AN EXPERIMENT AL AND MODELING STUDY OF FUNDAMENTAL SPRAY COOLING HEAT TRANSFER FOR TERRESTRIAL AND IN SPACE CRYOGENIC STORAGE TANK CHILLDOWN By JUN DONG 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 2019
Â© 2019 Jun Dong
To my parents
ACKNOWLEDGMENTS First of all, I would like to acknowledge my advisor, Dr. Jacob N. Chung for his support, guidance, and understanding through my graduate research. I appreciate his effort to accommodate my communication skills and faith to let me lead several projects in the lab. I want to thank the other faculties on my committee, Dr. Neil Sullivan, Dr. S. A. Sherif, and Dr. Ryan W. Houim, for the effort and time they contribute d to my dissertation. Dr. Houim helped me a lot in correcting the mistake in this dissertation . Also, I would like to express my great appreciation to my graduate associate Hao Wang for his great contribution to this project. Hao spent a great amount of time assisting me to perfo rm the ground and flight tests, prepare for the flight test and perform FEA analysis for the test rig structure. Without his effort, a lot of projects in the I want to thank Dr. Samuel Darr for his help, guidance and covering during the early stage of my Ph.D. study. I also want to than k Dr. Uisung Lee who equipped me a lot of skills and k nowledge on perfo r ming experimental study. Lastly, I want to thank Neil Glikin for his great effort to the 2015 flight project. His contribution towards the DAQ system makes up the corn er stone of my DAQ system design throughout all these years.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 15 Storage and Transfer of Cryogenic Fluids ................................ .............................. 15 Motivation and Purpose of Work ................................ ................................ ............. 19 Scope of Dissertation ................................ ................................ .............................. 20 2 BACKGROUND AND LITERATURE REVIEW ................................ ....................... 23 Background of General Sp ray Cooling ................................ ................................ .... 23 Nozzle Type ................................ ................................ ................................ ..... 23 Spray Hydrodynamics ................................ ................................ ...................... 24 Boiling Curve and Chilldown Curve ................................ ................................ .. 25 Literature Review ................................ ................................ ................................ .... 27 Experimental Studies ................................ ................................ ....................... 29 Simulation Studies ................................ ................................ ............................ 33 3 SPRAY COOLING EXPERIMENT SYSTEM ................................ .......................... 36 Experiment Overview ................................ ................................ .............................. 36 Experiment Setup ................................ ................................ ................................ ... 37 Data Acquisition and Control System ................................ ................................ ...... 41 Experiment Procedure ................................ ................................ ............................ 42 Experiment Conditions and Test Management ................................ ....................... 45 4 CRYOGENIC SPRAY COOLING UNDER TERRESTRIAL CONDITION ............... 53 Spray Parameters ................................ ................................ ................................ ... 53 Spray Angle ................................ ................................ ................................ ...... 53 Quality ................................ ................................ ................................ .............. 55 Heat Transfer of Cryogenic Spray Cooling ................................ ............................. 58 Plate Temperature Profile ................................ ................................ ................. 58 Heat Flux an d Heat Transfer Coefficient ................................ .......................... 61
6 Optimization of Chilldown Performance ................................ ................................ .. 68 Enhancement by FEP Coating ................................ ................................ ......... 68 Intermittent Spray ................................ ................................ ............................. 71 Intermittent Spray o n FEP Coated Surface ................................ ...................... 76 5 CRYOGENIC SPRAY COOLING UNDER REDUCED GRAVITY CONDITION ..... 93 Reduced Gravity Chilldown Curves ................................ ................................ ........ 95 FEP Coating Enhancement in Reduced G ravity Condition ................................ ..... 96 Gravity Effect on Chilldown Efficiency ................................ ................................ ..... 97 Gravity Effect on Film Boiling ................................ ................................ .................. 99 Film Boiling HTC Correlation ................................ ................................ ................ 101 6 CONCLUSIONS AND RECOMMENDATIONS ................................ ..................... 118 Conclusions ................................ ................................ ................................ .......... 118 Recommendations for Future Investigation ................................ .......................... 119 APPENDIX A NOMENCLATURE ................................ ................................ ................................ 121 B REDUCED GRAVITY TEST DATA ................................ ................................ ....... 124 LIST OF REFERENCES ................................ ................................ ............................. 136 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 139
7 LIST OF TABLES Table page 3 1 Thermo physical properties of saturated nitrogen and oxygen at 1 atm. ............ 47 4 1 Nozzle spray angles at 80 and 100 psi tank pressure. ................................ ....... 78 4 2 The chilldown time and LN 2 consumption of all continuous two plates spray tests. ................................ ................................ ................................ ................... 78 4 3 Effect of FEP coating on c hilldown time and LN 2 consumption. ......................... 78 4 4 The chilldown efficiencies of the downward facing orientation tests. .................. 78 4 5 The chilldown efficiencies of the vertical facing orientation tests. ....................... 79 4 6 The chilldown rates of the downward facing orientation tests. ............................ 79 4 7 The efficiency of inte rmittent spray on coated and uncoated plate performed at 80 psi tank pressure by the large nozzle with a period of 1 s. ........................ 79 5 1 Tests performed in the parabolic flight. ................................ ............................. 107 5 2 Chilldown efficiencies of continuous spray tests. ................................ .............. 107 5 3 Chilldown efficiencies of the flight tests. ................................ ........................... 107
8 LIST OF FIGURES Figure page 1 1 Concept of cryogenic propellant transfer in LEO. ................................ ............... 22 1 2 Simple 3D drawing showing a typical tank chilldown by spray nozzles. ............. 22 2 1 Spray types. ................................ ................................ ................................ ........ 35 2 2 Chilldown curve (l eft) and boiling curve (right) for a typical pipe chilldown. ........ 35 3 1 Single and double nozzle configurations of the test section. .............................. 48 3 2 The schematic of fluid system. ................................ ................................ ........... 48 3 3 Exploded view of the test section. ................................ ................................ ...... 49 3 4 Locations of the thermocouples. ................................ ................................ ......... 49 3 5 CAD model of the test system (main rig). ................................ ........................... 50 3 6 Front view of the mai n test rig and gas cylinder rig ................................ ............. 50 3 7 Ba ck view of the main test system ................................ ................................ ...... 51 3 8 Electrical diagram of the test system. ................................ ................................ . 52 4 1 Photos of the spray angles .. ................................ ................................ ............... 80 4 2 Degree of subcooling of two typical tests. ................................ .......................... 81 4 3 Temperature measurements inside the test section. ................................ .......... 82 4 4 Resulted equilibrium quality when saturated LN 2 dropped to 1 atm. .................. 82 4 5 Equilibrium qualities before and after nozzle orifice. ................................ ........... 83 4 6 Mass flow rate of a typical test. ................................ ................................ ........... 83 4 7 Thermocouples and heaters location. ................................ ................................ . 84 4 8 Chilldown curves for all 25 thermocouples. ................................ ........................ 84 4 9 Chilldown curves by the average temperature. ................................ .................. 85 4 10 Comparison of the chilldown curves of individual TC and average TC groups. .. 85 4 11 Temperature distribution across the test plate by interpolation. .......................... 86
9 4 12 Heat flux again st time at seven TC group locations ................................ ............ 86 4 13 Heat flux against degree of superheating at seven TC group locations. ............. 87 4 14 Heat flux against degree of superheating of TC22. ................................ ............ 87 4 15 Heat flux against degree of superheat ing of a typical pipe chilldown test. .......... 88 4 16 HTC against degree of superheating at seven TC group locations. ................... 8 8 4 17 Biot Number against superheat at seven TC group locations. ............................ 89 4 18 Chilldown curves of FEP coated surface (up) and regular SS surface (bottom). ................................ ................................ ................................ ............. 89 4 19 A rectangular waveform of 20% duty cycle and 1 s period. ................................ 90 4 20 Mass flowrates of two pulse spray tests at the same duty cycle. A) 1 s period, B) 3 s period. ................................ ................................ ................................ ...... 90 4 21 Configuration of test section for upward facing test plate. ................................ .. 91 4 22 Normalized efficiency of the downward facing tests by small nozzle. ................. 91 4 23 Normalized efficiency of the downward and vertical facing tests by large nozzle at 80 psi tank pressure. ................................ ................................ ........... 92 4 24 Normalized efficiency of the vertical facing tests by small nozzle at 80 psi tank pressure. ................................ ................................ ................................ ..... 92 5 1 Gravity level vs. time characteristics of the parabolic flight. ............................. 108 5 2 The mass flow rate measurement of the second test in parabolic flight. .......... 108 5 3 The mass flow rate measurement of the third test in parabolic flight. ............... 109 5 4 The average chilldown curves of the uncoated plate of the third test in parabolic flight. ................................ ................................ ................................ . 109 5 5 The mass flow rate measurement of the fifth test in parabolic flight. ................ 110 5 6 A typical set of full 25 chilldown curves from a bare surface stainless steel disk for flight test 1. ................................ ................................ ........................... 110 5 7 A simplified set of chilldown curves from a bare surface stainless steel disk for flight test 1. ................................ ................................ ................................ .. 111 5 8 Average temperature curves of the first test in the flight. ................................ .. 111
10 5 9 Average temperature curves of the second test in the flight. ............................ 112 5 10 Average temperature curves of the seventh test in the flight. ........................... 112 5 11 Chilldown efficiencies of continuous spray tests. ................................ .............. 113 5 12 Measured and corrected mass flow rate of the flight test 3. A) Measured flow rate, B) Corrected flow rate. ................................ ................................ .............. 113 5 13 Measured and corrected mass flow rate of the flight test 4. A) Measured flow rate, B) Corrected flow rate. ................................ ................................ .............. 114 5 14 Nu vs. Re c . ................................ ................................ ................................ ........ 115 5 15 Nu vs. Re d . ................................ ................................ ................................ ....... 115 5 16 Schematic of the film boiling model. ................................ ................................ . 116 5 17 Model prediction compared with test data during film bo iling for T6 and T7. .... 116 5 18 Model prediction compared with test data for all film boiling experimental data. ................................ ................................ ................................ ................. 117 B 1 Simplified chilldown curves for Martian G, 80 psig tank pressure, continuous flow (Case 1), Teflon coating disk. ................................ ................................ .... 124 B 2 Simplified chilldown curves for Martian G, 80 psig tank pressure, continuous flow (Case 1), bare SS surface disk. ................................ ................................ 124 B 3 Mass flow history for Martian G, 80 psig tank pressure, continuous flow (Case 1). ................................ ................................ ................................ ........... 125 B 4 Pressure history for Martian G, 80 psig tank pressure, continuous flow (Case 1). ................................ ................................ ................................ ..................... 125 B 5 Simplified chilldown curves for micro G, 80 psig tank pressure, continuous flow (Case 2), Teflon coating disk. ................................ ................................ .... 126 B 6 Simplified chilldown curves for micro G, 80 psig tank pressure, continuous flow (Case 2), bare SS surface disk. ................................ ................................ 126 B 7 Mass flow history for micr o G, 80 psig tank pressure, continuous flow (Case 2). ................................ ................................ ................................ ..................... 127 B 8 Pressure history for micro G, 80 psig tank pressure, continuous flow (Case 2). ................................ ................................ ................................ ..................... 127 B 9 Simplified chilldown curves for micro G, 80 psig tank pressure, 40% DC, 1s period, Teflon coating disk. ................................ ................................ ............... 128
11 B 10 Simplified chilldown curves for micro G, 80 psig tank pressure, 40% DC, 1s period (Case 3), bare SS surface disk. ................................ ............................. 128 B 11 Mass flow history for micro G, 80 psig tank pressure, 40% DC, 1s period (Case 3). ................................ ................................ ................................ ........... 129 B 12 Pressure history for micro G, 80 psig tank pressure, 40% DC, 1s period (Case 3). ................................ ................................ ................................ ........... 129 B 13 Simplified chilldown curves for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4), Teflon coating disk. ................................ ................................ 130 B 14 Simplified chilldown curves for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4), bare SS surface disk. ................................ ............................. 130 B 15 Mass flow his tory for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4). ................................ ................................ ................................ ........... 131 B 16 Pressure history for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4). ................................ ................................ ................................ ........... 131 B 17 Simplified chilldown curves for micro G, 60 psig tank pressure, continuous flow (Case 5), Teflon coatin g disk. ................................ ................................ .... 132 B 18 Simplified chilldown curves for micro G, 60 psig tank pressure, continuous flow (Case 5), bare SS surface disk. ................................ ................................ 132 B 19 Mass flow history for micro G, 60 psig tank pressure, continuous flow (Case 5). ................................ ................................ ................................ ..................... 133 B 20 Pressure history for micro G, 60 psig tank pressure, continuous flow (Case 5). ................................ ................................ ................................ ..................... 133 B 21 Simplified chilldown curves for micro G, 90 psig inlet pressure, continuous flow (Case 6), Teflon coating disk. ................................ ................................ .... 134 B 22 Simplified chilldown curves for micro G, 90 psig inlet pressure, continuous flow (Case 6), bare SS surface disk. ................................ ................................ 134 B 23 Mass flow history for micro G, 90 psig inlet pressure, continuous flow (Case 6). ................................ ................................ ................................ ..................... 135 B 24 Pressure history for micr o G, 90 psig inlet pressure, continuous flow (Case 6). ................................ ................................ ................................ ..................... 135
12 LIST OF ABBREVIATIONS CHF CHV Critical heat flux Charge hold vent DC Duty cycle EDS Earth departure stage FEP Fluorinated Ethylene Propylene GN 2 HTC IHCP Gas nitrogen Heat transfer coefficient Inverse h eat c onduction p roblem LEO Low earth orbit LCH 4 Liquid methane LH 2 Liquid hydrogen LOX Liquid oxygen MAE Mean absolute error NTR Nuclear thermal rocket ONB Onset of nucleate boiling PTFE Polytetrafluoroethylene SEP Solar electric propulsion SS TC Stainless steel Thermocouple VI Virtual instrument
13 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 AN EXPERIMENTAL AND MODELING STUDY OF FUNDAMENTAL SPRAY COOLING HEAT TRANSFER FOR TER RESTRIAL AND IN SPACE CRYOGENIC STORAGE TANK CHILLDOWN By Jun Dong May 2019 Chair: Jacob N. Chung Major: Mechanical Engineering The extension of human space exploration from a low earth orbit to a high earth biggest challenges for the new millennium. Integral to this is the effective, affordable, and reliable supply of cryogenic fluids. The efficient and safe utilization of cryogenic fluids in thermal management, power and propulsion, and life support systems of a spacecraft during space mi ssions involves the transport, handling, and storage of these fluids in terrestrial, reduced gravi ty and microgravity conditions. Before these cryogens can be transfer, chilldown of the storage tank and transfer is necessary to ensure the transfer and stor age of pure liquid cryogens. Chilldown is the process of introducing the cryogen into the system to cool the hardware down to the liquid cryogen temperature. The main objective of this dissertation research is to build an accurate combined terrestrial and microgravity cryogenic chilldown heat transfer database for the fundamental study of spray cooling in the application of storage tank chilldown . Series of tests are performed at both t errestrial and reduced gravity conditions producing a broad experimental database of cryogenic chilldown through spray cooling. The
14 experiment results show the similarity between the spray cooling chilldown and pipe chilldown , and the chilldown process are categorized into vapor phase convection, film boiling, transition boiling, and nucleate boiling regimes by the heat flux behavior with respect to the wall superheat. This dissertation also investigates the methods to improve the chilldown performance and they are FEP coating, intermittent spray method, and combined pulse spray and FEP coating. These methods are effective in reducing LN 2 consumption and increasing the efficiency for a chilldown process under both the terrestrial and reduced gravity conditio ns. Based on this study of elementary spray cooling heat transfer, the film boiling heat transfer correlation and chilldown engineering approaches are available for NASA and Aerospace engineers to design the optimal in space tank chilldown processes.
15 CHAPTER 1 INTRODUCTION Storage and Transfer of Cryogenic Fluids Cryogenic fluids are extensively utilized in material processing, superconducting, and space exploration. In space application s , cryogens are mainly use d for thermal management, life support systems and most important ly as the propellants to generate power and propulsion . For example, liquid hydrogen (LH 2 ) and /or liquid methane (LCH 4 ) are used as the fuel a nd Liquid oxygen (LOX) as the oxidant for chemical rocket engines . Compared to other fuel options such as kerosene and ethanol, the combination of LOX/LH 2 offers the highest specific impulse among all chemical rocket engines. Nuclear thermal rocket (NTR) engine, another promising propulsion e ngine, also employs LH 2 as working fluid and provides the specific impulse as high as twice of the LOX/LH 2 engine can produce  . Even though solar electric propulsion (SEP) can achieve higher efficiency, the thru st it can provide is very disappointing. Due to the advantages of high thrust capability and specific impulse, LOX/LH 2 and NTR engines are considered as the solution of the main propulsion system for manned deep space exploration vehicles  . For instance , LOX /LH 2 engine power ed vehicle is expected to offer the shortest trip duration for Mars landing mission compared to other options such as NTR, Nuclear electric, and SEP Chem engine. Mercer et al.  showed that the total trip durations by means of aforementioned propulsion systems are 880, 914, 980, and 1065 days, respectively. The shorter the total time ex pen de d on a deep space mission, the less the human explosion to space environment hazards. Hence, utilization of rocket engines powered by cryogenic propellant can promote deep space exploration success and reduce human risks in a space mission .
16 However, the advantages of cryogens powered engines come at a cost, a huge amount of propellant is required to complete a mission . To satisfy this requirement, the concept of refueling a space vehicle in low earth orbit (LEO) is raised   . Figure 1 1 illustrates the concept of transfer cryogenic propellant from fuel depot to exploration spacecraft in LEO. Currently, to launch a space exploration mission, the exploration vehicle is lifted to LEO first by a launch vehicle. Then, the exploration vehicle can travel out of LEO by itself. For the space exploration vehicle fueled by the cryogenic LEO must be the propellant. For example, Hanna et al.  showed that to travel to the surface of the moon from LEO with a payload of 40 t the total amount of LOX/LH 2 need ed is about 66.9 t while loitering at LEO. cryogens can be lost due to the boil off during the Earth Departure Stage (EDS) and it is estimated that the loss can be as high as 32% of total cryogen s carried  . However, the total mass including the propellant and payloads that can be l a unched to LEO is limit ed. Thus, the additional propellant comes at the cost of less dry mass available after arriv al at LEO. In the case of insufficient propellant i n LEO, the speed of t he space vehicle can travel is limited. Thus, sufficient propellant must be available before departing LEO for a deep space mission in order to fully take advantage of cryogens powered engine. Fueling the space exploration vehicle in LEO is one feasible solution. With the ability to top off the propel lant on LEO, a space exploration vehicle can be l a unched without any propellant of its own. After reaching LEO, the vehicl e can dock with the fuel depot i n LEO and fill up its own propellant tank to further travel out earth. In
17 this way, the space explorat ion vehicle is able to carry more dry mass and more propellant for deep space travel. Another aspect of utilizing cryogenic propellant is to effectively store and transfer LOX and LH 2 during the extended time of loitering in space. In other words, only sto re and transfer these cryogens in their liquid form avoiding boil off. First of all, much less storage volume is needed if the propellants are stored in the liquid form other than the gaseous form for the same amount of mass to be store d . Because the liqui d density is much higher than that of vapor. Second, chemical propulsion engines prefer liquid fuel for reliable and predictable operation. Last ly , the propellant entering the engine is required to be liquid in order to achieve hig her thrust as higher mass flow rate can be obtained if pure liquid is transfer red compared to the case of gas being transferred . Nevertheless, the temperature of the transfer hardware which including the piping, valves, as well as the receiver tank is normally much higher than the boiling temperatures of these cryogenic fluids even in the vacuum condition of space due to radiation from celestial objects such as S un and E arth. For instance, the thermal radiation from the Sun and Earth can keep the space vehicle warm from 250 K to 350 K. Nonetheless, the boiling temperatures of LH 2 and LOX are only slightly higher than their saturatio n temperatures, 20.3 K and 90.3 K respectively at 1 atm. If these low temperature liquids were introduced into the relatively hot hardware, the y will boil off immediately and become a mixture of liquid and vapor or completely gaseous. Since cryogenic prope llants can only be stored and utilized in the form of liquid and there is no effective way of separating liquid from vapor in space, the gaseous propellant and two phase The single
18 phase liquid transfer and storage can only happen when the hardware is c ooled well below the temperature of boilin g can happen. Thus, the hardware of transferring and storing cryogenic propellant s must be drastically chilled before the fuel can be transferred to space exploration vehicle from a n LEO orbiting fuel depot. The current work focu se s on a series technologies of cryogenic fluid transport , which will support the development of cryogenic propellant transfer scheme as well as giving estimations of the cost of time and amount of cryogen required for this process. This process of cooling the transfer hardware is called chilldown or quenching, which is a complex transient process involving two phase heat and mass transfer. To successfully transfer liquid cryogens from a supply tank to a receiver tank, all the hardware that the fluids flow through must be quenched, which includes straight tubes, tube fittings, control valves, nozzles, and the receiver tank. Since the geometri es and thermal mass of these fluid components are significantly different from each other , distinctive schemes are adopted to chilldown these subjects. For example, the tubes, tube fittings, and control valves are simpl y quenched by introducing cryogenic f luid flow through them. On the other hand, the relatively large tank is quenched by spraying cryogen onto the inner wall of the pressure vessel through an array of spray nozzles or spray bars . Figure 1 2 gives the configuration of nozzles for a typical tank chilldown by spray cooling. Although these components are quenched mainly by two different scheme, the behavior s of the heat transfer are similar. According to Shaef fer et al.  , the total time and cryogen required during a chilldown operation are spent mostly during the film boiling heat transfer regime between the cryogenic fluid and heat transfer component .
19 To achieve the goal of efficient, predictable, and reliable transfer of cryogenic propellant in space, the chilldown process in reduced gravity must be studied and understood. This dissertation focus es on the chilldown of a receiver /storage tank in both te rrestrial and reduced gravity environment . Motivation and Purpose of Work A series of ground tests and modeling work regarding to tank chilldown have been performed since 1967   . Various tank chilldown schemes were investigated in these preceding studies using a sizable aluminum tank as the receiver tank and LH 2 as the working fluid. The receiver tank can be q uenched by continuously spraying LH 2 from the top of the container. Impinging the LH 2 jet onto the inner wall of the tank is another viable method. More recently, a charge hold vent (CHV) tank chilldown procedure has drawn increasing attentions from NASA r esearchers. CHV consists of charging a small amount of liquid into the tank, holding the vent valve closed to allow the liquid to completely vaporize, and then opening the vent valve to relieve the pressure. This three step procedure is repeated until the tank wall temperature is sufficiently chilled down. However, only two sets of g round tests have been carried out that show the feasibility of a charge hold vent (CHV) procedure ,  . Due to that the gravity level plays an important role in two phase heat and mass transfer phenomenon, the results of these studies cannot accurately depict the real tank chilldown process in a weightless environment such as in space. The present work cover s this gap in the literature by experimental investigating cryogenic spray cooling in both reduced gravity and terrestrial conditions. T his work adopts a fundamental approach that a circular stainless steel (SS) plate is quenched by spray cooling instead of quenching a real tank . This is because that cryogenic chilldown by spray cooling is
20 the basic foundation of the cryogenic tank chilldown process regardless of the specific shape of the subject. The advantages of this fundamental case study are given as follows. First of all, the heat transfer mechanisms are the same for those two approaches though the shape of a plate is different from the cylin drical surface of a tank. Thus, the experimental data can be used to estimate the required cryogen consumption and cost of time for a real tank chilldown case. On this bas is , a more realistic design of a real tank chilldown system can be achieved with the knowledge of the fundamentals of cryogenic spray cooling . Second , the resulted data does not limit itself only to the application of tank chilldown but can be extended to other applications where drastic cryogenic cooling is required . Third , the base case test can be designed very flexible such that it can be tested at various test conditions. Plenty of spray cooli ng schemes and conditions are tested by this work such as the gravita tional acceleration, mass flux, coating enhancement, and intermittent sprays. To sum up, the presented work experimentally investigated the fundamental case of cryogenic spray cooling by a circular SS plate in one g and simulated reduced gravity environment at series of test conditions, which will give insight into the method of optimizing space cryogenic propellant transfer. Scope of Dissertation In Chapter 2 , the background of spray c ooling is introduced shortly including the category of nozzles, the hydrodynamic parameters of the spray, and the general heat transfer modes of a chilldown process. The n the previous experimental and simulation studies of the spray cooling are reviewed. T he differences between previous experimental work and current work are outlined. In Chapter 3, the design of the new experimental system and the procedures of performing the test during the flight are
21 presented. And test conditions as well as the test mana gement scheme are also given in this chapter. Chapter 4 gives the ground test results of the base case cryogenic tank chilldown by spray cooling. The heat transfer data and the parametric study of the spray cooling performance are discussed. The experiment data of the fundamental case of tank chilldown achieved under reduced gravity condition is given in chapter 5. In addition, a heat transfer model to predict the heat transfer coefficient for film boiling mode is given. Chapter 6 concludes this dissertatio n with a summary of the presented work and the recommendations in the future study.
22 Figure 1 1 . Concept of cryogenic propellant t ransfer in LEO . Figure 1 2 . Simple 3D drawing showing a typical tank chilldown by spray nozzles.
23 CHAPTER 2 BACKGROUND AND LIT ERATURE REVIEW Background of General Spray Cooling Nozzle Type Spray nozzles are used to disperse liquid into spray, which offers three major benefits: increased liquid surface area, wider distribution of liquid over an area, and increased impact on a solid surface. According to the performance and atomization mechanisms, the spray nozzles can be identified as two fluid nozzle, pressure nozzle, ultrasonic nozzle, and rotary nozzle. However, only the two fluid nozzles and press ure nozzles are normally used in spray cooling  . Two fluid nozzles or atomizers utilize a secondary high speed gas stream usually compressed air to break up the liquid stream and produce fine droplets. Dependi ng on mixing and external mixing nozzles. Since an additional gas stream is required for the two fluid nozzles which increase the chilldown omizers are not considered suitable for space applications. The pressure nozzles, however, achieve the liquid stream breakup solely by the momentum of the flow. The liquid is shattered into droplets by forcing the flow through the small orifice of the pres sure nozzle. If the pressure drop across the orifice is high enough, fine droplets size can be produced. Based on the distribution of the droplets sprayed on a surface, the sprays can be further divided into full cone, hallow cone, and flat fan  . The liquid emitted from the full cone nozzle makes up a conical shape in space and creates a circle of spray with liquid present throughout when meeting a surface. On the contrary, the hollow cone spray nozzle distribu tes most of the droplets
24 near the rim of the spray circle. A flat fan nozzle create s a sharply defined fan spray and the impact area usually takes the shape of an oval. Since the liquid is concentrated into a smaller area, the droplets impact is generally greater than that of a full cone or a hallow cone nozzle. Full cone pressure nozzles are preferable for the tank chilldown in a space condition due to the simplicity and the large area that a single nozzle can cover. One thing worthy of noting in cryogenic chilldown is that before the chilldown of the piping and pressure spray nozzles, the mixed vapor in the flow may assist the breakup of the liquid emitted from the orifice, which is the characteristic of a two fluid nozzle. Spray Hydrodynamics Sprays ca n be defined as dilute, intermediate, or dense spray based on fluid coverage on the spray surface. For a dilute spray shown in Figure 2 1 a, there is almost no liquid buildup on the spray surface and the interaction between droplets is low. On the other ha nd, a continuous liquid film forms on the spray surface in the case of a dense spray given in Figure 2 1 c, where the volumetric flux is high. In between is the case of an intermediate spray depicted in Figure 2 1 b , where a thin liquid film may partially cover the surface due to interactions between droplets but the liquid film will not continuously cover the whole spray area. To model the heat transfer behavior of spray cooling the surface fluid coverage conditions must be carefully differentiated. In the case of a dilute spray, the heat fluxes can be modeled as the sum of all the heat fluxes between the spray surface and individual droplets, while a continuous convective boiling model might be suitable for a dense spray. Typically, sprays are consisted of a wide range of droplet sizes. There are two commonly used definitions of droplet size distributions. One is the mass median diameter, and the other is Sauter mean diameter. The mass median diameter, d 0.5 , is
25 def ined as the diameter at which 50% of the droplets by mass are larger and 50% are smaller. Sauter mean diameter, d 32 , is defined as the droplet diameter that has the same volume to surface ratio as the entire spray. , ( 2 1 ) Where n i is the number of droplets with diameter d i . Among all the parameters of a spray, the mass flux is certainly the most critical one affecting the heat transfer performance. Three different types of mass flux are referred to in the literature. The most accurate definition of the mass flux is the local mass flux, which is the mass flux impacting on an infinitesimal area of the surface. The less accurate mass flux is the mean mass flux. The mean mass flux is the total mass flow rate of a spray divided by the area covered in the spray cone. The least accur ate mass flux is cal culated simply by dividing the total mass flow by the whole area of the heat transfer surface. Boiling Curve and Chilldown Curve The characteristics of heat transfer response of a surface to cryogenic spray chilldown can be described b y the boiling curve and/or chilldown curve. Figure 2 2 gives examples of a typically chilldown curve and boiling curve for a pipe chilldown process. In the chilldown curve , the temperature history at a certain axial location of the pipe is plotted as a function of time. The heat flux versus wall temperature is given by the boiling curve. Compared to a conventional boiling process which is controlled by the heat flux fr om external energy input, a chilldown process is transient and controlled by the surface temperature as well as the thermal mass of the heat transfer component. In chilldown , the only available heat that can be transferred to the fluid is from the stored
26 t hermal energy of the hardware. Consequently, the temperature of the heat transfer medium will drop continuously until reaching the fluid temperature while the heat transfer mechanism changes from four distinctive modes through this process. Cryogenic chill down is usually initiated with film boiling, which features a relatively low heat transfer rate between the fluid and solid surface. During film boiling, owing to the high degree of the wall superheat , the liquid cannot wet the solid surface and a continuo us vapor film is formed covering the heat transfer surface. The heat is first transferred to the vapor film and then absorbed by the liquid on top of the vapor. The film boiling ends when the solid surface is cooled to the rewet temperature, below which th e liquid can then touch the solid surface and nucleation is going to take place on the solid surface resulting a much higher heat transfer rate. After rewetting , the heat transfer goes into the transition boiling regime, where the heat transfer r ate increa ses rapidly and peaks at critical heat flux (CHF) point. As the cooling of the wall continues, the heat transfer rate at the surface would start to decrease as the heat transfer goes into the nucleate boiling regime. Nucleate boiling would continue until t he surface is cooled below the temperature of onset of nucleate boiling (ONB). After the cessation nucleation, the heat transfer is governed by single phase liquid convection without phase change taking place. Since the heat flux during the period of trans ition and nucleate boiling is extraordinarily greater than that of film boiling, the temperature curve plunges after rewetting point and resembles a vertical line. After nucleate boiling , the surface temperature does not drop significantly. Thus, film boil ing makes up the majority of a chilldown process in terms of time. On the other hand, a conventional boiling process is well kept at nucleate boiling regime where the heat flux is below CHF point,
27 to achieve maximum heat transfer rate and avoid the burn ou t of the heat transfer component. To sum up, cryogenic chilldown process features its changing heat transfer modes, large temperature variation, and low heat flux in the film boiling stage. Literature Review In gen eral, spray cooling is a popular topic c apturing great interest from the scientific communit y but also a general term referring to a method of cooling an object by spray ing cold fluid on its surface . Thus, even though numerous studies were conducted in this area, not every aspect of the spray co oling is covered by these studies. Based on the requirements of the applications, spray cooling has a very wide range of operating conditions consisting of different combinations of aforementioned heat transfer modes and spray types . For example, to cool d own an electronic device a dense spray is commonly used and the heat transfer mode is well controlled in the single phase liquid convection or nucleate boiling regime. However, to quench a metal part during heat treatment , the heat tra nsfer mechanism goes through all the possible modes described previously with film boiling taking up the most of the process. I n the application of cryogenic surgery on human skin, a concentrated and jet like spray is desired whereas full cone spray s with a large spray angle a re preferable to cool down a large surface. The application of spray cooling for in space cryogenic tank chilldown process requires following constraints: 1) the working fluid must be among the cryogenic fluids; 2) a dilute spray; 3) reduced gravity level; 4) transient heat transfer with film boiling taking up most of the p rocess. Indeed, many experimental results studying spray cooling are reported in the literature . However, none of existing experiment al data is useful for the spray cooling application t hat the proposed study aims at . First of all, the working fluids used by previous researchers are water, FC 72, FC 77, R 22, R -
28 134a and so on but no experiment s using cryogenic fluid are reported. Second , little effort was made towards the study of heat tr ansfer by a sparse spray. Third, most data is derived from the experiments using a heated surface with a constant heating rate , which implies a steady state heat transfer behavior . Last, though a handful of spray cooling experiments in reduced gravity condition through parabolic flight are available, the failure of satisfying the other requirements render them unsuitable for the proposed application . The coupling of the two phase heat transfer and spray hydrodynamics gives rise to th e complexity of modeling the heat transfer by spray cooling. M any factors are identified to play an important role in the performance of spray cooling . Kim  gave a brief review of the experimental studies of spray cooling for electronic cooling applications. It identifies the research interests of spray cooling as heat transfer mechanisms, heat transfer on enhanced surface, effect of spray hydrodynamics , and gravity effects. Liang and Mudawar ,  gave two very detailed and comprehensive review papers of spray cooling. Liang and mudawar divi ded th e spray cooling into two categories : the low temperature application and high temperature applicatio n . The fi rst paper is devoted to the low temperature spray cooling and the second paper covers the high temperature spray cooling . The low temperature spray cooling is defined as the left part of the CHF point o n a boiling curve and it includes single phase liquid convection, convective nucleate boiling as well as CHF point which serves as the boundary of these two cases. The high temperature spray cooling starts at the right of the CHF point and t owards the film boiling regime o n a boiling curve. The high temperature spray cooling include s transition boiling, rewet temperature, and film boiling. By reviewing the
29 experiment data and proposed model s available in the literature, Liang and Mudawar  summarized : contradictory con clusions drawn among researchers are mainly due to incorrect use of characteristic length when defining the Nusselt number, inability of assessing the heated surface temperature distribution, the com plexity of spray impact mechanism. Besides, heat flux measurement/calculation method and non uniformities in surface temperature distribution are not well addressed in most spray cooling literature. Experimental Studies Kato et al.  studied the gravity effect on liquid spray cooling using a nickel plated copper block in terrestrial and vari able gravity condition s . The vari able gravity condition was simulated through parabolic flight. The copper block was firs t heated by seven cartridge heaters embedded in the block to a prescribed temperature and then cooled down by spraying liquid onto the nickel plated surface which is only 19 mm in diameter. The working fluid s used were water and CFC 113. They observed that the heat transfer in the low heat flux regime below the CHF w as enhanced b y the reduction in gravity for both fluids. However, the effects of reduced gravity act differently on these two fluids at CHF. The CHF for CFC 113 was decreased in low gravity level whereas the CHF of water increase d . Kato et al. also reported the vanishing of the gravity effect at high spray volume flux. As a follow up study , Yoshida et al.  conducted a more com prehensive study of the effect of gravity on spray cool ing . Two different heaters were used in this study. One is similar to the copper block described by Kato et al  except that the surface was plated by chromiu m and 50 mm in diameter. The other was a glass prism plated with a thin transparent indium tin oxide (ITO) film as the heating element. This transparent glass heater was used in order to visualize the liquid deposition on the
30 heater surface for steady stat e spray cooling while a copper block was used for transient spray cooling test. The working fluids used were water and FC 72. Series of ground based tests and parabolic flight tests were performed by varying the test parameters such as working fluid, heate r surface orientation, heat flux at heater surface, mass flux of the coolant as well as heater types. They reported that gravity level had little effect on the nucleate boiling regime . Moreover, they suggested a coupled effect of gravity and spray volume flux on CHF. In the case of low spray volume flux, neither the magnitude nor the direction of gravity affects CHF. However, the CHF under reduced gravity is higher than the CHF under hyp ergravity by 10 percent. They also noted the significant influence of gra vity on the film boiling regime when We bber number is low. And they argued that the deterioration of the heat transfer during the film boiling in the case of low We bber number and red uced gravity condition is due to a lack of secondary impingement on the heater surface. Michalak et al.  investigated the effects of gravity levels on the performance of a closed loop partially confined spra y system. In this experiment, the variable gravity condition was created through the parabolic flight of a KC 135 aircraft. The gravity level was between 0.15 and 1.8 g. FC 72 was sprayed onto a resistor heater inside o f a test chamber and circulated in th e fluid loop by a pump. One thing worthy of noting from their experiment is that during the reduced gravity period, the mass flow rate is significantly decreased due to the pump drawing vapor from the fluid tank, which caused sever e overshoot of th e measur ed temperature. They foun d that the wall superheat decrease d as the gravity level decrease d for given heat flux. Besides, higher degree s of the coolant subcooling would also lower the wall superheat for a given heat
31 load. For all acceleration levels, the increase in volumetric flow rate would always lower the wall superheat. However, the boiling curves plotted by these researchers d id show a distinctive heat flux jump between single phase convection regime and nucl eate boiling regime. Since no effort was no conv incing evidence that can help determine the heat transfer mode of their experiment. Roisman et al.  studied the gravity effect on the liquid film produced by spray ing liquid on a heated surface and the spray cooling in both terrestrial and reduced gravity conditions. The water wa s sprayed on a convex shaped heated surface which was made of a truncate d sphere by a full cone pressure swirl nozzle. The diameter of the projected curved surface wa s 20 mm. Images were collected together with the temperature measurements allowing the study of liquid film thickness during spray cooling tests . They found out t hat the average film thickness increase d as the gravity level decrease d and postulated that this could be a reason to explain the decreasing performance of spray cooling in a reduced gravity condition. Besides, they argued that not only the gravity level b ut also the curvature of the impacted surface affect ed the liquid film thickness. A higher radius of curvature of a targeted surface would yield a thicker liquid film . They also observed an optimal liquid volumetric flow rate for spray cooling and identifi ed this value as 0.3 l/ min for water. Somasundaram and Tay  evaluated the performance of intermittent spray cooling by spraying water on a heated copper block. The area of the impact surface on the co pper block is 25 X 25 mm 2 . The spray cooling wa s controlled by closed loop feedback based on the measured temperature and set temperature values . The set
32 temperature values were chosen to be 5, 10, and 15 Â°C above the steady state temperature of a continuous spray at a given heat input. The water wa s sprayed on the heated surface when the measured temperature is 0.5 Â°C above the set temperature values and the spray wa s stopped wh en the measured temperature is 0.5 Â°C below the set temperature values . It was found that the surface temperature fluct uations increase d as the heat flux was increase d at low heat flux region and decrease d as the heat flux was increase d at high heat flux r egion. The maximum temperature fluctuation wa s reached at a moderate heat flux. In addition, the difference of spray efficiency between continuous spray and intermittent spray wa s negligible if the surface temperature wa s low corresponding to a low heat fl ux input. The spray efficiency of intermittent spray was only significantly higher than that of a continuous spray when the surface temperature is close to boiling temperature. Based on the reported temperature range by Somasundaram a nd Tay, 40 to 100 Â°C, it is likely that the majority of their spray data only corresponds to the single phase liquid convection heat transfer regime. Aguilar et al.  investigated the evolution of a single droplet emitted from a straigh t tube nozzle by conducting an experiment to measure the velocity, diameter, and temperature of droplets coming out of the nozzle. R 134a was used as the working fluid. They reported an extended region near the nozzle tip where droplet coalescence and rec irculation were observed. A theoretical model wa s raised to match the measured data from the experiment. The evolution of a single droplet wa s modeled by using conservation law of mass, moment um , and energy. However, no measurement of droplet velocity, diameter and temperature w ere available at the tip of the nozzle due to little atomization at the tip of the nozzle. Thus, the conservation equations were actually
33 solved by guessing the initial va lues of the droplet velocity, diameter, and temperature. By varying the initial guess es , the model can be best fitted to the experiment al dat a. C omparing the result of the model with convection effect with that of the model that does not take account of co nvection, they concluded that the droplets absorb a significant amount of heat from the surrounding air before impacting on a surface. Takeda et al.  experimental studied the Polytetrafluoroethylene (PTFE) c oated pipe chilldown process using LN 2 . The test results of uncoated pipe and three other pipes coated by PTFE of thickness 23 Âµm, 63 Âµm, and 91 Âµm are compared. It was found out that the temperature of minimum heat flux point is increased for the coated p ipes compared with the uncoated one, which promote d the transition of film boiling to nucleate boiling. As a result, the chilldown time is significantly decreased. In addition, the total amount of LN 2 can be saved by applying the PTFE coating is as high as 64%. They believe the re is a significant temperature gradient inside the PTFE layer due to the low thermal conductivity of PTFE . Thus, the coated surface contacting the fluid can be quenched to the t emperature at which liquid can contact the surface faster resulting in the earlier transition to nucleate boiling. Simulation Studies Selvam et al.  simulated the vapor bubble growth and droplet impact on thi n liquid film formed by liquid spray using the level set method. A 40 m thick liquid layer with one growing vapor bubble due to heterogeneous nucleation wa s created in a two dimensional planar area. Since the governing equations wer e non dimensionalized , t he thermo physical properties of a real fluid were not used. Instead , the following dimensionless numbers were used l v =20, Re=200, We = 1, Pe=1000 and Ja=0.1.
34 The properties of liquid FC 72 were used as reference values. According to Selvam et al., using the low density ratio of liquid to vapor phase results in a reduction in computation time and avoid ed numerical instability issues. However, the density ratio of liquid to vapor phase of a real commonly used fluid is usually higher than 100 , which im plies the simulation results may not represent the two phase interaction in spray cooling . First, the process of a vapor bubble grow ing from the heater surface until co llaps ed wi th vapor above the liquid film wa s simulated. They found a sharp increase in t he maximum velocity in the computation d omain as well as a peak in the plot of average Nu number when the bubble merge d with the vapor on top of the liquid film. They postulated that this increase in heat transfer wa s due to the enhanced liquid motion when liquid is spreading to the dry area previously occupied by vapor bubble. A impact on the thin liquid film wa s simulated as the second case. They observed a gradual increase of Nu sselt number after the collapse of the vapor bubble due to d roplet impact and the average Nu sselt number remained high for about 173.2 microsecond.
35 Figure 2 1 . Spray types. Figure 2 2 . Chilldown curve (left) and boiling curve (right) for a typical pipe chilldown .
36 CHAPTER 3 SPRAY COOLING EXPERIMENT SYSTEM This c hapter presents the design of the experiment al system to characterize the heat transfer of the base case ta nk chilldown process by cryogenic spray cooling. In this experiment , liquid nitrogen is sprayed onto a stainless steel disc, and the evolution of the disc temperature s at several locations is measured and recorded. The experiment system is designed in a fashion that it can be carried onboard a Boeing 727 200F aircraft such that this experiment can be performed in a simulated reduced gravity environment facilitated by the parab olic flight trajectory maneuver of the aircraft. Experiment Overview Stainless steel disc is chosen for the base case mainly because its simplicity and resulted heat transfer correlation s can be applied to the case of real tank chi lldown without significant errors . If a cryogenic storage tank is used for the purpose of storing propellant in space, the volume of this tank will be inevitably large to st ore the required amount of fuel. But the number of sp r ay nozzle s can be install ed inside the tank must be limited to avoid extra mass to be launched and unnecessary thermal mass to be quenched. Thus, a single spray nozzle must cover a relatively large area of the cylindrical ta nk body. Since the curvature of the cylindrical surface tha t a single nozzle covers is relatively small, the heat transfer correlation s from a flat surface will not deviate significantly from the case of a slightly curved surface . Besides, spay cooling itself is a very popular research topic which has a broad appl ication in material processing, cryogenic surger y, and electronic packaging cooling. Moreover, the experiment al set up is more flexible if a SS disc is used other than a real tank. Various nozzles with different flow capacities, mean droplet sizes, and spr ay distributions can be
37 easily varied on the current experiment set up. The spray angles of the nozzle can be determined by visualization which only requires minim al modification to the current set up. There are two configurations of the test section , see Figure 3 1 . The first one is the single nozzle configuration. In this case, only one nozzle is installed in the test section and a substantial amount of ground tests are performed by varying the orientations of the nozzle , tank pressure, and spray methods. The second case is the double nozzles configuration. Two nozzles are installed symmetrically in the test section and it allows the simultaneous chil ldown of two different plates in one test. This configuration is used to study the effect of the Teflon coating in terrestrial condition and for the parabolic flight tests. In this way, each set of flight test data includes the results of the chilldown of two test plates. Liquid nitrogen is chosen as the test fluid for certain reasons, especially owing to its similar thermo physical properties to liquid Oxygen when neither LH 2 nor LOX is not an option due to safety considerations . Table 3 1 gives the thermo physical properties of LN 2 and LO X . The thermo physical properties of nitrogen fall within 20% of the se of oxygen. Besides, LN 2 can be easily purchased at a n affordable price from Airgas Inc . Experiment Setup A schematic of the experiment system is given in Figure 3 2 . LN 2 is supplied to the system from an 80 L cryogenic liquid dewar with a relief valve set at 125 psig . A gaseous nitrogen cylinder initially at 2000 psig pressurize s the dewar to a set value for each test, which ranges between 6 0 and 100 psig . A pressure regulator is used to ma intain the dewar pressure at a preset value, which can fix the dewar pressure to within 5 psi of the set value. Before the liquid nitrogen can be sprayed on the SS disc inside the test section, it has to go through a pre cooler and flow meter. The flow
38 coming out of the test section must go through two vaporizers configured in parallel that ensures pure gas phase of the flow before it can be vented overboard . For the flight tests, the gaseous nitrogen is vented through three venting port s in the aircraft conne cted by three rubber hoses. A pre cooler is essential ly a tube shell heat exchanger. It serves two purp oses. The first one is to cool the liquid nitrogen slightly below its saturation temperature such that the liquid going into the test section is su bcooled. Secondly, it works as a liquid vapor separator to separate LN 2 from GN 2 . The liquid nitrogen flow s through a finned tube inside the shell of the pre cooler, which is submerged in a bath of liquid nitrogen. Since the pressure inside the fin tube is much higher than the pressure inside the shell, the liquid temperature inside the tube is higher than the temperature of the LN 2 bath. As the LN 2 flow s through the tube, the heat is transferred from the LN 2 inside the tube to the LN 2 bath. As a result, th e LN 2 in the tube side is cooled while the LN 2 in the shell side evaporate s and is vent ed out of the system from the vaporizer 1. To reduce the parasitic heat flux from the environment into the SS tubing between the test section inlet and cryogenic dewar , a cold shield made by SS tubing and hoses are wrapped around the fluid pass and flow meter. Before each test LN 2 flow s through the cold shield to cool down the tubing to the temp erature close to the saturation temperature of liquid nitrogen and eventually into the precooler, where the liquid can be separated from the vapor by gravity. After the chilldown of the cold shield, the precooler can be filled. Two thermocouple (TC) probes are inserted into the precooler to measure the temperature s inside the precooler which can be used to indicate the liquid level of the LN 2 bath. The pressure and temperature of the fluid inside the tube are also monitored.
39 A Coriolis flow meter (CMF025M from Micro motion) is installed to measure the flow rate of LN2. The flow meter is cryogenic ally rated and can accurately measure the flow rate with in 0.35% of total flow rate as long as the fluid goes through the flow meter is pure liquid . A total of two thermocouples are attached to the flow met er to monitor the temperature outside the flow meter shell . However, the electronics of the flow meter is not rated for cryogenic application and gives out error signal if it is chilled below 120 K. So, the flow meter is always kept between 120 and 140 K in the tests. The test section is a SS container enclosing the spray nozzles and test discs. It is essentially an off the shelf cubic vacuum chamber. The explode d view of the test section is give n in Figure 3 3 . Two SS discs can be mounted at any sides of the cubic chamber , for example , the front and back or top and bottom . Ea ch disc is sandwiched by two bored vacuum flanges. Two PTFE gaskets are compressed against the SS disc such that the inside of the chamber and the back of the disc can be sealed. The outermost flange on each test disc assemble is connected to the vac uum pu mp such that the back of the SS disc is insulated by vacuum to minimize the parasitic heat flux from the environment . For measuring the disc transient temperature history during the chilldown, 25 thermocouples are soldered to the vacuum side (back) of each SS disc. As shown in Figure 3 4 , a total of 23 TCs are distributed on 6 concentric circles (6 rings, in addition to one (TC25) at the center and one (TC5) on the outside near E ight MINCO film heaters (Hap 6945 and 6946) are attached to the back of the disc to heat up the plate. The temperature and pressure of the fluid are measured before the i nlet of the test section. The pressure inside the test section is measured by a pressure transduce r
40 located at the outlet of the test section. There are two solenoid valves (SV1 and SV3) placed at the inlet and outlet of the test section res pectively. The opening and closing of both valves are controlled by Labview virtual instrument (VI) . SV1 can be progra m med to be open ed continuously or intermittently , whereas SV3 is only opened if the LN2 is sprayed inside the test chamber. Depends on the test requirement, different nozzle configurations can be implemented in the test section , as shown in Figure 3 1 . Only one nozzle is installed in the test section for the case of the spray cooling in the terrestrial condition. To study to the effect of FEP coating on spray cooling cryogenic chilldown , t wo identical nozzles are installed. One nozzle spray s LN 2 on a regular SS plate. The other spay s LN 2 on a Teflon coated SS plate. Since the fluid paths coming out of the two nozzles are exact ly the same, the mass flow rate, the pressure, and the liquid tem perature can be considered the same for these two sprays. To study the effect of mass flux on spray cooling, two nozzles with different nomina l flow rates are tested in the terrestrial condition, and they are 1/8WL 3/4 90 and 1/4WL 1 1/2 90 from Bete fog n ozzle. These two nozzles are the WL series whirl low flow full cone nozzles. The first nozzle is made fraction number in the nozzle number is the pi pe connection size . denote the nozzle series. The second fraction number is the nominate flow rate in GPM at 40 psi for water. The last number is the spray angle of the nozzle in degree when water is the working fluid. To simplify the notations of these two nozzles , the nozzle with a smaller nominal flow rate will be called small nozzle and the nozzle with a higher
41 nominal flow rate will be called large nozzle. Only the small nozzles are used for the tests performed in the reduced gravity condition. There are three vaporizers in the experiment system . Two are in parallel at the downstream of the test section. One is downstream of the shell side outlet of the precooler. As required by the flight provider that no liquid shall be vented overboard, vapori zers are used to vaporize any liquid nitrogen coming out of the test section and precooler and then warm up the nitrogen gas to above 40 Â°C, below which the rubber hose s will become brittle and easy to crack . Each of the vaporizers is a shell and tube bun dle he at exchanger and is heated by a heating tape wrapped around the outer shell. All the fluid components , as well as the electrical components, are housed by 8020 aluminum structure as shown in Figure 3 5 . The nitrogen gas bottle is housed in another 8020 structure, which is not shown in the CAD model. For the flight tests, the test rigs are brought onboard the aircraft and secured to the floor attachment by both bol ts and straps. The photos of the main test rig and gas cylinder rig are given in Figure 3 6 and Figure 3 7 with the major components labeled by the yellow arrow. Data Acquisition and Control S ystem In this presented experiment the data acquisition and control system makes up the whole electrical system for the purpose of data measurement and system control , see Figure 3 8 . Electrical diagram of the test system. . The control system is designed to minimize t he actions need to be taken during a flight test. For example, the temperature all controlled through the progra m ming in the Labview VI. T here are a total of seven dir ectly measured data and they are gravity level, LN 2 flow rate, LN 2 temperature and pressure at the inlet of the test chamber, the pressure inside the test chamber,
42 temperature s of the test plate at vacuum side, and the temperature of the nozzle. Type T the rmocouples (Omega), both the thermocouple wires and profiles, are used for the cryogenic temperature measurement. The gauge size of the TC wires is 30 AWG and is insulated by Neoflon. Type K thermocouples (Omega) are used for monitoring the heaters tempera tures. Two different pressure transducers are used in the whole system. The inlet pressure is mea sured by the pressure transducer made by Omega Model number PX409 150A5V. Since this transducer is not rated for cryogenic temperature application, extension tubing is added between the measurement point and transducer to protect the pressure transducer from LN 2 . The test section pressure is measured by Kulite pressure transducer model number CTL 190M 140BARA . Triple Axis ADXL 335 accelerometer from Sparkfun i s used to measure the acceleration. Five NI 9214 thermocouple input modules are used to read the T type TC signals and the voltage signal s of Kulite pressure transducers. One NI 9205 analog input module is used to measure the voltage signal s from the Omega pressure transducer, Accelerometer, and Coriolis flowmeter . NI 9211 thermocouple input module is used to read the K type TC signals. NI 9472 digital output module is used to control the solenoid valves, tape heaters, film heaters. All these modules are ho used by NI cDAQ 9178 compact DAQ chassis. NI USB 6009 is used to gives rectangular wave form signal to the solid state relay achieving the intermittent spray. Experiment Procedure This experiment is performed both in terrestrial and reduced gravity (by par abolic flight) conditions. Since the test procedure for flight test is essentially the same as the ground test but with increased time limitations , only the test procedure for the flight test will be discussed. Before taking off , a few steps need to be car ried out to make sure
43 there is enough fluid to complete a certain set of tests during the flight test. Both the gas nitrogen cylinder and liquid nitrogen dewar need to have nitrogen fluid at their full capacity. For the nitrogen gas cylinder, this is simpl y done by putting a fresh nitrogen gas cylinder in the test rig and connect the cylinder to the fluid system (the pressure regulator). To fill up the liquid nitrogen dewar, the procedure goes as follows. First, one 180 L liquid nitrogen dewar from Airgas n eed s to be broug ht close to the test apparatus from the 180 L dewar . Second, the hose is connected to the fill port and GV4 is fully opened. Third, the filling process is initiated by o pening GV2, GV3, and the liquid valve on the 180 L dewar. As the 80 L test dewar being filled up with liquid nitrogen, the level indicator on the dewar will go up. Third , the GV2, GV3 and the liquid valve on the 180 L dewar are closed to stop filling proce ss when the level indicator shows that the 80 L dewar is full. Last, the 180 L dew a r is removed by detaching the fill hose from the test rig and closing GV4. After this point, the test system is ready for the flight test. After taking off and when the aircraft reaches the designated airspace for the parabolic maneuver, the experimenters can leave th eir seats and prepare the tests . There are 15 minutes for the preparation of the test before the first parabola begins . During these 15 minutes preparation, steps are carried out as follows. First, the data acquisition system is turned on. This is done by simply turning on the laptop and running the Labview VI program . Second , set fluid valves across the fluid system to the test condition. The valves on the ga s cylinder and liquid nitrogen (GV1 to GV3) are turned to open. The pressure regulator on the gas cylinder is adjusted to the set pressure of the test and the three way valve (3WV) is switched to position 13 such that
44 nitrogen gas can flow into the dewar f rom gas cylinder pressurizing the dewar. Third, chilldown fluid components upstream of the test section and fill up the precooler. The solenoid valve 2 (SV2) is energized letting the fluid flow into the precooler. The SV2 is closed when the precooler is fu ll. Fourth, pre chill the fitting connections inside the test chamber. This is simply done by flowing LN 2 into the test section meanwhile turn on the film heater. As the LN 2 flowing into the test section, the pipe, tee, and nozzles will be quenched as well as the test plates. After the temperatures of all these components are dropped the below 110 K, the LN 2 flow is stop pe d . After the test plates are heated back up to the desired temperature, th is step is complete. Lastly, set the pulse flow pattern. The pre determined combination of duty cycle and period are entered into the Labview VI. At this point, the test system is ready for the first chilldown test. The spray chilldown test can only begin after the start of parabolic maneuver. In the flight, the ZeroG staff will notify the flight crew before going into reduced gravity period. After notification, the experiment er need s to adjust his position in order to perform the test. After entering the reduced gravity peri od, the flight attendants will signal the start of the reduced gravity period. Besides, the gravity level is measured by a triple axis accelerometer and the reading of the sensor is shown in the VI. The test is started by click the Labview VI button on the laptop screen. Then, the LN 2 will be spray ed into the test section according to the pre set spray pattern controlled by the VI . The reduce gravity period lasts between 17 and 20 seconds. Before the end of reduced gravity period, the flight attendant s will signal the end of reduced gravity period to the flyers. For the ground test, t he chilldown test is complete when the temperature readings of the two discs do not drop anymore. For the flight test, however, the spray is
45 generally stopped after 25 seconds even though the test environment may be in the hypergravity condition . After the completion of the first test, the crew members will prepare the experiment apparatus for the following five parabola sets. This is done by turning on the film heaters that at tached to the back of the test plates to heat the SS plates back to the desired temperature and setting the tank pressure to the desired test pressure. The preceding procedures are repeated for the rest of the flight. After all the six parabola sets are fi nished, the system is required to be shut down and returned to the state before the test started. And this denotes the complet ion of the flight test. Experiment Conditions and Test Management There are many parameters that could be varied in t he base case tank chilldown experiment . Th ese parameters are the orientation of the test plate (for ground test) , nozzle flow capacity , the supply pressure of the LN 2 , the combinati on of the duty cycle and period for a pulse spray. For the ground test, the effect of the directions of gravity is studied by changing the test plate orientation . There are three possible orientations that the test plate can be placed and they are horizontal upward facing (1g), horizontal dow nw ard facing ( 1g), and vertical facing (tangential 1g to heat transfer surface). The volumetric flow rate can be varied eit her by changing the supply pressur e or by changing the nozzle . However, for a given nozzle the flow rate will not increase significantly by increa sing the liquid supply pressure. For example, the nominal flow rates of the small nozzle at 80 psi and 100 psi are 1.0 4 GPM and 1. 05 GPM , respectively. If the big nozzle is used, the flow rate can be achieved at 80 psi is 2.08 GPM . Hence, replacing the noz zle with a higher flow capacity one is a more effective way to increase the mass flux on the sprayed surface . To study the effect of pulse spray
46 the duty cycle can be set at 10 %, 20%, 4 0%, and 80%. And the period can be set at 1 s, 2 s, and 3 s. In total, the combination of different duty cycle s and period s gives a total of 12 different spray patterns. However, to limit the total amount of the tests, not every parameter is varied. In order to organize all the tests performed, each test is name d in Duty period , test number spray , the place for the duty period is left blank . For the two nozzle s configuration case, the letter D is used to denote double pl ates in the test section . Otherwise, it is not noted . For example, for a n upward facing single nozzle ground test with a tank pressure set at 80 psi and the LN 2 is sprayed according to the pulse pattern of 20% duty cycle and 1 secon d period, the 18WL3490D80PSI20%1sa A total of 310 ground tests are performed before the flight. For simplicity , the full test name of these tests will not be shown but the naming scheme will be referred to in the later chapter s .
47 Table 3 1 . Thermo physical properties of saturated nitrogen and oxygen at 1 atm . Thermo physical properties LN 2 GN 2 LOX O 2 Density [kg/m 3 ] 806.2 4.6 1141.3 4.5 Specific heat [kJ/(kgÂ·K)] 2.0 1.12 1.7 0.97 Viscosity [PaÂ·s] 1.6e 4 5.4e 6 1.9e 4 6.9e 6 Thermal conductivity [w/mÂ·K] 0.14 0.007 0.15 0.008 Surface tension [N/m] 0.0089 0.01 3 Saturation temperature [K] 77.3 90.2 Latent heat [kJ/kg] 199.2 213.1
48 Figure 3 1 . Single and double nozzle configurations of the test section. Figure 3 2 . The schematic of f luid system .
49 Figure 3 3 . Exploded view of the test section. Figure 3 4 . Locations of the thermocouples.
50 Figure 3 5 . CAD model of the test system (main rig). Figure 3 6 . Front view of the main test rig and gas cylinder rig. Photo courtesy of author , taken January 3 rd , 2019.
51 Figure 3 7 . Back view of the main test system . Photo courtesy of author , taken January 3 rd , 2019.
52 Figure 3 8 . Electrical diagram of the test system.
53 CHAPTER 4 CRYOGENIC SPRAY COOLING UNDER TERRESTRIAL CONDITION The chilldown experiment by cryogenic spray under terrestrial conditi on are investigated in this chapter. Before the test, visualization of the spray is attempted to determine the actual spray angle of the nozzle s when LN 2 is used as working fluid. To study the effect the gravity towards the heat transfer behavior, the orie ntati on of the test plate is varied by placing it at the top, bottom , and side of the cubic vacuum chamber. The corresponding orientations are 1g downward facing, 1g upward facing, and 1g vertical facing, respectively. Both the continuous and pulse spray s are performed to study the performance of cryogenic chilldown by spray cooling. The enhancement of FEP coating on the chilldown performance is verified through directly comparing the chilldown results of a FEP coated plate with the test results of an uncoa ted plate. Spray Parameters Spray A ngle Even though these two nozzles are labeled to have a 90Â° spray angle by the manufacture r , the actual spray angle is differen t from the specification . This is because the nozzle specification provided by the manufacturer is obta ined from the test result of water, which stays as liquid at room tempera ture. Nevertheless, the thermo physical properties of LN 2 are significantly different from those of water. Due to the substantial pressure drop a cross the orifice o f the nozzle, the LN 2 will instantly boil off and become the mixture of the liquid and vapor while no phase change would occur if water is used as working fluid. As a result, one could expect the actual spray angle of the nozzle when LN 2 is employed as wor king fluid would be different from the case where water is used. The actual spray angle for LN 2 is obtained by visualizing the spray coming out of the
54 nozzle tip. This is simply done by putting two flanged viewports on the cubic test section. One viewport is for the observation, the other one is for the illumination . After the sprays are fully established, t he photos of the spray are captured by a cellphone camera and they are given in Figure 4 1 . The spray angles of the two nozzles at 80 psig and 100 psig tank pressure are calculated according to the captured footages and are given in Table 4 1 . Other than these two nozzles, spray nozzles from other companies are also tested in searching for a large r spray angle nozzle. Unfortuna tely, the actual spray angle of these tested nozzles c annot match the specified spray angle. One thing worthy of attention is that these reported spray angles only applied to the case that the relatively stable spray is reached. In fact , before the relatively stable spray can be established , the spray angle v aries a lot and takes about 20 s to become relatively stable. This variation of the spray angle is mainly due to the pressure fluctuation resulted from the phase change upstream of the nozzle outlet and the changing phase composition of the flow, which is also the result of the phase change in the upstream . Due to the complexity of the two phase flow problem, the actual spray angle cannot be determined if the relatively stable spray is not established . Even though the pipe fittings and nozzles are quenched in the preparation steps and insulated by polyethylene foam, these components would still warm up to around 180 K before a test can start. For this reason, the transient spray will always exist in the first 20 s of a test and the actual spray angles are di fferent from the reported values. More detailed results regarding the phase change occurred upstream of the nozzle will be shown in the next section.
55 Quality Compared with the spray cooling by room temperature fluid and refrigerant, great difficulties a ris e for the LN 2 spray cooling as it is almost impossible to achieve a high degree of subcooling. This is because the saturation temperature of LN 2 is significantly much lower than room temperature. For example, without spending any effort room temperature wa ter would have a degree of su b cooling at least 75 K. Nevertheless, the Liquid Helium must be used to cool LN 2 to achieve the same degree of subcooling even if the nitrogen become solid . As a result of unable to achieve a high degree of subcooling, the flui d coming out of the spray nozzle would always be two phase mixture instead of s ingle phase liquid, which might contribute to the fact that actual spay angle of LN 2 is much less than the specified spray angle. Mainly, the contributions to the phase change c an be summarized as 1) the heat transfer from the fluid components before the test section; 2) the heat transfer from the fluid components inside the test chamber; 3) the high pressure drop across the orifice of the nozzle. Before the LN 2 can enter the tes t chamber, it must go through a series of fluid components such as liquid flow meter, solenoid valve, and pipe connections . Before a test can start, all these components are quenched during the precool method in the literature. Notwithstanding , the heat can still be transferred from ambient air to the fluid, even though all these components are covered by insulation material. This heat addition to the fluid reduces the achieved degree of subcooling from precooler or even boils off the cryogen to some degree . The state of the fluid can be determined at two locations by measuring the fluid temperature and pressure. One measurement station is inside the precooler. The o ther one is at the inlet of the test section. To determine the state of the flow, t he measured fluid temperature is compared with the
56 calculated saturation temperature of LN 2 base on the pressure measurement . If the measured fluid temperature is higher tha n the calculated saturation temperature, the fl ow is a two phase mixture. Depending on the test conditions and test operation, the fluid enter s the test chamber can be subcooled, close to saturation, and two phase mixture. Figure 4 2 gives the degree of subcooling of two different tests as an example . The flow is two phase mixture when the degree of subcooling is less than zero. I f the fluid is neither saturated nor sub cooled, the assumption that the fluid is saturated has to be applied since the phase comp osi tion of the flow cannot be determined . On the other hand, t he heat addition from the fluid components inside the test chamber can be estimated. The temperature s of the nozzle, tee, pipe, and the ambient air are measured by a single TC on each com ponent. The temperature measurements of these fluid components for a typical test are given in Figure 4 3 . The se components delivering the LN 2 are always at a much higher temperature than the fluid temperature when a test can begin . Thus, the chilldown of all these fluid components is inevitable. To estimate the heat transfer to the fl uid, three assumptions have to be applied. The first one is to assume that the temperature of each component can be measured by a single TC. This is equivalent to say that no temperature gradient exists in each of these component s . The second assumption is that the heat transfer between the fluid components and the GN 2 inside the test section can be ignored. Last ly , it has to be assumed that the change of flow rate is relatively small such that a simple energy balance can be used. After applying all these assumptions, the resulted fluid enthalpy and equilibr ium quality can be calculated :
57 , ( 4 1 ) , ( 4 2 ) w here is the mass flow rate of the fluid, m is the mass of the component, c is the specific heat, T is the temperature, t is the time, H f is the enthalpy of saturated liquid evaluated at fluid pressure , H g is the enthalpy of saturated va por evaluated at fluid pressure . Other than the heterogeneous boiling caused by the heat transfer from fluid components, homogenous boiling can occur due to the rapid drop in pressure. The saturated LN 2 at higher pressure can boil off immediately when the fluid pressure is dropped to a lower value prompt ly . This type of boiling happens throughout the fluid and does not require contact with any surface. To account the effect of this homogeneous phase change, the equilibrium quality of the fluid coming out of the spray nozzle can be calculated . Even if the fluid may not reach equilibrium before reaching the test plate, the calculated equilibrium quality would still give a good indication of the state of the flow before touching the test plate . To illustrate this , Figure 4 4 plots the resulted equilibrium quality against the fluid pressure for the case when saturated LN 2 suddenly dropped to 1 atm. In general, the resulted equilibrium quality increases monotonically as the pressure drop increases. Combining these two factors contributing to the boil off of LN 2 insid e the test chamber, the resulting equilibrium quality of the fluid coming out of the nozzle is , ( 4 3 )
58 w here H f ,cs is the saturated liquid enthalpy evaluated at chamber pressure, H g ,cs is the saturated vapor enthalpy evaluated at chamber pressure. One should note that t he equilibrium quality only indicates the potential of the phase change that the fluid can go throug h but not the actual quality. If the evaporation of the droplet due to the convection from the media inside the test chamber can be neglect, the equilibrium quality at the nozzle outlet is the same as that at the surface of the plate. The equilibrium quali t ies of these two stages are plotted in Figure 4 5 for a typical test. The first three peaks in these two equilibrium quality curves correspond to the chilldown of the nozzle, tee, and pipe respectively. After 18 s the contribution due to the fluid components inside the test chamber become negligible as the calculated quality, X mid , stays relatively close to zero, which is the result of the chilldown of all the f luid components in side test section. After this point in time, the mass flow rate of this test became relatively stable as shown in Figure 4 6 . From this aspect , the influence of the transient nature of the LN 2 spray cooling test is also asserted. Heat Transfer of Cryogenic Spray Cooling Plate Temperature Profile The test plate temperatures are measured by 25 thermocouples, which are soldered to the vacuum side of the test plate. A total of eight film heaters are attached to the test plate between the thermocouples. The schematic of the thermocouples and film heate rs attachment is given in Figure 4 7 . Based on the radi al location on the disc, the 25 thermocouples can be divided into 8 thermocouple groups. Each thermocouple in t he same group has the same radial distance to the center of the disc. The advantages of arranging the temperature measurement and film heaters are : 1) the temperature distribution across the test plate can be monitored; 2) the heaters can warm up the test
59 plate from liquid temperature to the desired temperature relatively evenly; 3) the error from the misalignment in installation can be reduced. The reason behind t he first two advantages is that most of the thermocouples and heaters are evenly distributed i n the Since the whole test apparatus is made/assembled in house, one should not expect the locations /alignments of all the parts inside the test section would be exact to the design. The most significant error is that the nozzle is not posi tioned to point to wards the center of the test plate with great accuracy . And this error can be spotted in the chilldown curves where the temperature curves in the same TC group deviate from each other noticeably . Since there are at least three thermocouples placed axisymmetric ally to the center of the plate in the thermocouple group s 2 to 7, the temperature profiles of these thermocouples can be averaged. This averag e temperature curve would give a better depictio n of the heat transfer behavior at the location of interest by reducing the error caused by the misalignment of the nozzle . Figure 4 8 gives the typical chilldown cur ve s from all 25 thermocouples. To illustrate the point that the average temperature should be used, the TC group 2 is taken as an example. As can be seen from the plot, in the second TC group the temperatures measured by TC2 and TC10 are almost identical t o each other . Nonetheless, they are notably different from the rest two temperature measurements in this TC group, TC14 and TC22. This is because the center of the spray zone of the plate. And one can expect that in the cas e these center points collapse with each other, the temperature profiles of the four TCs should be close and sitting somewhere between TC10 and TC14. Figure 4 9 gives the average temperature profile of the eight TC groups. Figure 4 10 plots the individual TC mea surement together with the average
60 temperature of each TC group . As it is shown, the average temperature T2 gives a better description of the temperature history pro file than the four individual temperature measurements at th is given location. In general, the initial temperature of the plate is not uniform when a test started. For the test shown in Figure 4 9 , the initial temperature across the plate varies from 285.9 to 346.3 K . The lower bound of plate temperature, 285.9K, is measured by the these flanges contacting the test plate weigh 37 lbs, their thermal mass is considered bi g enough such that the temperature of the flanges will not be affect ed by the cooling of the spray and is kept at room temperature throughout a test. Hence, at the outermost location of the plate, the initial temperature is low er than that near the center . To exclude the influence of the thermal mass from the flanges, the measured temperature T8 is not included in the calculation of the quantit ative results of plate chilldown but be used as a boundary condition. For example, whether the plate is quenched or not is decided by the temperature measurement of T7 instead of T8. If not including T8, the lower bound of the plate initial temperature is 320.1K. In order to simplify the initial c ondition of the plate, the mass averaged temperature is defined : , ( 4 4 ) w here T is the plate temperature at location r, r is the radial location on the disc, R7 is the radial location of T7. The numerator of (4 4) is calculated by integrating the interpolated initial temperature profile across the plate given in Figure 4 11 numerically by trapezoida l rule.
61 For this test, the mass averaged initial temperature is 338.4 K. In the following sections , if initial temperatu re is called, it means the mass averaged initial temperature of the plate from the center until the location of T7 . The chilldown curves of a ll the 25 TCs show a shape that is similar to the temperature curve of a pipe chilldown test in Figure 2 2 . A fter the test started, the temperature of the plate decreases relatively slow at first. Then, after the plate temperature is lowered to a certain point, the slope of the temperature curve suddenly increase d which corresponds to the rapid increase in chilldown rate . Shortly after this increased cooling rate period, the slope of the temperature curve start s to decrease to zero. Eventually, all the temperatur e curves are leveled out and close to each other. The final temperature of the plate after chilldown is lower than the inlet temperature of LN 2 and close to the saturation temperature evaluated at the chamber pressure. Comparing the chilldown curves at dif ferent locations, the measured temperature closer to the center of the disc always decrease faster than the temperature further away from the center. This indicates the chilldown rate depends on the radial location, and the center of the plate has the fast est chilldown rate. This observation is physically explainable as the distribution of the mass flux by a spray nozzle is always the highest at the center of the spray cone and decrease to zero at the location far away from the center. Heat Flux and Heat T ransfer Coefficient The heat flux on the heat transfer surface is of vital interest to characterize the spray cooling, which cannot be directly measured. With the measured temperature history on the vacuum side of the test plate, the heat flux can be estimated by applying man y approximations and assumptions. In general, this estimation of surface
62 temperature and surface heat flux by the measured temperature on the other side of a slab constitutes the Inverse Heat Conduction Problem (IHCP). Nonetheless, a simple solution to the two dimensional IHCP can be hardly found in the literature, not even to mention the three dimensional solution. If one assumes the temperature distribution across the circular disc is axisymmetric and neglect temperature gradient in the radius direction, the IHCP can be approximated as one dimensional IHCP which simple solutions to heat transfer surface temperature and heat flux can be readily found from Burggraf  . The surface temperature is given by , ( 4 5 ) w here T o is the plate temperature measured at the vacuum side, is the t hermal diffusivity of the plate, is the thickness of the plate, t is the time. The heat flux at the spray surface is given by , ( 4 6 ) w here T o is the plate temperature measured at the vacuum side, is the d ensity of the plate material, c is the specific heat of the plate material, is the therma l diffusivity of the plate, is the thickness of the plate, t is the time. If one relax es the assumption that negligible temperature gradient in radi al direction but assume the temperature changes little across the two surfaces of the test plate, the lump ed body analysis can be used to estimate the heat flux. This estimation can take account into the heat conduction in the radius direction which is significant near the region of quenching front but can be ignored far away from this region.
63 ( 4 7 ) Taking only the first three terms of (4 6) , the equation takes the form of: ( 4 8 ) To take account the temperature gradient in the radi al direction but ignoring the temperature difference across the thickness direction of the plate, (4 7) can be re written as: ( 4 9 ) (4 9) can be divided into two parts. The first term on the right hand side is the first term of (4 8) , and it considers the change of thermal energy with respect to time at the given location of the measurement. The second term calcul ate s the heat flux in the radial the Piecewise C ubic Hermite P olynomials are used to interpolate the temperature profile across the radi al direction with known temperature s at measured location s at each tim e step . Then, the temperature derivatives in the r direction can be calculated numerically by central difference according to the temperature curves. In the calculation of the heat flux, the average temperatures are used instead of the single TC measuremen t. Figure 4 11 gives the temperature profile s across the plate in the radial direction throughout one test. The temperature profiles are plotted every 4 s after the t est started. For the first 8 second s direction even though the locations closer to the center chilled faster than the locations f urther away from it. How ever, one can clearly see a sharp temperature gradient
64 between the locations of T2 and T3 at 12 s while t he temperature profile is relatively flat before and after th is interval . From Figure 4 9 , one can tell that the part of the plate at T2 location is almost quenched while at the location of T3 has no t yet reached Leidenfrost point at 12 s. Thus, it is concluded that this sharp gradient in the temperature profile rep resents the quenching front of the spray. Moreover, as time increases this quenching front moves towards the outer edge of the test disc. Owing to the resolution of the TC s placement , the exact location of the quenching front cannot be determined. The heat flux es calculated by both two methods at seven TC locations are plotted against time in Figure 4 12 . Typically, the calculated heat flux es by (4 8) and (4 9) do not deviate from each other significantly but not iceable difference does exist in the plot. Since the heat flux curves acquired by these two methods follow each other in terms of both the trend and magnitude, the heat flux calculated by (4 9) will be used for further discussion without concerning the justification of method. The boiling curves at seven TC locations calculated by the lumpe d body method are given in Figure 4 13 . For illustration purpose, the boiling curve calculated by single TC 22 is plotted in Figure 4 14 and a typical boiling curve from pipe chilldown test is given in Figure 4 15 . The degree of superheat for spray cooling is defined as the surface temperature subtract ed by the LN 2 saturation temperature evaluated at the chamber pressure. On the contrary , for the pipe chilldown test the saturation temperature is evaluated at the fluid pressure. Essentially , the cryogenic chilldown of a plate by spray shows comparable he at transfer characteristics as the quenching of a pipe by flowing through LN 2 . Just as the boiling curve of the pipe chilldown test, the spray chilldown test can also be divided into four
65 distinctive region s according to the response of heat flux to the su rface super heat . First of all, the heat flux increases while the wall superheat is decreasing right after the test is began, as labeled by the arrow A in both Figure 4 14 and Figure 4 15 . However, this period in the pipe chilldown test is much short than that in the plate spray chilldown test. During this period, it is considered that the quality of the fl ow is so high or even complete ly in vapor phase that the flow exhibits a similar cooling effect as the single vapor phase convection . The increase in the heat flux is ess enti ally the result of the increasing temperature difference between fluid and surface while the heat transfer coefficient changes little. This point can be explained by an example of an infinitely long tube. In the case of LN 2 flowing into an infinitely l ong tube, by the time the nitrogen reached the exit of the tube its temperature would be the same as the initial temperature of the tube. Thus, the heat flux on the tube surface is zero all the time. However, at the location that is far away from the inlet but not infinitely away from the inlet the temperature of the tube will stay unchanged for a long time, but it will even tua l ly start to drop as the upstream of this location is quenched . D uring this period the fluid phase changes from the single vapor phase to the vapor liquid mi xture phase and the bulk fluid temperature drops from tube initial temperature to the vicinity of the saturation temperature of the cryogen. On the other hand, this scenario will never happen at the location that is infinitely c lose to the inlet because the fluid will always be in liquid state . Essentially, the less er the thermal mass upstream of the interest ed location, the faster the flow can transit from the single vapor phase to the mixed phase . Similarly , one ca n conclude th at the first heat transfer mode on the boiling curve of a cryogenic chilldown test represent s the heat transfer transit from the single phase vapor convection to the
66 vapor liquid two phase convection. And the duration of this transition depends on the ther mal mass upstream to the location in question. For the pipe chilldown test, this transition is so short that it is usually ignored . However, due to the added thermal mass of fluid components in the test chamber of the spray cooling test, this transition lasts longer than the line chilldown test. When the fluid bulk temperature is close enough to the saturation temperature, this transition period labeled by letter ends and the heat transfer enters the second regime labeled by letter where the boili ng curve can be approximated by a straight line. Film boiling is considered as t h e heat transfer mode of this regi me . Owing to the special flow structure of the spray, the film boiling in the spray cryogenic chilldown shows some unique features. First of a ll, the heat flux is comparably high compared to the critical heat flux at the center of the spray cone. This is shown by the boiling curve , qn1 in Figure 4 13 , which is measured at the center of the plate. For instance, the second peak of qn1 is even higher than the critical heat flux near the end of the boiling curve. The second feature is that the dependence of heat transfer on the radial location. By comparing the b oiling curve s at all seven locations, one can find that the heat flux of the same superheat during film boiling period is always higher at the TC that is closer to the center of the spray . And this is also shown in the temperature curves Figure 4 9 as the curves always drop faster at the TC location that is closer to the center. Third , the heat transfer behaves differently according to whether the loc ation is inside the spray zone or not. For this test, it is estimated that the spray zone is 2.17 inches in radius. From Figure 4 7 , one can determine that this area s hould cover up to the location of T5. To see this difference , one has to examine the fluctuations of all the boiling curve s in Figure 4 13 . The most notable sign woul d be the
67 three peaks in qn1 occurred at the begin ning of the test. This fluctuation in heat flux is usually the result of the fluctuation of the mass flow, see Figure 4 5 and Figure 4 6 . These three peaks are still detectable in curves qn2 to qn4 even though the magnitude is greatly decreased. However, in the curves qn5 to qn7 these peaks can no longer be distinguished from the regular heat flux fluctuation. the film boiling reions of the heat flux curves qn4 to qn7 are relatively more stable than those of the curves of qn1 to qn3. Based on these observations, it is considered that the heat transfer inside the spray zone mainly is the combin ed effect of the droplets impact and the convection of the fluid filament or droplets from the center of the spray zone. Ho wever, for the part of the plate that is not covered in spray cone , the heat transfer mainly is the sole effect of the fluid convection from the spray zone. After the surface is cooled to the exten t that the vapor film between the fluid and surface can no long er survive, the liquid would be able to rewet the surface intermittently. At this point which is also called Leidenfrost point, the film boiling ends and transition boiling starts. The corresponding temperature is called rewet temperature. In the transition boiling the heat flux increases rapidly as the superheat decreases until a maximum value is reached which is called critical heat flux. The critical heat flux point separates the transition boiling from the nucleate boiling where the heat flux decreases as the superheat decreases. The nucleate boiling ends at the temperature where the superheat can no longer support the nucleation of the fluid. In Figure 4 13 , there are more than one peaks in the region of nucleate boiling and transition boiling in some boiling curves. This is the result of averaging the heat flux curves of several TCs in one TC group . A typical heat flux should only be o ne peak as indicated in Figure 4 14 and Figure 4 15 .
68 The Heat Transfer Coeffi cient (HTC) is calculated by , ( 4 10 ) w here T sat ,cs is the saturation temperature of the LN 2 evaluated at the test chamber pressure. The plot of the HTC against the surface superheat is given in Figure 4 16 . With the HTC known , the Biot number, which is the ratio of heat transfer resistance inside of and at the surface of a body can be calculated as , ( 4 11 ) w here h is the HTC, k is the thermal conductivity of the stainless steel , L c is the characteristic length and the thickness of the plate is used for characteristi c length. The calculated Bi number is also plotted against the superheat and is given in Figure 4 17 . The Bi number at the T3 to T7 during film boiling period is less than 0.05 while at T1 it could be as high as 0.13 during film boiling. In general, the lumped body assumption is good before the transition boiling for the estimation of the heat flux except at the location s of T1 and T2. Optimization of Chilldown Performance Enhancement by FEP Coating To demonstrate the effect of the FEP coating on the enhancement of the cryogenic spray cooling, the test section is changed to the two spray nozzles and two test plates configuration. Two nozzles that identically the same are installed on a tee and oriented to spray at the opposite directions. This allows these two nozzles to spray LN 2 onto two test plates that are in the front and back of the cubic test section
69 simultaneously. Since the fluid path s of these two nozzl es are symmetric other reason to consider the flow coming out of these nozzle s will be significantly different from each other . In other words, the thermodynamic and hydrodynamic properties of the flow coming out of these two nozzles should be the same at any time. Besides, t he flow rate of each nozzle is ideally half of the measure d value by the liquid flowmeter. Moreover, the two test plates are prepared in an identical fashion ex cept that one is coated by Teflon 959G 203 on the surface . This means that the TCs and heaters attached to the back of the plates at the vacuum side are at the same location. The only difference between these two plates is that one of these plate s is coated by FEP paint . In addition, these two plates are also placed in the test section in a way that they are symmetric to each other about the center of the test section . Considering all these symmetrical and identical features of the nozzles and test plates, one can conclude that at any time the sprays onto these two plates are identical. This set up of the test section allows the direct comparison of the chilldown curves of the coated plate to those of the baseline case, the uncoated plate . If one plate can be chilled faster than the other, it can be concluded that th e FEP coating has an effect on the chilldown performance. The chilldown curves of test 14WL11 / 290DD80PSI100% a are given in Figure 4 18 . The SS plate coated by FEP is labeled as plate A, and the uncoated plate is labeled as plate B. The initial temperature s of plate A and plate B are 300.9 and 305.6 is relatively small compared to the temper ature drop of a chilldown test, which is about 200 K. Thus, these two test plates can be considered as having the same initial temperature for the convenience of direct ly comparing the chilldown time and LN 2 consumption. The
70 criterion of the completion of chilldown has to be defined before one can compare these two parameters that de scribe the performance of cryogenic chilldown . The completion of the chilldown of a test plate is defined as the moment that the calculated heat flux at T7 first drops to or bel ow zero. For the given test , the heat flux of AT7 drops below zero after 23.25 s. T he heat flux of B T7 drops below zero after 35 .00 s. Then, it can be said the chilldown time of test plate A that coated by FEP is 23. 2 5 s. And the uncoated test plate B quen ched in 35 s. The total amount of LN 2 sprayed in the test section by two nozzles is defined as the inte gral of the mass flow respect to the time: , ( 4 12 ) w here m LN2 is the total amount of LN 2 flowed into the test section , is the mass flow rate measured by the flow meter, t is the chilldown time, is the time . Then, the total amount of LN 2 sprayed onto one test plate is 0.5m LN 2 , as the flow rates of these two nozzles are considered the same at any time. Given the chilldown time, it is found that plate A use s 0.7345 kg LN 2 while plate B use s 1.2955 kg LN 2 . The chilldown time and LN 2 consumption of comparison tests performed for continuous spray are given in Table 4 2 . In general, the FEP coated test plate chills faster and use less LN 2 than the uncoated test plate. And the percentage of saving in chilldown time and amount of LN 2 are given in Table 4 3 . The FEP coating decreases the chilldown time by a round 4 0 % and the LN 2 consumption by around 45% for all the comparison tests. Besides, the enhancement by FEP coating may also depend on the nozzle as the savings in time and LN 2 consumption by the small n ozzle tests are slightly higher than that by the large nozzle test s . For the saving in time, the small nozzle tests show about 10% more saving than the large nozzle tests. For the saving in LN 2 consumption, the
71 small nozzle tests show about 5% more saving than the large nozzle tests. Since these nozzles are of the same model but have different nominal flow rate s , it is postulate d that the enhancement of FEP coating is more significant for the small flow rate tests. Due to the small difference between the savings from different tank pressure tests, there is no strong evidence shows that the spray pressure also has an influence on the performance of the FEP coating. Intermitte nt Spray Other than introducing the LN 2 continuously to quench the test plate, it can also be chillded by spraying the LN 2 intermittently. This method of non continuously spray LN 2 is also called pulse spray. In the previous study of transfer line chilldown , it is shown that the pulse flow can significantly increase the efficiency of the LN 2 utilization at the cost of prolonged chilldown time . Thus, the intermittent spray method is investigated in the study of the base case tank chilldown to shed lights on the optimal method of quenching a tank for the in space application . A pulse spray can be best described by a rectangular waveform. A rectangular waveform of 20% duty cycle and 1 s period is given in Figu re 4 19 . The period is the length of time that the waveform takes to repeat itself from start to finish. The period that the amplitude of the waveform is greater than zero is called the pulse or mark for a rectangular waveform. In addition, the period when the amplitude of the rectangular waveform is zero is called space. In electrical signal, the pulse and space periods can also be called on time and off time, respectively. A pulse spray is r ealized by sending a rectangular waveform to the solenoid valve before the spray nozzle. During the on time, the solenoid valve is energized such that the LN 2 can pass the solenoid valve and come out from the spray nozzle. When in the off time, the solenoi d valve is de energized. As a
72 result, the flow is cut off and can no longer flow out of the spray nozzle. To define a pulse spray, two parameters are required. One is the period and the other is the Duty cycle (DC). The duty cycle is defined as the percent age of time that the rectangular waveform is in the pulse period ove r the period time of one cycle ( 4 13 ) For the pulse spray, the duty cycle is the percentage of time that the LN 2 is sprayed over one period. With these two parameters, the shape of the rectangular waveform that a pulse spray follows is determined . Though the electrical signal controls the opening and close of the valve follows a rectangular waveform, the flow rate does not resemble the shape of the rectangular waveform. And the measured mass flow rates of two pulse spray tests of the same duty cyc le of 20% but different period are given in Figure 4 20 . For the test that has a period of 1s, the measured mass flow rate mass flow rate only goes down to zero when the duration of the off period is long enough in the case of a longer period. Besides, the test that has a longer period shows significant spikes in the measured mass flow rate than the test with a short er period . These spikes in mass flow are considered as the result of frequent variation of the downstream pressure due to the unsteady and unstable features of a pulse flow. To evaluate the performance of the pulse spray compared to the continuous spray, the chilld own efficiency and chilldown rate have to be defined first. The chilldown efficiency is defined as the ratio of the thermal energy removed from the test plate to the total available latent heat from the LN 2 that is sprayed in the test chamber. To bette r depict the chilldown efficiency, the thermal energy removed from the fluid components
73 inside the test section is subtracted from the total available latent heat of the LN 2 flowed into the test section. T he equation of calculating the efficiency is define d as , ( 4 14 ) w here r is radial location on the dis c, R7 is the r coordinate of T7, SS is the density of the SS, is the thickness of the plate , c SS is the heat capacity of the s ta inless steel, T mavg,F is the mass averaged final temperature of the test plate, T mavg,I is the mass averaged initial temperature of the test plate, m LN2 is the total mass of the LN 2 , h fg,cs is the latent heat of LN 2 evaluated at the t est chamber pressure, T min is the minimum temperature. T I is the initial temperature. The chilldown rate is the change of the mass averaged temperature divided by the chilldown time ( 4 15 ) The pulse spray tests are performed for all three orientations. However, the results of the upward facing test will not be included in the discussion due to the configuration of the test section ( see Figure 4 21 ) . In this case, the outlet of the test section is placed at the side of the cubic vacuum chamber. As a result, any liquid sprayed into the test section will stay on the bottom of the chamber, the test plate, which would give a highly increased efficiency for pulse spray. In other words, this situation is identical to the case that quenching a tank under normal terrestrial gravity condition and there is n o need to spray the LN 2 on the tank wall to increase the efficiency. Thus, only the results of the vertical facing and downward facing pulse spray tests are included in
74 this section. Since there are so many parameters can be varied for the pulse spray test s, these parameters are selectively studied to limit the total amount of the tests that have to be performed. The chilldown efficiencies of all these tests are given in Table 4 4 and Table 4 5 . The chilldown rates of the downward facing tests are given in Table 4 6 . The dependence of the chilldown ra te on the test conditions is straightforward that the higher mass flux over the target surface gives a faster chilldown rate for continuous spray. For the pulse sprays, the chilldown rate decreases as the duty cycle decreases. Besides, at the same duty cyc le, the chilldown rate of the case that has a longer period is much slower than that of the case having a shorter period. This is because the cooling rate during the off period of the spray is greatly reduced due to reduced mass flux or even does not contr ibute to the chilldown at all in the case the flow is completely cut off. Thus, the chilldown time is prolonged as the off period takes a higher portion of the whole chilldown process. On the other hand, the dependence of the chilldown efficiency on the spray pattern is more complicated as the efficiency is affected by m any factors such as the orientation of the test plate, the spray angle, and even the random nature of the spray. For example, the continuous spray of the small nozzle at 80 psi tank pressure shows higher efficiency when the plate is placed vertically than the case when the plate is downward facing. And this is also true for the large nozzle at 100 psi tank pressure. However, little difference is found for the rest two con tinuous tests at two different plate orientations. In addition, the spray itself is somewhat s tochastic that many spray parameters hav e to be determined statistically such as the average droplet diameter and the radial distribution of the mass flux. Thus, only averaging a large number of test results can give a relatively accurate depiction of
75 the tests. For this study, the results of the continuous spray are usually averaged by three to five tests, and the shown results of the pulse spray are usually the a verage values of two to three test s. Even though the sample size of the current study is small that the quantitative result may not be accurate enough, it is believed that the qualitative conclusion will not be distorted due to the small sample size. In th is study, the chilldown efficiency by the pulse spray is only compared with that of the continuous spray. To better see the effect of the spray pattern on the chilldown efficiency, the efficiency of each pulse spray test is normalized by the efficiency of the continuous spray test at the same tank pressure, nozzle type, and plate orientation, they are plotted against the duty cycle in Figure 4 22 to Figure 4 24 . In general, intermittent spray always favors the chilldown efficiency no matter the orientat ion of the gravitational acceleration towards the surface when the small nozzle is used . In essence, pulse spray always promotes chilldown efficiency when the mass flux of the spray is low. For instance, most of these data points show the trend that the ch illdown efficiency increases as the duty cycle decrease when the period is kept at the same. For the test that has a downward facing test plate conducted at 100 psi tank pressure with spray pattern of 10% duty cycle and 1 s period, it shows an increase in chilldown efficiency as high as 54%. On the other hand, the pulse spray shows a dependence on the relative direction of the gravity towards the surface when the mass flux of the spray is relatively high. When the test plate is placed downward facing, the c hilldown efficiencies of the pulse spray are smaller than the efficiency of the continuous spray tests, and this can be seen in Figure 4 23 . For the vertical orientation c ase, the pulse spray by the large nozzle show s a similar trend as the small nozzle that the chilldown efficiency is increased by reducing
76 the duty cycle. Nevertheless, the effect of the period of the pulse pattern cannot be determined as the test data does not give a clear trend. It is preferable to assume the period of the pulse pattern plays little role in the chilldown efficiency. Based on the above observation, it is concluded that the chilldown efficiency of the pulse spray is mostly determined by the duty cycle. For a highly efficient in space tank chilldown , the combination of small duty cycle and small cycle period is preferred. However, the life span of the valve and chilldown rate might be undermined by pursuing a high chilldown efficiency . Intermittent Spray on FEP C oated S urface The FEP coating and intermittent spray method are proven effective in enhancing the chilldown performance. The combined performance of the FEP coating and intermittent spray method is studied, and the experiment data are selectively reported. This is because a detailed experimental parametric study was not allowed due to the limitation of time. In addition, eve n if accurate experiment result is available from this test set up, it may not be accurate for another test set up or application due to the different test section as well as test conditions. Thus, this section only report s the test results of pulse spray performed by the large nozzle at 80 psi tank pressure with a period of 1s. The tests are performed using the double nozzle configuration of the test section. Thus, the results of the coated and uncoated plate are acquired simultaneously. The chilldown efficiencies and normalized efficienc ie s are reported in Table 4 7 . As shown, very high chilldown efficiency can be achieved by combining the intermittent spray method together with the FEP . For e xample, the 10% duty cycle test yields almost 150% increase in the chilldown efficiency compared with the continuous uncoated spray test .
77 In conclusion, combining the intermittent spray method and FEP surface coating gives an elevated high chilldown performance. However, a detailed investigation of this combined effect is encouraged for the following study.
78 Table 4 1 . Nozzle spray angles at 80 and 100 psi tank pressure . Tank pressure (psi) Small nozzle Large nozzle 80 44.5Â° 48.4Â° 100 55.4Â° 52.8Â° Table 4 2 . The chilldown time and LN 2 consumption of all continuous two plates spray tests. Continuous test Plate A Plate B Initial temperature (K) Chilldown time (s) LN 2 consumption (kg) Initial temperature (K) Chilldown time (s) LN 2 consumption (kg) Large60PSI 312.5 27.6 0.815 318.6 44.4 1.4525 Large80PSI 305.7 22.8 0.7605 309.4 34.8 1.3225 Large100PSI 325.8 19.6 0.9506 335.36 32.2 1.7113 Small60PSI 294.8 28.5 0.7435 291.3 53.4 1.4235 Small80PSI 309.9 25.1 0.6555 306.5 49.2 1.326 Small100PSI 308.8 20.1 0.689 304.4 37.1 1.298 Table 4 3 . Effect of FEP coating on chilldown time and LN 2 consumption. Continuous test Percentage saving in time (%) Percentage saving in LN 2 (%) Large60PSI 37.84 43.89 Large80PSI 34.48 42.495 Large100PSI 39.13 44.45 Small60PSI 46.63 47.77 Small80PSI 48.98 50.565 Small100PSI 45.822 46.918 Table 4 4 . The chilldown efficiencies of the downward facing orientation tests. Nozzles Tank pressure (psi) Period (s) 10 % DC(%) 20 % DC(%) 40 % DC(%) 80 % DC(%) 100 % DC(%) Small nozzle 80 1 13.40 13.69 12.55 10.60 9.69 2 14.88 13.52 11.86 10.44 3 12.83 11.38 12.86 10.74 100 1 15.14 13.83 12.24 11.19 10.38 2 16.03 14.00 12.97 11.65 4 14.40 14.06 12.42 11.54 Large nozzle 80 1 11.47 10.17 8.28 8.01 8.92 2 10.27 8.91 7.43 6.49 3 10.97 8.20 7.68 6.58 100 1 10.85 9.41 8.05 7.24 5.75 2 11.26 9.64 8.15 7.14
79 Table 4 5 . The chilldown efficiencies of the vertical facing orientation tests. Nozzles Tank pressure (psi) Period (s) 10% DC(%) 20% DC(%) 40% DC(%) 80% DC(%) 100% DC(%) Small nozzle 80 1 17.22 16.40 15.11 14.47 13.11 2 18.59 17.41 15.57 14.44 3 17.64 16.63 14.77 13.06 100 1 16.82 15.73 14.15 12.26 10.30 2 16.99 15.37 11.69 3 18.12 16.41 12.87 12.70 Large nozzle 80 1 11.40 10.50 9.39 8.89 8.27 2 11.95 11.12 9.74 8.21 3 12.33 10.20 9.45 8.51 100 1 12.28 10.68 9.30 8.31 7.01 2 11.65 10.68 9.44 Table 4 6 . The chilldown rates of the downward facing orientation tests. Nozzles Tank pressure(psi) Period(s) 10% DC( K/s ) 20% DC (K/s) 40% DC (K/s) 80% DC (K/s) 100% DC (K/s) Small nozzle 80 1 4.78 5.25 5.41 5.67 5.92 2 4.75 4.70 5.19 5.93 3 3.81 4.34 5.03 5.46 100 1 5.33 5.98 6.24 6.88 6.75 2 4.97 5.42 5.80 6.73 4 3.90 4.37 5.36 6.64 Large nozzle 80 1 5.69 6.84 7.43 8.11 8.66 2 5.37 5.91 6.77 8.14 3 4.66 5.27 6.62 8.33 100 1 7.22 7.69 8.59 9.95 9.72 2 6.25 7.00 8.17 9.81 Table 4 7 . The efficiency of intermittent spray on coated and uncoated plate performed at 80 psi tank pressure by the large nozzle with a period of 1 s. Duty cycle (%) Uncoated plate Coated plate Efficiency ( %) Normalized efficiency Efficiency ( %) Normalized efficiency 10 10.26 1.26 20.38 2.49 20 9.47 1.16 19.23 2.35 40 8.22 1.01 16.70 2.04 80 7.15 0.88 16.95 2.07 100 8.17 1 15.28 1.87
80 A B C D Figure 4 1 . Photos of the spray angles . A) Small nozzle at 80 psi tank pressure , B) small nozzle at 100 psi tank pressure , C ) large nozzle at 8 0 psi tank pressure , D ) large nozzle at 10 0 psi tank pressure . Photo courtesy of author, taken May 16 th , 2018 .
81 Figure 4 2 . Degree of subcooling of two typical tests.
82 Figure 4 3 . Temperature measurement s inside the test section. Figure 4 4 . Resulted e quilibrium quality when saturated LN 2 dropped to 1 atm .
83 Figure 4 5 . Equilibrium qualities before and after nozzle orifice. Figure 4 6 . Mass flow rate of a typical test.
84 Figure 4 7 . Thermocouples and heaters location. Figure 4 8 . Chilldown curves for all 25 thermocouple s .
85 Figure 4 9 . Chilldown curves by the average temperature. Figure 4 10 . Comparison of the chilldown curves of individual T C and average TC groups.
86 Figure 4 11 . Temperature distribution across the test plate by interpolation. Figure 4 12 . Heat flux against time at seven TC group locations. qn is calculated by lumped body method, and qb is calculated by Burggraf equation .
87 Figure 4 13 . Heat flux against degree of superheating at seven TC group locations. Figure 4 14 . Heat flux against degree of superheating of TC 22.
88 Figure 4 15 . Heat flux against degree of superheating of a typical pipe chilldown test . Figure 4 16 . HTC against degree of superheating at seven TC group locations.
89 Figure 4 17 . Biot Number against superheat at seven TC group locations. Figure 4 18 . Chilldown curves of FEP coated surface (up) and regular SS surface (bottom).
90 Figure 4 19 . A rectangular waveform of 20% duty cycle and 1 s period. A B Figure 4 20 . Mass flowrates of two pulse spray tests at the same duty cycle. A) 1 s period , B) 3 s period.
91 Figure 4 21 . Configuration of test section for upward facing test plate. Figure 4 22 . Normalized efficiency of the downward facing tests by small nozzle.
92 Figure 4 23 . Normalized efficiency of the down ward and vertical facing tests by large nozzle at 80 psi tank pressure. Figure 4 24 . Normalized efficiency of the vertical facing tests by small nozzle at 80 psi tank pressure.
93 CHAPTER 5 CRYOGENIC SPRA Y COOLING UNDER REDUCED GRAVITY CONDITION This chapter reports the results of the cryogenic spray chilldown tests performed in the simulated reduced gravity condition. For the flight test, the test section was set to two nozzles configuration. Two test plates were placed in a position that face each other and the LN 2 were sprayed on by two separates nozzles. As a result, two sets of chilldown data can be obtained simultaneously in one parabola. One was SS plate coated with low th ermal conductivity Teflon 959G 203, and the other was the bare surface SS test plate without any coating to serve as the baseline case for evaluating the coating effects. Table 5 1 g ives the summary of the tests that are performed in the parabolic flight. The experiment is conducted onboard a Boeing 727 200F aircraft. And the reduced gravity environment is achieved through maneuvering the aircraft in th e parabolic trajectory. The flight service is provided by Zero Gravity Corporation at Sanford I nternational A irport , Sanford , Florida. In one flight, a total of six parabola sets are provided, and each parabola set consists of five consecutive parabolas. T he first three parabolas in the first set simulated the Martian gravity level which is about 0.38 g. The last two parabola s of this set simulated the lunar gravity level. The micro gravity condition is simulated in the rest of the parabolas in this flight. Figure 5 1 provides the information on t he parabolic flight gravity level versus time characteristics. The microgravity period is always sandwiched by two 1.8 g periods. The microgravity period nominally lasts between 18 25 seconds. For the research flight, in order to maintain the acceleration levels within Â±0.01g, the microgravity is around 18 second. In total, seven tests were attempted in the flight. However, some of these test results are not usable due to two phase LN 2 mixture is drawn from dewar. When the
94 two phase mixture is flowing throu gh the Coriolis flowmeter , it gives error signal that the flow rate reading is zero meanwhile the chilldown rate is significantly decreased. In previous cryogenic pipe chilldown flight test, on average, at least five tests can be achieved from each flight using the same method to deliver LN 2 from the dewar . However, each spray chilldown test performed in flight uses at least twice of the amount of LN 2 used in the pipe chilldown test since each spray test quenches two test plates. Besides, extra LN 2 is used to keep the Coriolis flowmeter close to the LN 2 temperature. The first two tests are considered perfect si nce the dewar is relatively full and no issues are found from the test result . Figure 5 2 gives the flow test performed in the parabolic flight. As it shows that the mass flow is continuous and nothing unusual occurred . However, the difficulty of drawing pure LN 2 start ed to sho w up from the third test. The measured mass flow rate of the third test of the flight is plotted in Figure 5 3 . Between 2 and 6.5 s, the mass flow suddenly drops to zero. This is not because of that the flow is stopped but the two phase mixture entering the flow meter and error signal is given. Corresp ondingly, the chilldwon rate at the center of the uncoated plate between 3 and 5 s econds is notably decreased and the temperature curve shows the shape of a plateau in this time interval as shown in Figure 5 4 . This unusual change of chilldown rate indicates the vapor takes a very high portion of the fluid that is sprayed in the test section. The measured mass flow rate of the fifth test is given in Figure 5 5 , which indicates the vapor liquid mixture flowed into the test section throughout the whole test. The quality of the tests that do not have proper mass flow measurement are co nsidered marginal since the state of the flow is unclear .
95 Reduced Gravity Chilldown Curves A chilldown curve is defined as the transient temperatur e history recorded by a thermo couple during the chilldown process. So, these curves are the plots of the temperatures measured by the TCs soldered on the back of the disk that document the disk back surface temperature history during chilldown. Figure 5 6 shows a typical set of full 25 chilldown curves for test 1 where the Martian gravity was in effect. The chilldown curves shown were from the bare stainless steel disk. Figure 5 6 is a simplified version of Figure 5 7 where only the medium chilldown curve for each ring is plott ed in addition to the center TC25 a nd outer boundary TC 5. A medium chilldown curve is the one with a chilldown time in the middle among those of the four curves from the same ring. At any instant, the back surface is the warmest and the front surface is the coolest, therefore, chilldown is complete only when the back temperatures do not change anymore . Chilldown time is therefore defined based on the disk back surface temperatures. As seen in Figure 5 6 , th e chilldown curves are divided into four groups identified by red, green, yellow and purple double headed arrows. The color is also used to identify the TCs. For example, TC s 2, 10, 14, 22 and 25 belong to the red group. It is very clear that the closer to the center the faster the ra te of chilldown. The center (TC 25) is the fastest chilldown poin t. A ll the chilldown curves for all seven cases are given in the Appendix B . For each case, there are two sets of curves, one for the bare surface stainless steel disk and the other for the Teflon coated disk. All the sets show the characteristics of spray cooling where the local rate of chilldown is inversely proportional to the distance between the local point and the center of the disk. In general the coating substantially expedited the chilldown process.
96 FEP Coating Enhancement in Reduced G ravity Condition Since the duration of the reduced gravity condition achieved through f lying the aircraft in a parabolic trajectory is relatively short compared with the chilldown time obtained by the ground test, the two test plates cannot be quenched within the reduced gravity period . In general, the Martian g period lasts 20 seconds and t he Micro g period lasts 17 seconds. However, it takes at least 40 second s to chilldown the uncoated test plate, see Table 4 2 . It is decided to stop the test 25 second s after the test started in the flight test regardless whether these two test plates are quenched or not . The average temperature curves of both plates for the first and second test are given in Figure 5 6 and Figure 5 7 . These two tests are unaffected by the issue of drawing liquid vapor mixture from the dewar. The spray is set to be continuous and the tank pressure is set at 80 psig for these two tests. The first one is performed in Martian g condition and the second one is conducted in the micro g condition. The temperature curves labeled by FEP coated plate. The other plot is the me asurement of the uncoated plate. Note AT8 and BT8 are not included in the criteria of determining the completion of chilldown but used as the boundary condition. As shown by the se plots, the uncoated test plate B cannot be quenched in 25 second s , and it is only quenched to the place where the TC group 4 is located at 20 second s . On the other hand, the FEP coated plate is typically quenched in 17 second s before the end of low g period of the parabola. The average temperature profile s of t he seventh test in t he flight are also given in Figure 5 8 . As given by Table 5 1 , this test is performed at 90 psig tank pressure with the pulse p attern of 50% duty cycle and 1s period. Even though this test is considered as failed since the fluid sprayed onto the test plates contains a high portion of nitrogen vapor, it somehow further
97 confirms the enhancement of the FEP coating on cryogenic spray cooling in the respect of the phase composition . C ompared with the first two tests, the fluid sprayed on the two test plates contains higher portion of vapor phase in the seventh test given the reason that the flow meter gave the error signal. Still, the plate coated by FEP is till chilled much faster than the uncoated test plate. Thus, one can draw the conclusion that the FEP coating enhances the cryogenic chilldown performance even though the spray has a high portion of vapor phase. To sum up , the enhancement effect of FEP coating on the cryogenic spray cooling is confirmed regardless of the gravitational acceleration level through the tests that performed in Martian g, Micro g, and 1 g condition. Gravity Effect on Chilldown Efficiency To stu dy whether the orientation of the gravity affects the chilldown efficiency, the chilldown efficiencies of 12 different test conditions are tabulated in Table 5 2 . The test s of the same flow rate conditions (the same tank pressure and same nozzle size) share the same test number. In this way, the orientation effect can be compared by the tests having a similar flow condition. These calculated chilldown efficiencies are also plotted in Figure 5 11 with the error bar. As shown, for the test condition 2 and 3, the efficiencies of different orientations are so close to each other that they are a ll within the error bars of the three tests. For smallest flow rate tests, test 1, difference of the chilldown efficiencies from the vertical test and downward facing test are notable. However, the error bar of these two data points still overlaps with eac h other. Thus, it is possible that at this flow condition the chilldown efficiencies from three orientation conditions are still close to each other. Only at the highest flow rate condition, the downward facing and upward facing test data are outside the r
98 error bar. It is conclude that the relative orientation of the gravity towards the surface does not affect the chilldown efficiency or the influence is marginal when the mass flux on the surface is not very high. Only at the highest f low rate condition invested, the downward facing test has the lowest chilldown efficiency among all three directions and upward facing test has the highest chilldown efficiency. Similarly, the chilldown efficiencies of all the flight tests are given in Table 5 3 . For each gravity condition, there are results for bare stainless steel disk and Teflon coated disk. For test 3 and 4, the estimated efficiency intervals are given. The upper limit of the efficiency is calculated using the measured flow rate. The lower limit of the efficiency is calculated by the corrected mass flow rate, see Figure 5 12 and Figure 5 13 . This is because the flow rate information of these tests are not complete. The correction is made based on experience and educated guess. For tests 5, 6, and 7, chilldown efficiencies were not available for micro g conditions due to inconsistent mass flow rates that were caused by drawing two phase mixture in the supply tank. Comparing the results from test 1 and test 2, one can finding the chil ldown efficiencies of the same test plate are very close to each other regardless the gravity level . Recall the uncertainty in the chilldown efficiency calculation is about 20% of the calculated value, which is relatively large. It is concluded that the ma gnitude of the acceleration force does not contribute to the overall chilldown efficiency or the magnitude of the gravity slightly depress the chilldown efficiency. Comparing test 2, test 3, and test 4, one can find that the lower the duty cycle, the highe r the chilldown efficiency. Even though the uncertainty ranges of the Teflon coated plate of test 3 and test 4 overlaps with each
99 other, the results of uncoated plate of test 3 and test 4 support this conclusion. Thus, the intermittent spray improve s the c hilldown efficiency no matter the gravity conditions. Gravity E ffect on F ilm B oiling Since the duration of the reduced gravity period simulated through the parabolic flight is relatively short, the nucleate and transition boiling data in reduced gravity c ondition cannot be acquired through the test for some TC measurement s . In addition, compared with film boiling, the nucleate and transition boiling play a less important role in cryogenic chilldown process due to the high heat flux and short duration. Thus, only the film boiling heat transfer data is stud ied and compared for the different gravitational acceleration condition. The heat transfer coefficient in the convective film boiling regime is widely cons idered being affected by the mass flux of the heat transfer media. In order to further reduc e the data, the dimensionless number is used in the analysis. The Reynold number (Re) is used to characterize the effect of mass flux, and the HTC is converted into Nusselt number (Nu). To distinguish the effect of the convective heat transfer in the radial direction and the heat transfer due to the droplet impingement inside the spray cone, two Reynold numbers are defined, Re c and Re d . The Re c , Re d , and Nu are , ( 5 1 ) , ( 5 2 ) , ( 5 3 )
100 Where r is the radial location of the measurement point, R s is the radius of the spray zone, is the measured mass flow rate, Âµ is the dynamical viscosity of vapor , k is the conductivity. All the thermal physical properties are the saturated vapor properties evaluated at the pressure of test chamber. In the above definitions , the characteristic length is chosen to be the radial locati on of the measurement points except for the center point . The argument of it comes two folds. First of all, the general used characteristic length for the spray, d 32 , is not available in the current study. Thus, the characteristic length has to be the length scale that can be measured. In this case, the radial location is the best choice among all the lengths that can be used. Secondly, the heat fluxes at all the measurement locations show a dependence on th e radial location. By introducing the radial location into the dimensionless number, this location dependence can be combined with mass flux as one parameter. The length scale for the center point is the half of the characteristic length of the measurement point at T2. The Nu is plotted against Re c for all the TC measurement s except for the center one and it is given in Figure 5 11 . The Re d versus Nu is given in Figure 5 10 for all the measurement points within the spray cone for all continuous test s . The data points shown in these two plots are quite scattered. Nevertheless, the genera l trend of Nu as a function of Re c or Re d can be recognized. In Figure 5 9 , Nu increases as the Rec increases until 25,000. After this value, the Nu flat out or even decreases for the downward facing tests. The physical explanation could be that as the mass flux exc eeds a certain value, the bouncing of the droplets become significant. For the downward facing test, once the droplets bounced away from the touch the surface again. However, the vertical facing
101 tests results do not support this. Hence, it cannot be ruled out the possibility that this decreasing trend of the downward facing tests is actually due to the scattering of the data. The similar trend i s also displayed in Figure 5 10 . When Re d is less than 2 ,000 the Nu increase s with respect to Re d . After 2 ,000 , the Nu no long er increases but level out in regard to Re d . For the ground test data , the gravity shows little influence on the heat transfe r . This is b ecause all the data points from different gravity direction s are mixed together a nd there is no clear boundary to distinguish the test data performed in one gravity orientation from the other in the Figure 5 9 and Figure 5 10 . However, compared with the ground test, the test data from the reduced gravity condition s shows a weak dependence on the magnitude of the gravitational acceleration. I n Figure 5 9 , most of the data points from Martian g and M icro g condition are slightly below the average of the ground test data points at the same Rec. On the other hand, f or the Rec range less than 1.3Ã—104, the Martian g data somewhat overlaps with the Micro g data. The Micro g data points are significantly below all the ground test data a t Rec between 1.4Ã—104 and 1.6Ã—104. Nevertheless , considering the reduced gravity test data only consists of one test for each gravity level, the contributions of data scattering cannot be ruled out easily. Besides, about half of the reduced g data points are still within the lower bound of the ground test region having the same Rec on the Nu versus Rec plot. Therefore, it is concluded that the direction of the gravity relative to the heat transfer s urface does not promote or depress the heat transfer of spray cooling during the film boiling. But the magnitude of gravity may enhance the film boiling heat transfer with a narrow margin . F ilm B oiling HTC Correlation Even though the cryogenic spray chill down also shows a prolonged heat transfer period that is similar to the film boiling period of the cryogenic pipe chilldown , the film
102 boiling of the spray chilldown is different from the film boiling of the pipe chilldown in several aspects. The first one is the va por liquid phase structure. Secondly, the vapor phase takes up the most of the flow in volume when the spray contacting the heat transfer surface ind icating that the heat transfer on the surface is mostly due to the vapor convection . At last, the droplet impingement effect is greatly elevated . It is recognized that a continuous vapor blanket forms between the heat transfer surface and liquid core in the cryogenic pipe chilldown . However, for cryogenic spray cooling , no continuous vapor blanket can exist since the fluid is atomized into droplets by the pressure nozzle. Rather, the vapor separate s the small liquid chunk or the droplet s from the heat transfer surface as the result of the Leidenfrost effect. The liquid chunk may break into smaller drople ts. Or the droplets may coalesce into liquid chunk. As it is shown in the previous chapter that the phase change takes place due to the heat transfer from the fluid components as well as the sudden drop in the saturation temperature, the quality of the spr ay is inevitably high . Subsequently, the majority of the heat transfer surface is covered by the vapor phase suggesting the vapor convection is the main mechanism of the heat transfer. Take the case where the spray has an equilibrium quality of 0.2 kg/kg , as take n from Figure 4 5 , for example. The voi d fraction can be calculated as , ( 5 4 ) w here S is the slip ratio. Taking the physical properties of nitrogen at 1 atm and assuming homogeneous for the slip ratio, the void T his calculation indicates the
103 vapor phase takes up 97.7% of the spray surface in the case of sparse spray . In addition, the Leidenfrost effect also separates the liquid from the hot surface, which further supports the i dea of only considering the heat removal by the vapor phase . The last aspect is very forthright that all these shattered droplets would impinge on the heat transfer surface at least once. Upon the very moment that the liq uid touches the surface, a small portion of the liquid from the droplets flash es to form the small vapor layer separating the liquid from the surface. The heat removal due to the flashing of the droplets should be accounted for . Moreover, the impinging droplets would also disturb the vap or layer and increase the heat transfer between the vapor and the surface. To sum up, the film boiling heat transfer model is constructed based on two arguments: 1) the vapor convection is the primary mode of heat transfer; 2) the droplet impingement adds the extra heat removal due to flashing. The schematic of the film boiling heat transfer model is given in Figure 5 11 . To build the film boiling HTC correlation, several assumptions are used: 1) the mass flux is assumed to be uniformly distributed i nside the spray cone. 2) the spray angle can be depicted by constant value. 3) the vapor temperature is assumed as the saturation temperature at the test chamber pressure . 4) n o direct heat transfer between the liquid and the surface . 5) at the center of the spray heat transfer is mainly by droplet impingement . 6) outside the spray zone the heat transfer is solely by convection. 7) the heat flux at the surface is the sum of the heat flux due to convection and droplet impingement as given by (5 5) below. ( 5 5 )
104 The approximate solution of flow over a flat plate with constant surface temperature is proposed to be the basic form of the HTC cor relation that accounts the heat flux resulted from the vapor convection, and it is given below , ( 5 6 ) w here the Nu x is the local Nusselt number, Re x is the local Reynold number , Pr is the Prandtl number. The coefficient 0.332 of (5 6) is changed to correlate the experimental data . In addition, the ratio of liquid temperature and saturation temperature at the inlet is added to the exponent of Re to improve the accuracy of the correlation slightly. The measured data from the T6 and T7 is used to determine the coefficient . The final form of the Nu for the convection effect is given by (5 7) , and it gives a Mean Absolute Error (MAE) of 12. 77% when fitted to the heat transfer data from T6 and T7 . Nu from the test is plotted against the predicted Nu by (5 7) in Figure 5 12 . ( 5 7 ) The energy balance of the heat flux from the droplet impingement can be written as ( 5 8 ) Where the F is the term that accounts for the percentage in weight droplets would flash upon contacting the hot surface. And its value can be determined by fitting the experimental data. (5 8) can be further reduced into dimensionless form to give the Nu for the droplet term as shown below.
105 ( 5 9 ) (5 9) has to be modified before it can be used for the data fitting to find out F. The ratio, r/R s , is put into an exponential function. This is because if zero is used for r, (5 9) would indicate no heat transfer at the center, which is incorrect. This problem comes from the fact that better characteristic length, average droplet diameter, is not available for current study. Finally, the proposed correlation for the overall HTC is determined by fi tting all the film boiling data including all gravity conditions, and it is given as follows. ( 5 10 ) Where the Âµ f is the viscosity of the saturated liquid , k g is the conductivity of the saturated vapor , h fg is the latent heat, We is the Webber number. The Webber number (We) is calculated as: ( 5 11 ) f saturat ed liquid. (5 10) gives a MAE of 14.89% and Ro ot Mean Square Error of 658.62 for all film boiling data . The model predicted Nu is plotted against measured Nu for all film boiling test data and is shown in Figure 5 13 . This empirical correlation gives a very acc urate MAE of all the test data. However, this a ccuracy is only for a large sample of data point s . For some data points, the error could be very high. In summary , the HTC correlation for the film boiling heat transfer of the spray chilldown is given by (5 10) . The first term accounts the heat transfer of droplet
106 impingement, and the second term account s for the convection of the cold gas. If only (5 7) is used, it still can give a relatively good prediction for the HTC except for at the center of the spray.
107 Table 5 1 . Tests performed in the parabolic flight. Test Simulated gravity level Tank pressure Duty cycle (%) Period (s) Test quality 1 Martian 80 100 Excellent 2 Microgravity 80 100 Excellent 3 Microgravity 80 40 1 Good 4 Microgravity 80 70 1 Good 5 Microgravity 60 100 Marginal 6 Microgravity 90 100 Marginal 7 Microgravity 90 50 1 Marginal Table 5 2 . C hilldown efficiencies of continuous spray tests. Test Number Test Name Downward (%) Vertical (%) Upward (%) 1 18WL3490D80PSI100% 9.69 13.11 11.50 2 18WL3490D100PSI100% 10.38 10.30 11.09 3 14WL11290D80PSI100% 8.92 8.27 8.36 4 14WL11290D100PSI100% 5.75 7.01 8.82 Table 5 3 . Chilldown efficiencies of the flight tests. Test Simulated gravity level Tank pressure (psi) Duty cycle (%) Period(s) Plate A (%) Plate B (%) 1 Martian 80 100 16.91 12.38 2 Microgravity 80 100 19.50 13.43 3 Microgravity 80 40 1 27.07 33.73 20.67 25.73 4 Microgravity 80 70 1 22.04 24.41 12.13 13.91 5 Microgravity 60 100 6 Microgravity 90 100 7 Microgravity 90 50 1
108 Figure 5 1 . G ravity level vs. time characteristics of the parabolic flight. Image Credit: Zero Gravity Corporation , retrieved from  . Figure 5 2 . The mass flow rate measurement of the second test in parabolic flight.
109 Figure 5 3 . The mass flow rate measurement of the third test in parabolic flight. Figure 5 4 . The average chilldown curves of the uncoated plate of the third test in parabolic flight.
110 Figure 5 5 . The mass flow rate measurement of the fifth test in parabolic flight. Figure 5 6 . A typical set of full 25 chilldown curves from a bare surface s tainless steel disk for flight test 1.
111 Figure 5 7 . A simplified set of chilldown curves from a bare surface stainless steel disk for flight test 1. Figure 5 8 . Average temperature curves of the first test in the flight.
112 Figure 5 9 . Average temperature curves of the second test in the flight. Figure 5 10 . Avera ge temperature curves of the seventh test in the flight.
113 Figure 5 11 . Chilldown efficiencies of continuous spray tests. A B Figure 5 12 . Measured and corrected mass flow rate of the flight test 3. A) Measured flow rate, B) Corrected flow rate.
114 A B Figure 5 13 . Measured and corrected mass flow rate of the flight test 4 . A) Measured flow rate, B) Corrected flow rate.
115 Figure 5 14 . Nu v s . Re c . Figure 5 15 . Nu vs. Re d .
116 Figure 5 16 . Schematic of the film boiling model. Figure 5 17 . Model prediction compared with test data during film boiling for T6 and T7.
117 Figure 5 18 . Model prediction compared with test data for all film boiling experimental data.
118 CHAPTER 6 CONCLUSION S AND RECOMMENDATIONS A new experimental system is designed, built, and tested t o study the spray cooling i n the application of terrestrial and in space cryogenic tank chilldown . LN 2 is employed as the working fluid . Prior to the test runs, visualization of the spray is done to determine the actual spray angles of the nozzles when LN 2 is the working fluid. A substantial amount of tests are conducted to cha racterize the performance of the spray cooling in the application of the cryogenic chilldown process. The m ass flux of the spray is varied either by changing the supply tank pressure or replacing the nozzle with one having larger flow capacity. The relative direction of the gravity to heat transfer surface is varied by putting the test plates at the top, bottom, and side of a cubic vacuum chamber for the ground test. The spray cooling data in the reduced gravity condition is achieved b y performing th e test onboard an aircraft flying parabolic trajectory , which simulated t he gravity level on Mars and in space. The heat transfer of the film boiling mode is modeled for all tests conditions. In addition, the methods to optimize the chilldown performance are investigated and they are FEP coating on the surface, intermittent spray method, and the combination of the coating and pulse spray. Conclusions The test results show that the spray cooling cryogenic chilldown is similar to the convective cryogenic pi pe chilldown . The whole chilldown process can be divided into four distinctive regions based on the response of heat flux with aspect to the wall superheat. They are vapor phase convection , film boiling, transition boiling , and nucleate boiling. The chilld own rate of the plate is not uniform but depends on the distance to the center of the spray. The closer to the center of the spray, the higher the cooling effect. The final temperature of the chilldown is close to the saturation temperature of LN 2 evaluated at the pressure of the spray container instead of the liquid temperature at the inlet. The FEP coat ing on the heat transfer surface is proven to be effective in promoting the chilldown performance both in terrestrial condition and reduced
119 gravi ty condition. The saving s in chilldown time and LN 2 consumption are around 40% and 45% respectively for all the continuously ground test performed. The intermittent spray method can give a higher chilldown efficiency, however, at the cost of extended chill down time. The combination of the coating and intermittent spray is the most superior method for the terrestrial and in space cryogenic chilldown and worthy of detailed investigation. Considering the transition boiling and nucleate boiling are relatively less critical to the whole chilldown process in contrast to the film boiling, only the gravity effect on the film boiling is analyzed by virtue of dimensionless number, Re c , Re d , and Nu. When Re c is less than 25,000, the Nu increases as Re c increases. High er than this value, Nu no longer increases with respect to Re c . In addition, the relative orientation of gravity does not play a role in the film boiling heat transfer. T here is no strong evidence to show that the magnitude of gravity plays a crucial role in film boiling heat transfer as data points from tests of different gravity conditions do not separate from each other significantly. The film boiling heat transfer of the spray chilldown is modeled and a correlation to predict the HTC is given. This correlation gives a MAE of 14.89% when fitting all the film boiling HTC from the ground and flight tests. Recommendations for Future Investigation Cryogenic chilldown by spray cooling is excee dingly complicated to investigate experimentally. Prior to the completion of current work, no similar study can be found in the literature. The difficulties arise due to the following facts of current work: The working fluid has to be LN 2 . The transient f eature of a cryogenic chilldown process. The test can be performed in the reduced gravity condition. Nevertheless, as a result of limited time and resources, the current experimental study cannot cover every aspect of this process . To further investigate in this subject, the suggested work s and improvements are listed below: Characterization of the nozzle. First of all, a nozzle with larger spray angle is preferred for the application of in space tank chilldown process. Nevertheless , when LN 2 is employed a s working fluid, the actual spray angle is usually much smaller than the specification of the manufacturer. Thus, the effort is required in the selection of the nozzles. Secondly, many factors contribute to the actual spray angle such as pressure, mass flo w, and vapor phase composition. As a consequence, delicate and adequate tests are required to have a better
120 description of the spray angle. In addition, the mass flux distribution of the spray is required to have a more accurate description of the hydrodyn amic condition at the heat transfer surface. Larger test plate . The size selection of the current experiment set up is a trade off based on many considerations. To have the capacity of carry ing the whole experiment system onboard an aircraft, the system ha s to be compact and lightweight . This limited the size of the test section. However, the designed flow rate cannot be t o o low due to the unsatisfying quality of the insulation of the fluid components. If the flow rate is too low, the phase change resulted from the parasitic heat of the environment before the fluid reaching the nozzle will become significant. Nonetheless, a sparse spray is desired for the tank chilldown application which can be achieved by either reducing the mass flow or covering a large area. Hence, the larger test sample is recommended for the future studies to cover the heat transfer region where the mass flux is extr emely low. Pure liquid spray. As discussed in this work, it is inevitable that two phase mixture is spray ed by the nozzle during most of a te st. Efforts should be made towards achieving the condition that pure liquid is sprayed by the nozzle . Thus, the re sulted data will be cleaner and excluded from other factors disturbing the chilldown process. The suggested measure is to jacket all the fluid components and keep them at liquid temperature throughout testing , which poses great challenge on the test system design . Larger test sample size. To characterize spray chilldown performance, a large test sample size is required to have a statistical description of the performance due to the stochastic nature of the cryogenic spray chilldown .
121 APPENDIX A NOMENCLATURE B i Biot number c Heat capacity DC Duty cycle d 32 Sauter mean diameter d i Droplet diameter H Enthalpy H mid Enthalpy of the fluid before the nozzle orifice H f Enthalpy of the saturated liquid H g Enthalpy of the saturated vapor h Heat transfer coefficient H fg Latent heat k conductivity L c Characteristic length Mass flow rate of LN 2 m Mass Nu Nusselt number n i Number of the droplets Pr Prandtl number P Pressure Heat flux Re c Reynold number for convection heat transfer Re d Reynold number for droplet impingement heat transfer R s Radius of the spray zone
122 r Radial location S Slip ratio T Temperature t Time x Quality Void fraction Thermal diffusivity Chilldown time Thickness of the plate Chilldown efficiency Âµ viscosity Chilldown rate Time Subscripts c Convection conv Convection term cs Saturation state evaluated at the test chamber pressure d Droplet impingement drop Droplet term eq Equilibrium F Final f Saturated liquid g Saturated vapor mavg Mass averaged
123 min Minimum value Nozz Nozzle I Initial i Heat transfer surface in At the inlet of the test section o Surface insulated by vacuum Pip Pipe SS Stainless steel sat Saturation sate Tee Tee
124 APPENDIX B REDUCED GRAVITY TEST DATA Figure B 1 . Simplified chilldown curves for Martian G, 80 psig tank pressure, continuous flow ( Case 1), Teflon coating disk. Figure B 2 . Simplified chilldown curves for Martian G, 80 psig tank pressure, continuous flow (Case 1), bare SS surface disk.
125 Figure B 3 . Mass flow history for Martian G, 80 psig tank pre ssure, continuous flow (Case 1). Figure B 4 . Pressure history for Martian G, 80 psig tank pre ssure, continuous flow (Case 1).
126 Figure B 5 . Simplified chilldown c urves for micro G, 80 psig tank pressure, continuous flow (Case 2), Teflon coating disk. Figure B 6 . Simplified chilldown c urves for micro G, 80 psig tank pressure, continuous flow (Case 2), bare SS surface disk.
127 Figure B 7 . Mass flow history for micro G, 80 psig tank pressure, con tinuous flow (Case 2). Figure B 8 . Pressure history for micro G, 80 psig tank pressure, con tinuous flow (Case 2).
128 Figure B 9 . Simplified chilldown curves for micro G, 80 psig tank pressure, 40% DC, 1s period , Teflon coating disk. Figure B 10 . Simplified chilldown curves for micro G, 80 psig tank pressure, 40% DC, 1s period (Case 3) , bare SS surface disk.
129 Figure B 11 . Mass flow history for mic ro G, 80 psig tank pressure, 40% DC, 1s period (Case 3) . Figure B 12 . Pressure history for micro G, 80 psig tank pressure , 40% DC, 1s period (Case 3) .
130 Figure B 13 . Simplified chilldown curves for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4), Teflon coating disk. Figure B 14 . Simplified chilldown curves for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4), bare SS surface disk.
131 Fig ure B 15 . Mass flow history for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4). Figure B 16 . Pressure history for micro G, 80 psig tank pressure, 70% DC, 1s period (Case 4).
132 Figure B 17 . Simplified chilldown curves for micro G, 6 0 psig tank pressure, continuous flow (Case 5 ), Teflon coating disk. Figure B 18 . Simplified chilldown curves for micro G, 6 0 psig tank pressure, conti nuous flow (Case 5 ), bare SS surface disk.
133 Figure B 19 . Mass flow history for micro G, 6 0 psig tank pressure, continuous flow (Case 5 ). Figure B 20 . Pressure history for micro G, 6 0 psig tank pressure, continuous flow (Case 5 ).
134 Figure B 21 . Simplified chilldown curves for micro G, 90 psig inlet pressure, continuous flow (Case 6) , Teflon coating disk. Figure B 22 . Simplified chilldown curves for micro G, 90 psig inlet pressure, continuous flow (Case 6) , bare SS surface disk.
135 Figure B 23 . Mass flow history for micro G, 90 psig inlet pressure, continuous flow (Case 6) . Figure B 24 . Pressure history for micro G, 90 psig inlet pressure, continuous flow (Case 6) .
136 LIST OF REFERENCES  S. K. Borowski, D. R. McCurdy and T. W. Packard, "Nuclear thermal propulsion (NTP): A proven growth technology f or human NEO/ mars exploration missions," in 2012 , . DOI: 10.1109/AERO.2012.6187301.  B. G. Drake, S. J. Hoffman and D. W. Beaty, "Human exploration of mars , design reference architecture 5.0," in 2010 , . DOI: 10.1109/AERO.2010.5446736.  C. R. Mercer, S. R. Oleson and B. Drake, "A combined solar electric and storable chemical propulsion vehicle for piloted mars missions," in AIAA SPACE 2013 Conference, San Diego, CA ;, 2013 , .  N. J. Bahr, "Developing a safe on orbit cryogenic depot," in The Second Conference on Lunar Bases and Space Activities of the 21st Century, Houston, TX, 1992, pp. 95 100.  M. J. Kaszubowski and J. K. Ayers, "The transportation depot: An orbiting vehicle support facility," in The Second Conference on Lunar Bases a nd Space Activities of the 21st Century, Houston, TX, 1992, pp. 83 93.  J. C. Fikes, J. T. Howell and M. Henley, "In space cryogenic propellant depot (ISCPD) architecture definitions and systems studies," in 57th International Astronautical Congress, Va lencia, Spain, 2006 , .  C. McLean et al , "Simple, robust cryogenic propellant depot for near term applications," in 2011 , . DOI: 10.1109/AERO.2011.5747488.  S. G. Hanna et al , "Cryogenic propulsion stage (CPS) configuration in support of NASA's multi ple design reference missions (DRMs)," in Space Propulsion Conference, Bourdeax ; France, 2012 , .  R. Shaeffer, H. Hu and J. N. Chung, "An experimental study on liquid nitrogen pipe chilldown and heat transfer with pulse flows," International Journal of Heat and Mass Transfer, vol. 67, pp. 955 966, 2013 . . DOI: https://doi.org/10.1016/j.ijheatmasstransfer.2013.08.037 .  G. Liang and I. Mudawar, "Review of spray cooling Part 1: Single phase and nucleate boiling regimes, and critical heat flux," International Journal of Heat and Mass Transfer, vol. 115, pp. 1174 1205, 2017 . . DOI: https://doi.org/10.1016/j.ijheatmasstransfer.2017.06.029 .  J. Kim, "Spray cooling heat transfer: The state of the art," International Journal of Heat and Fluid Flow, vol. 28, (4), pp. 753 767, 2007 . . DOI: https://doi.org/10.1016/j.ijheatfluidflow.2006.09.003 .
137  G. Liang and I. Mudawar, "Review of spray cooling Part 2: High temperature boiling regimes and quenching applications," International Jour nal of Heat and Mass Transfer, vol. 115, pp. 1206 1222, 2017 . . DOI: https://doi.org/10.1016/j.ijheatmasstransfer.2017.06.022 .  M. Kato et al , "Spray cooling character istics under reduced gravity," J. Thermophys. Heat Transfer, vol. 9, (2), pp. 378 381, 1995. Available: https://doi.org/10.2514/3.675 . DOI: 10.2514/3.675.  K. Yoshida et al , "Spray cooling under reduced gravity condition," Journal of Heat Transfer, vol. 123, (2), pp. 309 318, 2001.  T. E. Michalak et al , "Acceleration effects on the cooling performance of a partially confined FC 72 spray," J. Thermophys. Heat Transfer, vol. 24, (3), pp. 463 479, 2010.  T. Gambaryan Roisman et al , "Gravity effect on spray impact and spray cooling," Microgravity Sci Technol, vol. 19, (3 4), pp. 151 154, 2007.  S. Somasundaram and A. Tay, "Comparative study of intermittent spray cooling in sing le and two phase regimes," International Journal of Thermal Sciences, vol. 74, pp. 174 182, 2013.  G. Aguilar et al , "Theoretical and experimental analysis of droplet diameter, temperature, and evaporation rate evolution in cryogenic sprays," Int. J. Heat Mass Transfer, vol. 44, (17), pp. 3201 3211, 2001.  D. Takeda, K. Fukiba and H. Kobayashi, "Improvement in pipe chilldown process using low thermal conductive layer," International Journal of Heat and Mass Transfer, vol. 111, pp. 115 122, 2017 . . DOI: https://doi.org/10.1016/j.ijheatmasstransfer.2017.03.114 .  R. P. Selvam, L. Lin and R. Ponnappan, "Direct simulation of spray cooling: effect of vapor bubble gro wth and liquid droplet impact on heat transfer," Int. J. Heat Mass Transfer, vol. 49, (23 24), pp. 4265 4278, 2006.  J. J. Brogan and R. M. Vernon, "Study of cryogenic container thermodynamics during propellant transfer," United States, Tech. Rep. MFS 14310, 1968.  J. E. Anderson, P. M. Czysz and D. A. Fester, "No vent fill testing of liquid hydrogen," in Advances in Cryogenic Engineering , R. W. Fast, Ed. 1991, Available: http s://doi.org/10.1007/978 1 4615 3368 9_58 . DOI: 10.1007/978 1 4615 3368 9_58.  D. Chato and R. Sanabria, "Review and test of chilldown methods for space based cryogenic tanks," in 27th Joint Propulsion Conference Anonymous 1991, Available: https://doi.org/10.2514/6.1991 1843 . DOI: doi :10.2514/6.1991 1843.
138  R. H. Flachbart et al , "Large scale liquid hydrogen tank rapid chill and fill testing for the advanced shuttle upper stage concept," U nited States, Tech. Rep. NASA/TP 2013 217482, M 1358, 2013.  R. H. Flachbart, A. Hedayat and K. A. Holt, "Modeling and test data analysis of a tank rapid chill and fill system for the Advanced Shuttle Upper Stage (ASUS) concept," AIP Conference Proceed ings, vol. 613, (1), pp. 1284 1291, 2002. Available: http://aip.scitation.org/doi/abs/10.1063/1.1472156 . DOI: 10.1063/1.1472156.  S. C. Honkonen, J. R. Pietrzyk and J. R. Sc huster, "Analysis of pulsed injection for microgravity receiver tank chilldown ," in Advances in Cryogenic Engineering , R. W. Fast, Ed. 1991, Available: https://doi.org/10.1007/978 1 4 615 3368 9_60 . DOI: 10.1007/978 1 4615 3368 9_60.  O. Burggraf, "An exact solution of the inverse problem in heat conduction theory and applications," Journal of Heat Transfer, vol. 86, (3), pp. 373 380, 1964.  Zero Gravity Corporation (2019). Parabolic Flight. [image] Available at: http://www.spaceadventures.com/experiences/zero gravity flight/# [Accessed 3 Apr. 2019].
139 BIOGRAPHICAL SKETCH Jun Dong obtained a Bachelor of Science degree in thermal energy and dynamic engineering from the Central South University , China in June 2012. In May 2014, he graduated with a Master of Science in mechanical engineering from the University of Florida, with a focus on thermal and fluid sciences. Finally, in May 2019, Jun finished h is doctorate in mechanical engineering from the University of Florida.