to 50kW
---50to60kW --10to60kW

Figure 5-1. Case 1: Infinite reformer operation when undergoing step variations in fuel
flow rate










0.3
Infinite Reformer Transient Efficiency

0.2


0.1


S0.0 -
-00

-0.1
1 0kW:1.10%
20kW: 2.20%
30kW: 3.31%
-0.2 40kW: 4.42%
50kW: 6.63%

-0.3
Run Time (s)

10 to 20kW 20 to 30kW ---- 30 to 40kW ..... 40 to 50kW
---50 to 60kW -- 10 to 60kW


Figure 5-2. Case 1: Reformer efficiency changes due to change of fuel flow rate for the
infinite reformer

Case 2 The Medium Sized Reformer

As with the infinite reformer, the medium sized reformer was subjected to transient

variations in fuel flow. The changes in efficiency of the reformer are depicted in Figure

5-3 and 5-4. Figure 5-4 shows these changes in terms of the efficiency difference

variable. Table 5-2 summarizes these results. As with the infinite reformer, large changes

in non-dimensional flow result in large overshoots, and changes between lower non-

dimensional flows results in lower overshoots.

The magnitude of the efficiency undershoots for the medium reformer were

significantly lower than those of the infinite reformer. In addition, as changes in non-

dimensional flow rates occurred between higher non-dimensional flow rates, the

efficiency undershoot was decreased and even disappeared.







68


Table 5-2. Case 2: Summary of the transient response of the medium size reformer to
changes in fuel flow rate
Flow ND Change in ND Flow Overshoot
Initial Final
2.22 4.43 2.22 0.066%
4.43 6.65 2.22 0.129%
6.65 8.87 2.22 0.211%
8.87 11.08 2.22 0.292%
11.08 13.30 2.22 0.372%
2.22 13.30 11.08 1.406%



123.4
Medium Reformer Transient Efficiency
123.2 -
IjDL Fi- pl-,
123 t. .- '*v' 2 22"'..
123 -- - --- --- -
122. 20- 4 4
122.8 l :'. ."..
.. . . . . 4 .... "..
122.6 11 ..


S122.2 -------------------------------------------------30o
LU
122
121.8
121.6
121.4
0 20 40 60 80 100
Run Time (s)
10 to 20kW -.-.20 to 30kW ..... 30 to 40kW
---40 to 50kW 50 to 60kW 10 to 60kW


Figure 5-3. Case 2: Transient efficiency of the medium size reformer under transient load
changes

Case 3 The Short Reformer

As with the infinite and medium sized reformer, the short reformer was run under

transient load conditions as shown in Figure 5-5 and 5-6. In contrast with the larger

reformers, the short reformer did not exhibit efficiency undershooting, except between

step changes in non-dimensional flow between 4.44% and 8.88%. Besides this effect, the







69


results follow a similar pattern to those obtained for the infinite and medium reformers.

The results of the transient run of the short reformer are given in Table 5-3.


1.6

1.4

1.2

1

0.8

) 0.6


Medium Reformer Transient Efficiency
rJ Fi.:. h- ,

2,'kI f -J -i
:.r"/if 6 6?.
-J'Ik\,l -
.n, kI\ l I I n ; -
16,,rw f I. ;.,,


0.2 -
O0 -- -- --- -

0 n -0 40 '.0 O0 10
/ I


Run Time (s)
---.20 to 30kW
- 50 to 60kW


..... 30 to 40kW
10 to 60kW


Figure 5-4. Case 2: Reformer efficiency changes due to change of fuel flow rate for the
medium size reformer


Table 5-3. Case 3: Summary of the transient efficiency response of the short reformer to
changes in fuel flow rate
Flow ND Change in ND Flow Overshoot
Initial Final
4.44 8.88 4.44 0.36%
8.88 13.31 4.44 0.73%
13.31 17.75 4.44 1.06%
17.75 22.19 4.44 1.27%
22.19 26.63 4.44 1.29%
4.44 26.63 22.19 9.33%


10 to 20kW
40 to 50kW







70


124
ND Flow Key:
\10kW: 4.94%
122 20kW: 8.88%
30kW: 13.31%
i 40kW: 17.75%
120 ---...... 50kW: 22.19%
S.....................................................60kW : 26.63%

118
a ~ -~~~ -''--------------------------------------
I 116


114

Short Reformer Transient Efficiency
112 .
0 20 40 60 80 100
Run Time (s)

10 to 20kW ----20 to 30kW ----- 30 to 40kW ---40 to 50kW
50 to 60kW 10 to 60kW


Figure 5-5. Case 3: Transient efficiency of the short reformer under transient load
changes

Carbon Monoxide Concentration

Generally, temporary increases in CO concentration in the reformate gas occurred

during periods of increased premix flow rates. This phenomenon occurs because lower

fuel flow into the reformer leads to higher average steady-state temperature. This increase

in temperature results from less fuel being reformed, thus lower overall associated

endothermicity. High temperatures promote greater CO and water formation in the water

gas shift reaction. High temperatures typically occur near the exit of the reactor. When

there is an increase in fuel flow into the reformer, the average temperature of the reactor

does not drop immediately and this contributes to better transient conversion (i.e.

increased efficiency), and higher transient CO concentrations. As time progresses the

average temperature of the reactor drops because of the increased endothermicity










associated with the increased level of methanol reacted. Eventually the lower average

reformer temperature results in a lower steady-state CO level at higher fuel flow rates.

An off-steady-state CO concentration is defined as


[CO]o ,,,ey =[CO- [CO]L,, (5-2)


1.4
Short Reformer Transient Efficiency
1.2 ND Flow Key:
10kW: 4.94%
1 20kW: 8.88%
30kW: 13.31%
40kW: 17.75%
0.8 50kW: 22.19%
60kW: 26.63%


0.4 i



0 --
0-------------------- --^"^"-.-.-^^^
0 20 40 60 80 100
-0.2
Run Time (s)


10 to 20kW ---20 to 30kW ----- 30 to 40kW --- 40 to 50kW 50 to 60kW


Figure 5-6. Case 3: Reformer efficiency changes due to change of fuel flow rate for the
short reformer

Figures 5-7 through 5-13 show the transient CO concentrations and off-steady-state

CO concentrations for the infinite, medium, and short reformer triggered by step changes

in fuel flow into the reformer.

The CO spike is defined as the initial increase in CO concentration that occurs after

a step increase in fuel flow occurs. This spike occurs before there is a gradually

decreasing approach of the transient CO concentration towards the steady-state CO

concentration. In all instances, the larger the non-dimensional flow change the greater the









CO spike. However, between equivalent changes in non-dimensional flow, larger CO

spikes occurred between lower non-dimensional flows. For example, in the infinite

reformer, a CO spike 328 ppm larger is recorded for a change in non-dimensional flow

from 1.10% to 2.21% as compared to a change from 2.21% to 3.31%.

CO spikes, as defined here, are a measure of the transient deviation of the CO

concentration as compared to the steady-state condition for the same flow. This deviation

was highest for small reformers, which in order to obtain a certain steady-state hydrogen

flow rate required a large change in non-dimensional flow. However, if Figures 5-8, 5-10

and 5-12 are examined it can be seen that the off-steady-state CO concentration also

includes the unsteady, smooth drop in CO concentration associated with the increased

fuel flow into the reformer. Changes in flow between low non-dimensional flows caused

greater disruptions to this smooth drop. In a fuel cell system these flow changes may

cause variations in fuel cell performance and a degree of CO poisoning during the

transient event if appropriate actions are not taken. From this standpoint, fuel flow

variations in reformers operating with high non-dimensional flows (i.e. small reformers)

should exhibit greater linearity in power output during load changes. The short reformer,

the change of non-dimensional flow from 22.19% to 26.63% exhibited a very mild CO

drop disruption.

In Figure 5-13, the CO concentrations associated with a similar change in reformate

hydrogen power output, namely from 40 kW to 50 kW (lower heating value based), is

depicted. For this case, the short reformer produced a lower steady-state and transient CO

concentration change than the larger reformers. This occurs because the catalyst







73



effectiveness of the small reformer is much higher than for the larger reformers, which


leads to lower average catalyst bed temperatures.


0.0195


0.0190


0
0.0185

U
.4-

S0.0180
0


0.0175


0.0170


20 40 60 80
Run Time (s)


S10 to 20kW 20 to 30kW ---- 30 to 40kW ... 40 to 50kW
--- 50 to 60kW 10 to 60kW

Figure 5-7. Infinite reformer transient CO concentrations for various for changes in
power output


2000



1500

E

= 1000
o

0
I 500

0
o.


0



-500


50 to 60kW --- 40 to 50kW .---- 30 to 40kW
10 to 20kW -- 10 to 60kW


20 to 30kW


Figure 5-8. Infinite reformer transient off-steady-state CO concentrations for various
changes in reformer power output











0.0200


Medium Reformer Reformate CO Concentration


~ 0.0195
o
U
2 0.0190
4-

0.0185
C

S0.0180


r 0.0175
o
S0.0170
- 0.0170


0.0165


. -- .. -- -- ------------
.......... -




0 20 40 60 80 100 120
Time (s)


10 to 20kW
40 to 50kW


-- 20 to 30kW
--50 to 60kW


-..-- 30 to 40kW
-10 to 60kW


Figure 5-9. Medium reformer transient CO concentrations for various changes in power
output


3000 Medium Reformer CO Spikes
Medium Reformer CO Spikes


ri F,:, Fi e
IOn J 2 22"..
270 kj 4 4,."..
? 0 / iJ ,!. 'i .5": ..
&40k' J 8 8 ..
5O0uJ 11 08"..
i:.O' J 1 ?. ?.".


.* '



0 20 40 60 80 100 120

Time (s)


10 to 20kW
---40 to 50kW


--20 to 30kW
- 50 to 60kW


..... 30 to 40kW
-10 to 60kW


Figure 5-10. Medium reformer off-steady-state CO concentrations for various changes in
reformer power output


2500

CE
- 2000


1500


o 1000











0.0185
Short Reformer Transient CO Concentration
c, 0.018 ND Flow Key:
.2 1 O J: 4.94%
0.0175 0r:8.88%
S-'-. ?.Ok J: 13.31%
S. 40k J: 17.75%
0.017 ........... 50kW: 22.19%
OC .. ....." 60kW: 26.63%

0 .0 1 6 5 - - - - - - --.- - - -
0.0165


0.016

O
0 0.0155


0.015
0 20 40 60 80 100 120 140
Time (s)


S10 to 20kW ----- 20 to 30kW --- 30 to 40kW 40 to 50kW 50 to 60kW

Figure 5-11. Short reformer transient CO concentrations for various changes in power
output


3000
Short Reformer CO Spikes

2500 ri FI,, Ke,

C.



o 1500 i '- '
2000 ;,'i vv i:, :'I'*





8 1000 -
0
U 1000


500


0
0 20 40 60 80 100 120 140
Time (s)

S 10 to 20kW ----- 20 to 30kW --- 30 to 40kW --40 to 50kW 10 to 60kW


Figure 5-12. Short reformer off-steady-state CO concentration for various changes in
reformer power output











0.018 Comparison of CO Concentrations for Different Size Reformers
Operated Between 40 to 50kW Steady State Power Outputs

0.0175

Q.
0.017-



0
ND Flow ranges:
Short Reformer: 17.75% to 22.19%
o m CO Medium Reformer: 8.87% to 11.08%
St Infinite Reformer: 1.10% to 6.63%
0
0.016-


0.0155 -..
0 20 40 60 80 100 120 140
Run Time (s)
Medium CO Short CO Infinite CO


Figure 5-13. Transient CO comparison of various size reformers operating between 40
and 50 kW hydrogen power output

Hydrogen Concentration

Transient and steady-state changes in hydrogen concentration for all reformers

were such that the performance of PEMFCs or PAFCs should not be drastically affected.

Even under the harshest transient conditions studied, in which a change in fuel flow was

induced to generate a steady-state hydrogen yield between 10 kW and 60 kW, the

reformers hydrogen yields did not deviate much more than 5% (Figure 5-14). The lowest

hydrogen concentration was slightly below 60% wet molar concentration, and

corresponded to the small reformer producing 60 kW hydrogen yield.



















1D Flo.. Changes.
Infinite: 1.10 to 6.63
Medium: 2.22 to 13.30
Short: 4.44 to 26.63


- Infinite (10kW to 60kW)


Medium (10kW to 60kW) Short (10kW to 60kW)


Figure 5-14. Comparison of transient changes in H2 concentration induced by step
changes in fuel flow corresponding to hydrogen yields from 10 to 60 kW
(lower heating value based) for various size reformers


0.66

0.65


S0.64
0Q
S0.63 -
o \
0.62

' 0.61


0.60

0.59


Run Time (s)


\















CHAPTER 6
CONCLUSIONS

Based on the experimental and modeling results presented in this study the

following conclusions may be expressed.

Steam Reformer Modeling

* A method of modeling a methanol steam reformer was proposed, demonstrated and
compared to actual reformer performance

* The reformer model results matched general reformer behavior observed in actual
reformers

* Based on this model formulation a new set of reformer similarity variables were
proposed and the use of space velocity and aspect ratio as reformer design variables
was demonstrated to be insufficient

Steady-state Reformer Operation

* Non-dimensional flow, catalyst effectiveness, and non-dimensional conduction
were introduced as potential reformer similarity variables. These variables
effectively characterized the steady-state reformer performance of reformers.

* Based on these variables, the results showed that the steady-state CO concentration
decreases with increased non-dimensional flows

* The steady-state H2 concentration decreased with increased non-dimensional flows

* Higher non-dimensional flows resulted in decreased steady-state reformer
conversion

Unsteady Reformer Operation

* Transient changes in fuel flow into steam reformers produced non-linear variations
in hydrogen and CO concentration in the reformer outlet

* These variations are mostly caused by transient changes in reformer catalyst bed
temperature









* Step increases in non-dimensional fuel flow produced a transient increase in CO
concentration followed by an eventual drop in the CO concentration

* This transient increase in CO concentration was greater when large changes in non-
dimensional flow occurred

* Step increases between large non-dimensional flows produced lower transient
increases in CO concentration than the same step changes between lower non-
dimensional flows.

The implications of the results and methodology presented in this study are that

steam reformer chemical behavior, in the transient and steady-state case, has been

expressed in mathematical, non-dimensional, terms. This leads to the proposal of

similitude expressions that can be used to better characterize, compare, understand and

design packed bed steam reformers. Also, this study demonstrates the mechanisms that

produce carbon monoxide spikes during period of flow changes.

Generally, from a design standpoint, there seems to exist a balance between the

design of an infinitely long reformer that would have excellent efficiency (conversion)

but would also have high outlet CO concentrations in the steady state and transient

regime, and a short and "stubby" reformer with low efficiency but low transient and stead

sate CO concentrations for the same fuel flow rate. The optimum point will depend on the

fuel cell and its tolerance of CO. Design should be made such that during the highest

changes in non-dimensional flow, the CO spike does not exceed a maximum

concentration that would drastically reduce fuel cell performance through poisoning.














CHAPTER 7
SUGGESTED FUTURE WORK

The analysis provided in this study is largely based on experimental results

obtained from a single diameter reformer (UC Davis reformer). The numerical models

developed generally matched the performance of this reformer. A greater body of

experimental data, with reformers of varying lengths and diameters should be obtained to

determine the practical limitations of the analysis presented here.

In addition, a set of non-dimensional performance variables has been introduced to

the field of reformer design and analysis. These variables worked extremely well at

describing reformer performance based on the models developed. A study should be

conducted to evaluate whether these variables (or a variation of these variables) are

capable of determining similarity between distinct reformers.

Expansion of this study into reformer controls design, fuel cell controls design, and

overall fuel cell system controls design is a logical expansion of this body of work.

Understanding the transient nature of fuel cell systems is critical to their effective control.

It is hoped that this work forms the foundation for the development of controllers that are

capable of producing fast transient load changes in future fuel cell systems with minimal

efficiency penalties.















APPENDIX A
COPY OF FEMLAB SOLUTION REPORT


Lull
:: ^ I


File name: C IDocuments and SettingsWAdministrator\My
Documents\Daiiel\Disseration\Reformer Model'FEMLAB\Shori Reformer'Srn all
Reformer Model Min to Max Parametricfl
Application modes and modules used in this model:
- Geoml (2D
Heat Transfer by Conduction
Confectlon and Diffusion (Chemical Engineering Module)



1. Model Properties


Property
Model name
Author
Company
Department
Reference
Saved date
Creation date
FEMLAB version


2. Constants


Name
Cc
Void
rhoc


Value






Feb 27,2005 6:56:42 PM
Feb 7, 2005 12:09:00 PM
FEMLAB 3.1.0.158


Expression
900
0.5
1983


Value
900
0.5
1983












r '-1 ii.i

q ri S. -. ip 1

MWh2o
fi'rr e..h
". 1/" :._


Mvh2


r- c:.Ii
IRh.-i



P. :h
In1, ,'. i r I


T .I II
T..,. -il l


r i.-I [ri i -_in





rm T:.


.ITI 'c 'll1r':l
I.- ni-h


11 ':;e- .
55 91C i iz.':- 1C .
2, 4 -.'.I '" 1 i
2+1,
4+1 -1:,
1 1 ,-, 2
13-1.7,
2
3. : C 4 .I e
Fu r.1 i...
Ri ru. h ,
P r.LI ; U
P r.rv..,-.-..
Pu r.T..h











I .: .1 ii i
1pi.i 1:r"' TwaII 1 3 _i iq'e 1;V
' I. 1 ",1


4 51


1,.: 171 1


1 :1
.1 .' 1c-7





44


2
:i -.51
z .: 2 4 : c -

461 917 '-
T- 66136


4157 255
:. -7811I..- 1





r 1 l





* i .i 1: -


4.51


3. Geometry


'"-JI I il:er of LIe.. i --. 1







83



4.1. Scalar Expressions

Name Expression
r 'r'-1 3 J-t-' :' T,
alpha ..lrl ._ _*Kp^2- ':lf' F r. il--.4 +Ki, 2+4 I ..+4
d :Ill.h,?-, F-, o'E ,-5"* .,,+ .' i : .i : 1))
a I+ I .:.- .:1

b d-2
c -'. .
l i,. I '.'_ ,_' .-, ._, *' I L I
x e o h I '1I:; i !i 11 .- ,-, .: 1-h 1 i n o-. ,:, II + I- 1: ,: 1:11- 1:- +: .i ,I


, ,. .,C I: '. I: t : r i.: i :. .. ..




m iI:. '.f- i r'.
rm 1 :,' :.- (T /5 1 ., ,,, ) .,.,'-l ._-,: I-:,_,'T i .. ,e -h ,'1 .
, :,:1 ,.- -.ti ,. _' a n :1
qr i i ._ s tp I'.: ir I,,, .. ',' 1 ,: ', i
r i, i, .:: r r l i ,1 il I l 1 i1 -fic h : ,l.I ,i ,,. :, II ', I ii l llr ill 1 11:. I ,i.,.1 ..., .


mn i -l ,.; -, I 5 m : r !
rni 'i'" 'in ir i *qprn ix_ **^'' .''' 1
u ln ,.,, I ; r ,.' r. ,i +'_ in*n ... .' '. 'I I.T- li'.- "'1 '.i,, i .',h


4.2. Expressions

4.2.1. Boundary E.pressic'rin,

-- ,-ijr, I ., I, 4

- I _XCO -1 .. 1 L 1] lJe1 I

..-.: xco d -,. : : ". : i d Jd
avg_ -f afwg_:' .-: _Nexit
.1 j_ Z. .-i .._ '.. ,_Nexitd
:.q 'h-:1 : :i :1- '.C : rN .itd:
avg nmeoh av _rii-.i -i:i. rI.










avg_cco2/avg_Next


4.3. Mesh


4.3.1. Extended mesh

Number of degrees of freedom

4.3.2. Base mesh

Nu m ber of boundary elements
Nu m ber of elements
Minimum element quality


12780



263
933
0.0310


4.4. Application Mode: Heat Transfer by Conduction
Application mode type: Heat Transfer by Conduction
Application mode name: ht
4,4.1. Application Mode Properties

Property Value
Default element type Lagrange Quadratic
Weak constraints Off


4.4.2. Variables


avg_xco2












_- -- ., n.1 r t r f.:-.'-.le : T
in I.: -- i-It variables: x, y, z
.-:l.- -. f, Jni li. r. i Ila .,I: ]T
Interior boundaries not active

4.4 : 3 BEound.r, .S elttirig


Type
Inward heat flux


, i 1 1 I- l


External
fer i: e,3_Fure- ; Tin
I ,- r I -
dependent
constant (,-'.:,r, :. I,
Ambient
I : I l I L. i ,i .I
,T:w& i
t rT l [.'er .i u r I ..' ,
weakconstr
I i, r f I l I i ,. -i I -.


lIr. '_Ir .1i.,ii order

In riI i l e I n- i i


1
Temperature
0


1'
Thermral ,II.: ii i~~
0


4.4.4. S-itbdoirinin Settinr z.l

,. !.,..,: Iin -. 1
1l",._ l_, Ie IIf- 10_ _1



1'll.- -i -ii.,,.'i o rd e r I'n 1. '..['. e' ,
Constraint order ".[..., ir'l-r i
T Ii: rlI- .I-I : :. I uI : i .i., I :
T I l 111 i .-,i -. I J, 1 .1 .' .' ir: 7i .,7 .:


1
:,1ll "_- ll1, T ., .. .I I ,' L l "l ,_': h i :, lI,_ -|:I

-.Mi.l :I.1 : .- I

4
2
5
: C-. ,.1i 400}


3


0


0



0



Twall
1



2


0











.1. I-.



Heat.: :41 "', r-, i,.-.)
Heat source :.Di
CcIn e:ti.e heat transfer .:-erti:ienrr
Shr .--inr )
External temperature T
User-defined constant C .r ii- )
-'Irr-l. ie- lei-rn e:.l _iitu r (Tarmbtrans)

Subdomain ri.ti lI value
Temperature iTi


ISO


rhoc
900
-rm*rhoc*2.285e8
0


0
0
0

1
Ti-


4.5. Application vMode: Convection and Diffusion

A ph. I T.l:'i mode 14.. C r -. i..'r, and Diff I i.ri (Chemical Er i r. i %e-ri. MoIdu e
;'I-.pII U l-ii mode name: cd

4 5.1. Ap:.pl: .a lon PMode PiFoperties


Property
_ '- ul' .jil *-le ,111 l..,
EC usli.:.rn I r'ni
Weak constraints


Value

rL.- ir-r 1- i. r a i r i .

I f'r


4 V5.2. variables
. Je .rer,. i ii. variables: *.: .'h' ii ..i .. ., 2 .. ,i,2
i- l:.e -die -it .Lriable:. x, y, z





4 5 3 Eo .uria i y Selt i s
Iraleio tsondrie nt ~.IlJ


I, J .* j. i
T.;-


1
' ,' i e l l ,.,


Inward flu. ir. ': .1
C ..''-, rl'il.-l.'[i '.:'0 'i {'1 ,' (y_ h2 o+ 1)'10 1 :: .. i '"'
3e31
IRlI'Tirl': y_h2o!


4
',..],f i..<-,, li',. flux


I.', .i .7. 1.,.1:


hI .- i.ll l ,ll 1 ': l lli l.l












(y_h2o+1)'101.
3c3t(Ru' Tin)' O;0:0)
weakconstr 1
f.f, ;.- functions rl


Irh"te,:.-iririn order 2
(wcgporder)
Initial value l'.. ir '. i i. ,.i. .

4,5.4, SLlbd1oirl-Ii Eri 'ttin-I

Subdomrain
' L -: .I' D 1 i.n .-n n I II C'



Inbe,-jrir.-rin order i,.r:. _-r.:;er'
C .-. n r"; 11.-,nit order : .cic: rlL, i"L
.''i ff r ,-. i'.r ,i ..ffi. ip r.- I' i .,
D iff sion r ffi. ;{1, ii I, r,- i I .



,h T r -ate IF.',



Tin 1:-l- ,i rlir um 1. .111. )
S-..i..' rr, F c


1 ,lr .:.i I t I ii i : .. ii I' :, 'i
Tuinhi1i parameter i:'.: lh id
? rr^ .,'lr, lin; "Jl lhi.: i. -.r : 4... il h f. .1- -ii-
"tre -nlilni- difft sion [ ., _- i l. '.,e
T uni i ,.:. 11.3 r 31 l n et r I." l*. .I I
,r:. '"i*.'idJ :litfutL I r .'. il:h (cdon)
Crosswind iiffi,.'- ,.r. t'~pe (cdtype)
Tuiriiin 1 i... r -ir I r 1 r (delcd)

Subdomain initial vue

:Ic,. rllerl..,r cm eoh ,',: '
, ,:, ni : ,- r'i Lr3 l, :,r- r _', I ti -i ,-.: 1


" .i i ..,",


I 1I 1 .I I


1
I Jh
I i -ii I" T Il -.: I l' .-rn ..- shlag
, ", -, .'1 ,-, : I :i .- '. : .. .' ** i" l, .,. 'c c o ')
.:. hi1l 1i: 'ch2')
44444
22222
(0:0:0;0:0}
" 1 ,I .. 1 }):;{1 ,.r .: 1 ,; I 1. 1 1 .. 1 1 [, i.
II 1 |
i ;


{'rm' Void'rhoc':'rrn Void rhoc'(a-
y_ h2o)'.'rrm'Void'r hoc b','rrn'Void'rhoc
c':'rn'Vold'rhoc'd'
" ;1 ;1 1'
{'uin':'Lin ,'Ll n'-u'in':'uin'





:i', I I' l '*








1
1 I .., 1 I I .i .. 11)
h .: h 2,,+1 1' 1 ':, -.i.. ,.'T... Ii







88


C. r .. r r- I -Ir1 C ~ 2 r ).. k_ 0
L ,-r'7.-rr ,.ri*rl.'11.i CCO i'c,.::i. 0
- .: ri.' i I .- n : .- 0



5. Solver Settings

Solke .i:.i I 1 a script off

Auto selectsover on
:_ ,, I ,'' r F I= --1 .. i .- .ri -
SI I f: .I rl : ,fl
:,'- .- -.ri, I, fr "n general
' ,i ri eIr ,. off
.,1 i|- h i'n Off


5.1. Direct (UM FPACK)
.-,i..;r r.t'r Linear sster solve r

Parameter Value
F I .'1 lir- I 1 I J 0 .1
r. lem n :.r. all ltcl.-lr factor


5.2. Nonlinear

Parameter Value
F' I .- t o e ce 1 .I.., --,
Maximum number :Ir ,IrilI. I: :'i
Dan luriii: .:1 ,I: ir, : r 1 Lar rr eret.- -It
H i-- |l,: 1 i-,,,n lin -. a ,-,-.l-ic n ,--ff
N I f I [1 I]i r 1i
Initial tI:,,i ;I., factor 1
Minimum l.,r. i iri factor 1.0e 4
FPe rir:I.I-l f. r size: S -Jl: j 11 .


5.3. Parametric

Parameter Value
Name r,1' [-ir -n1'eler C'I rr i.
L is o r. -ir r i-- I i u t: I : -
List :' r iii ei : ': ''
= r .:' l.- l. :,r 1