Citation
Multivariable Expansion of Portable Control Lab

Material Information

Title:
Multivariable Expansion of Portable Control Lab
Creator:
Conner, Jason Christopher
Publication Date:
Language:
English

Notes

Abstract:
The purpose of the project was to expand on the Portable Control Lab and include multivariable controls with coupling and decoupling. The goal was to introduce a three tank system with a hot, cold, and control tank. Fluid flow was controlled via aquarium pumps and two heat exchangers heated and cooled corresponding levels of the control tank. By heating and cooling the hot and cold tanks, temperature and fluid flow both become input variables for a corresponding set point. Overall control of the temperature of the interface was the single output. Due to issues with the pumps and Arduino coding, controllers were not designed. The interface was observed and found to be controllable. The project is under further development. ( en )
General Note:
Awarded Bachelor of Science in Chemical Engineering, magna cum laude, on May 8, 2018. Major: Chemical Engineering
General Note:
College or School: College of Engineering
General Note:
Advisor: Spyros Svoronos. Advisor Department or School: Chemical Engineering

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Jason Christopher Conner. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

UFDC Membership

Aggregations:
UF Undergraduate Honors Theses

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

Page 1 of 23 Honors Project Expansion of Undergraduate Portable Controls Lab to Incorporate Coupling of Multivariable Controls Jason Conner UF Chemical Engineering April 26, 2018 Project Advisor: Dr. Svoronos

PAGE 2

Page 2 of 23 Table of Contents Introduction ................................ ................................ ................................ ................................ ................. 3 Theoretical Background ................................ ................................ ................................ ............................. 3 Evolution of Design ................................ ................................ ................................ ................................ ..... 7 Materials, Equipment, and Constru ction ................................ ................................ ................................ 9 Materials with Specifications ................................ ................................ ................................ ................. 9 General Materials and Tools ................................ ................................ ................................ ................ 12 Construction ................................ ................................ ................................ ................................ .......... 13 Results and Discussion ................................ ................................ ................................ .............................. 18 Conclusion ................................ ................................ ................................ ................................ ................. 23 Acknowledgements and Sources ................................ ................................ ................................ .............. 23 Figure 1. A simple diagra m of a undefined process showing the inputs, outputs, disturbances, and the control logic loop. ................................ ................................ ................................ ................................ ......... 4 Figure 2. A more detailed diagr am for a SISO control system with a feedback loop. ................................ .. 4 Figure 3. Tangent Line and Three Point Method Diagram. ................................ ................................ ........... 6 Figure 4. Frame cuts and shelf instillation. ................................ ................................ ................................ 13 Figure 5. Design of fill and drain system. ................................ ................................ ................................ .... 14 Figure 6. Mounting of heat exchangers. ................................ ................................ ................................ ..... 15 Figure 7. Placement of tanks and cutting of installation "trench". ................................ ............................ 16 Figure 8. Installation of the fans. ................................ ................................ ................................ ................ 16 Figure 9. Improper mounting of PC power supply. ................................ ................................ ..................... 17 Figur e 10. Full 21 hour test results. ................................ ................................ ................................ ............ 19 Figure 11. Test run showing a step change from 10% to 20% heater on. ................................ .................. 20 Figure 12. Comparison of steady state temperatures over test run. ................................ ......................... 21 Figure 13. Testing to find time to reach a true steady state. ................................ ................................ ..... 22 Table 1. List of equipment that has specifications pertinent to the system design. ................................ .... 9 Table 2. List of general materials and tools required for construction. ................................ ..................... 12 Table 3. Steady state tem peratures at multiple % heater on's. ................................ ................................ 20

PAGE 3

Page 3 of 23 Introduction The initial purpose of this project was to expand on the original Portable Control Lab designed by Dr. Svoronos. The original experiment consisted of a single tank with a heater located at the top of the vessel. Two fans were mounted flush with the bottom of the tank. An Arduino cont rolled the heater and fans by using Pulse Width Modulation (PWM) to tune the rate of the heater and fans. The percentage of time the heater was on and the speed of the fans were used to control the temperature within the vessel. This model was a Single Inp ut Single Output system as the manipulated input was the percentage of time the heater was on and the measured output was the temperature of the water in the container. While the fans were able to be manipulated, the only manipulated variable that was unde r automatic control was the heater. The project was designed to include multiple variables with multiple temperature readings. This would allow for multiple inputs and multivariable controls. By having a multiple tank system, when the system was designated as the control system, there we re four possible inputs, with stream input at the bottom of the tank, a temperature gradient would be formed and an interface would exist between the inputs. B y measuring the temperature at the top of the control tank, or a be achieved. Each measured output would have two directly related manipulated inputs. Thus, a m ultivariable control system could be designed. To incorporate coupling, the idea of controlling the temperature of the interface was created. Now, by having to set points, the temperature of the he control tank would fight for control over the temperature of the interface. A multivariable, couple controls problem with multiple manipulated inputs and a singular measured output. Theoretical Background P rocess control theory takes concepts from ch emical, industrial, and control engineering to design systems to provide stable control of processes automatically, consistently, and optimally. Any and all processes have inputs and outputs. Inputs can be defined as manipulated inputs or disturbances. Man ipulated inputs are controlled, or measured, inputs to the process. Disturbances are any input that changes the system that is not measured. An example of a manipulated input is a feed stream flow rate and an example of a disturbance is a decrease in react or volume due to a sudden leak. Manipulated inputs have a desired value, or a set point, that can be changed. Outputs are all streams/signals exiting the process. There is a rule of thumb that for every measured output, a measured input is required. Each process also has state variables. These variables can be a concentration of chemical in a reactor, a volume of fluid, or even a temperature. In general, a state

PAGE 4

Page 4 of 23 variable is matched with both a manipulated input and a measured output, yet multiple inputs ca n act on one state variable. Figure 1 A simple diagram of a undefined process showing the inputs, outputs, disturbances, and the control logic loop. Many controller designs use feedback control In feedback control, the output is measured versus the set point of the controller. Based on the error of the desired output, the manipulated input is either increased or decreased to change the output in the desired direction. Figure 2 A more detailed diagram for a SISO control system with a feedback loop. Undergraduate courses focus on single input, single output (SISO) controller design. This simplifies the controller design by simply having only one input and one output. The objective of controller design is find a transfer function that relates the measured output to both the input and disturbances. The desired transfer function is defined below:

PAGE 5

Page 5 of 23 G u (s) represents the transfer function between the output and the measured input and G di (s) represents the transfer function between the disturbances and the measured output. Multiple disturbances may still exist in SISO systems. Please note that the overall transfer function does not have a dependence on the state variable. The derivation of the transfer function is completed by setting up a state space model as the time derivative of the state variables as a function of the input, the manipulated variables and the disturbances. A measurement equation is found to be function of the input variables that is time delay dependent. These values are found at nominal state (steady state) and then deviation variables are presented as the present value minus the nom inal value. The model is then linearized and a La Place transform is used to find the overall transfer function. Experimental data can also be used to find transfer functions. To find a First Order Plus Time Delay (FOPTD) model from experimental data, G z is found where z is either u or d i : Two common methods are used to find FOPTD models for experimental data: the Tangent Line method and the Three Point method. The tangent line method is presented in the image below. The experimental is graphed as the measured variable versus time. A tangent line is drawn at the point of inflection. The current and new steady states have an imaginary line continued from their values. The point of the intersection between the tangent li ne and the line continued from the original steady state is called t corner a total change of 0.632y is found and a horizontal line is drawn at this value. A time value called t 63 is found. This value is associated with the calculation of tau. From this method, K, tau, and the delay are found.

PAGE 6

Page 6 of 23 Figure 3 Tangent Line and Three Point Method Diagram. The three ues are arbitrarily chosen as P1 and P2. The same strategy is applied, yet the results are found to yield the following: FOPTD Delay models for experimental data can also be calculated by introducing pulses and observing the resulting output. This is called FOPTD with Pulse Change. This method calculates a tau, delay, and K by modeling the function as an peace wise exponenti al function. Once the model is created, the sum of errors square is minimized by iterating values of the three variables and creates the closest model to the experimental data possible. Further modeling can be done to calculate automatic Proportional Integ ral (PI) and Proportional Integral Derivative (PID) controllers. Common controller design methods done in this way are the Ziegler Nichols, Cohen Coon, and ITAE methods. Each of these methods have coinciding constants that calculate new values for K c Tau i Tau D and the delay. Each of these design parameters can be found online or in the chemical engineering literature.

PAGE 7

Page 7 of 23 Evolution of Design Initial conceptual design was driven by the desire to incorporate prior Chemical Engineering course work and material relevant to the heat transfer problem while still presentin g a multivariable control system A three tank system was found to provide a viable solution. The overall task of the control system would be to control the temperature of a hot/cold wate r interface by controlling the inside the tank. These heat exchangers would be located approximately one third and two thirds from the bottom of the tank in a ho rizontal orientation parallel to the bottom of the tank. The hot tank would be heated with a suitable heating element and water would be pumped into the topmost heat excha nger. The cold tank would be cooled with the use of multiple fans and pumped through the bottom exchanger Both the heating element and the fans would have the option of pulse width modulation via the control of a switch. In the case of the heater, a solid state relay could be activated via a small power signal sent from an Arduino Uno co ntroller board. The fans would be powered in a similar fashion, yet with a power consumption of 12V versus the required 3V for the heating element relay, a different power source would be required. Initial designed held that a sufficient power bank of batt eries, such as a small motorcycle battery, could provide enough power to run the experiment for sufficient time. The heat exchangers would be operated in a counter flow orientation to provide the most efficient heat transfer between the elements. Conceptua lly, the heat tank. In between the two heat exchangers would exist an interface where the hot and cold water would be interacting with each other. The transfer of heat from the hot water would cause the cold water to want to rise, yet the hot water would give up its heat and begin to sink. Initial design held that this in terface could be observed while the hot and cold tanks reached steady state. Control of the temperature of this interface could then be achieved by coupling the control systems for both the hot and cold tank. This would depend on the consistency of heat tr ansfer of the heat exchangers. Aluminum automatic transmission oil coolers where thus chosen for their efficient transfer of heat and small size. and the Ard uino Elegoo Uno R3 board. Use of this board and a general purpose bread board was found to be successful during the controls lab section. The Arduino board comes with PCS USB cable to supply power and communicate between the Arduino and the computer. Data collection from a DS18B20 thermometer are carried out via software coding programed by Dr. Svoronos. The software allowed communication of the probe to Microsoft Excel through PXL DAQ (courtesy and thanks to: Jonathon Arndt) through a CP2102 USB to TTL cab le. Initial design above the top heat exchanger in the control tank, 3) between the heat exchangers, 4) below the five probes were planned to be reading at successive intervals to provide real time data for temperatures of the two controlled tanks and the temperatures of the interfaces three stages. The Arduino is capable of a 3.5V and 5V output with a 50mA and 500 m A, respectively, current draw. By supplying the fans w ith power from a battery bank, the

PAGE 8

Page 8 of 23 total power draw from the 5 temperature probes and switches would be possible to be supplied from the Arduino. A secondary goal when designing the set up was to allow for modular integration. The controls lab designed by Dr. Svoronos allowed stude nts to get hands on and see the fruits of their labor Individual hands on lab are currently only present in the controls lab and is a missing part of the current curriculum. A design that allowed integration of a fluid flow and heat transfer system would thus allow students to begin designing and working with their hands on a project during the initial core classes. This project would ultimately lead up to designing control sy stems. The inclusive project would be a system built and operated by the student from the beginning to the end of the program and would greatly increase student interest and participation while positively impacting learning goals for each class. Each tank thus was designed to have an individual fill and drain valve that could be independently operated. A singular fill and drain for the entire system would allow control of fluid flow to pass through the adding yet another control variable. For example, flow rate of heat removal from the control tank. Increasing the number of control variables would both increase the complexity of design and increase its versatility. Again, thi s was a secondary design goal, and would not be implemented at the sacrifice of the primary control system design. During the design of physical set up, multiple changes were made to the initial plans. Firstly, the battery bank was replaced with a PC Powe r Supply. The PC Power supply was capable of providing constant 5V and 12V power to the 120mm and 80mm fans. The power supply was also able to be connected to a standard 120V AC wall outlet removing the time constraint associated with a battery bank. While some PC power sup plies do require a motherboard return signal to allow the power supply to turn on, it is possible to create a wire jumper to bypass this safety requirement. Please note that bypassing this feature induces a small concern of the power supp ly failing. Students should pay close attention while operating the unit under these conditions. Installation of the PC Power supply did remove the use of pulse width modulation (PWM) of the fans. While control of the fans is still possible, the required p arts were not procurable within the time constraints. Possible switches were researched, but further testing and safety analysis was required. A soldering iron was also required for specific switches but was not available at the times needed. The Arduino w as also found to provide problems while programing for multiple probes. A background in Arduino coding would be extremely helpful when trying to set up the multiple probe inputs and outputs to Excel. Due an error while creating a code based of t he original code for the lab, the progress on developing a new code was lost. For sake of time, the code was reverted to the main code used during the original controls lab. This would allow data collection of pertinent data to set up the initial controls and calcula te parameters for multiple controllers but would only allow collection from one probe After researching the possibilities with the Arduino Uno R3, it was determined that using a larger Arduino, such as the Mega, would provide extra accessibility and wou ld be more feasible to expand the data collection of the program. A possible idea was introduced to attach a computer motherboard with built in graphics and RAM as a minimal test bench that would allow full computing power while running the control system. This was deemed uneconomical so it was not carried out, but if finances were not the determining factor, it would be the best solution. A small hard drive would allow use of Excel and storage of data gathered with

PAGE 9

Page 9 of 23 minimal IO ports allowing communication w ith all equipment, to include the fans. This would also allow for an expansion of the cooling system by using a water PC cooling system. This would also remove the safety risk of the PC power supply bypass. Possible increasing the number of fans was found to give a greater control of the cooling capabilities of the system. Future changes to the design can greatly improve upon the controls set up. Currently, high quality for successful Arduino operation. Installing a series of LED lights with proper labeling would allow easy visual e front of the controls set up to remove the need to look at every side of the equipment to ensure it is functioning. These changes, along with the possible increase of Arduino board size, computing power, and cooling abilities would be made to improve the project design. (Painting is optional, but show some school spirit!) Materials, Equipment, and Construction Materials with Specifications Table 1 List of equipment that has specifications pertinent to the system design. Equipment Specifications Elegoo UNO R3 ATmega328P microcontroller Input voltage 7 12V 5V Electric current: 500MA 3.3V Electric current: 50MA 14 Digital I/O Pins (6 PWM outputs) 8 Analog Inputs 32k Flash Memory 16Mhz Clock Speed https://www.elegoo.com/product/elegoo uno r3 board atmega328p atmega16u2 with usb cable/ Norpro 559 Immersion Heater 300 Watts 7.5 x 5.5 x 3.2 inches 9.1 ounces https://www.amazon.com/gp/product/B076QJBXKD/r ef=oh_aui_detailpage_o02_s00?ie=UTF8& psc=1

PAGE 10

Page 10 of 23 2x Sicce Mimouse Circulation Pump Specifications at 120V: Flow Rate: 82 US GPH (300l/hr) Power Absorbed 6.2W Ampere: 0.084 A Pump Head: 1.8 ft. Oulet: 13 mm http://www.sicce.com/prodottiDettaglio_eng.php/prodo tto=mimouse/idprodotto=4 2x TorqFlo STEER COOL Oil Cooler Thickness: 19mm Height: 63.5mm Length: 235mm Core: 146mm Fitting Size: 3/8 Inch (9.5mm) Aluminum Construction Weight: 2 lbs. Depth (Fin Width): 0.75 Inch https://www.autozone.com/external engine/power steering oil cooler/compressor works power steering oil cooler/577703_0_0 5x Diymore DS18B20 Temperature Probe Cable length: 100 cm. Stainless steel shell 6*50mm. Power supply range: 3.0V to 5.5V. Operating temperature range: 55C to +125C ( 67F to +257F). Accuracy over the range of 10C to +85C: 0.5C. Output lead: red (VCC), yellow(DATA) black(GND) https://www.amazon.com/gp/product/B01JKVRVNI/re f=oh_aui_detailpage_o00_s01?ie=UTF8&psc=1

PAGE 11

Page 11 of 23 Logisys 20/24 Pin 480W Power Supply pin) pin Molex Connector x SATA (Serial ATA) connector 115Vac Operation and CISPR22 230Vac Operation p ot & Function (Chroma)tested in under high Ambient Temperature(50¡C) http://elogisys.com/Show/SKU?SKUID=PS480D 2x 2 Insignia 80/120 mm Fans Airflow Volume: 44 ft^3per minute Fan Size 120 millimeters Fan Speed 1200 RPM Noise Level 22.5 decibels adjusted Bearing Type Dual ball bearing Fan Connector 4 pin Airflow Volume 24 ft^3 per minute Fan Size 80 millimeters Fan Speed 3600 RPM Noise Level 22.3 decibels adjusted https://www.bestbuy.com/site/insignia 80mm case cooling fan black/5976302.p?skuId=5976302

PAGE 12

Page 12 of 23 General Materials and Tools Table 2 List of general materials and tools required for construction. General Materials 1 3 Gallon Publix Water Container 2 50 oz Hanover Green Bean Cans PVC Primer PVC Sealant 7 11 90 o 2 4 way PVC Crosses 10 Crescent Wall Mounts Funnel Wood Screws: (Purchase 50 Piece Packs) #4 x 1/2" #6 x 1 #8 x 2 10x GE Type II Silicon (Lots) Oatey Epoxy Putty OOK 59204 Picture Hanging Kit 25 lb General Purpose Wire Elegoo 17 Value Resistor Kit Elegoo 30 piece Dupont Wire Pack Chanzon 100 piece Rectifier Diode Pack 5x N Channel Power Mosfet (P30N06LE) JBTek Acrylic Arduino Mounting Board Terminal Optimizer Breadboard 100 pack Assorted Size Zip Ties Twist on Wire Connectors Extension Cord 5 socket Surge Protector Blue and Orange Spray Paint Tools Required Power Drill PVC Pipe Cutters Assorted Screwdrivers Saw (Your preference) Assorted Saw Blades (Metal/Wood) Caulking Gun Vise Grips Wire Strippers Drill Bit S et 3/4" Bore Saw 1 Paperclip Electrical Tape Razor Blade

PAGE 13

Page 13 of 23 Construction Construction began with a frame constructed of the general purpose 1x1 boards. The length of the frame was cut to 24 inches. The height of the frame was cut to 20 inches. The overall depth of frame cut to 12 inches. Please note, all these measurements to include for the width of the frame boards and any board joints. First, the four legs were cut to the total height of 20 inches. 2 horizontal trusses were cut at 20 inches and 2 more were cut at 22 inches. 2 depth trusses were cut at 10 inches and 2 more were cut at 12 inches. The short horizontal trusses and the long depth trusses were cut to frame the top of the structure while the long horizontal trusses and short depth on the left and right side of the structure, while still being held within the frame. Please see pic ture for further clarification. Figu re 4 Frame cuts and shelf instillation. legs acted as the four corners of the frame. Then, the top trusses were attached lining the end of the leg up with the top of the trusses. #8 2 hole was bored through the left side of the t op b oard. This hole is drilled where the inlet is desired and can be moved to whatever location the builder requires. Three trenches are drilled, equidistance between the two shelf walls on the bottom board. This is accomplished by drilling 3, 1 a line parallel

PAGE 14

Page 14 of 23 to the shelf boards. The remaining material between the bores is cut with the saw, leaving a rectangular shaped trench, with an approximate length of 4 inches. This trench allows for installation and removal of the tanks if required. The to p and bottom plywood sheets can be attached with the #6 As with all plumbing, designing of the fill and drain system is done on site and custom for each specific job. First, open the aluminum cans, hopefully filled with green beans. You may eat these, they are delicious. Wash the cans thoroughly and find the center of the bottom of the can. Drill a PVC pipe. Drilling a smaller hole and then opening with a razor is highly recommended. Repeat this process for the second can and the water container. It is also recommended to slightly offset holes based on the exact locations for each trench and where t he tanks will best fit within the set up. It is a touch and feel type of thing, so slight adjustments are required. With the locations of the cans set, begin mapping out the location for each of the valves. Proper placement of valves, elbows, and crosses s hould be completed in that order and allow minimal complications during construction. All PVC joints must have a piper that is able to go at least half way the length of the fitting. Each pipe should thus be cut to a length where it is needed, adding half the length of the fitting to both sides of the pipe. A dry fit of the PVC system is recommended to ensure a correct fit. Then slowly mark each piece for correct positioning, and take apart the system. Begin rebuilding the system while applying PVC primer a nd sealant to each piece as instructed on the materials. Complete this process in a well ventilated area and cover all surfaces that will be worked on, Based on personal preference, the design for the drain and fill system is presented in the below picture s. Figure 5 Design of fill and drain system. Once the fill and drain system have been installed, install the two aluminum cans in the correct locations. Use silicon to create a water tight seal around the PVC pipe and any gaps in the can. Next mark out two lines approximately 1/3 and 2/3 from the bot tom of the water container. These will be a guideline for heat exchanger installation. Cut off the top of the water container. The depth

PAGE 15

Page 15 of 23 of this cut is personal preference, but if too much is cut off the guidelines will need to be moved. Once a proper heig ht and location of heat exchangers has been established, center the heat exchangers in the tank at the correct height and mark a location for the inlet and outlet on the water container. Note, the inlet and outlet will be attached outside the water contain er and the pipes will go through the plastic. Once these locations are marked, drill out small holes that require a bit of effort to push the heat exchangers pipes through. Test fit carefully to not break the plastic or bend the heat exchanger fins. Once t he holes are prepared, install the water container into its tank location and seal the connection to the PVC pipe with silicon. Use a lot of silicon while sealing all three of these connections. Allow the silicon to cure overnight and test for leaks. If le aks are present, attempt repairs or redo the steps for the tank with the associated leak. If no leaks are present, fill the bottoms of the can with silicon, level to the top of the PVC going into the can. Apply a heavy bead of silicon to the bottom of the outside of each container to give it a small amount of support. Fill as much of the space from the bottom (drain trench) of the tank with silicon as well. Allow the silicon to cure for three days. Figure 6 Mounting of heat ex changers. Once the tanks are secured, carefully install the heat exchangers into the water container. Use the 25 lb wire to hold them in place and support their weight on the edges of the tank. Silicon around the inlet and outlet pipes sticking through th e plastic. Silicon both sides of these connections but leave enough room to attach the rubber houses that will feed the heat exchangers. Once this silicon has cured, fill the system with water and test for leaks. Repair any leaks to the best of your abilit y. This might require backing up and redoi ng sections of the construction. Attach tubing between heat exchangers and route them into the corresponding tank. The feed tube should be able to reach almost to the bottom of the tank while the return tube should be cut off about two inches past entering the tank. By pumping from the bottom and returning to the top of the tank, on opposite sides if possible, mixing is introduced. Assuming the system is well mixed removes temperature gradients within each of the two controlled tanks allowing the control system to focus on the

PAGE 16

Page 16 of 23 middle tank. The pumps can be attached and installed a t this point. Ensure the use of the hose clamps and tight fits to stop disconnects and leaks. Figure 7 Placement of tanks and cutting of installation "trench". Install the 120 mm hobby b oard and screwing them too both the fan and the plywood bottom. Exact positioning of the fans is personal preference but they were installed directly in front of and behind the cold tank. Air flow was directed towards the tank from both fans. The two 80 mm fans are installed above the tank, blowing air down and towards the right hand shelf of the set up. This is done to keep air flow away from the control tank and to provide direct air flow over the surface of the water. Using the hooks and wire provided in the picture hanging kit, install the fans hanging above the tank as preferred An example is presented in the picture below. 25 lb wire can be used to reinforce the fans position and hold it stable during fan operation. Figure 8 Installation of the fans.

PAGE 17

Page 17 of 23 The installation of the PC power supply is not completed as recommended. Proper installment involves the use of a Power supply installment cage procured from an PC parts store. This cage can be mounted in the same location as the power supply, and the power supply simple slide into the cage and screwed down. Bypassing the motherboard signal requirement is dependent on the specific power supply. Wiring diagrams and manuals from the manufacturer should be consulted prior to carry ing out this step. As it can be a safety concern, instructions will not be discussed on how to complete the bypass to ensure safety notes are read while determining if it is possible. Figure 9 Improper mounting of PC power suppl y. The Arduino installation can be complex. Installation of the breadboard and controller board onto a mounting board greatly increases the ease of installation. The solid state relay must also be installed in close proximity. Again, personal preference g uides the installation process. An example i s provided in the picture below. Wiring diagram and Arduino code are courtesy of Dr.

PAGE 18

Page 18 of 23 Results and Discussion The construction of the set up was very difficult and the time required was much greater than expected. The design was completed from scratch, with each step involving decisions to improve or take away function of the set up. It is important to note that a t this stage, the project could never be a 100% success. This is because the manipulation of the pump speed was found to be implausible under current conditions. While PWM could be used to manipulate the speed of the pump, rapid switching for the pump was found to cause sticking issues with the impeller. Once the impeller would stick, the pump would have to be taken apart and the impeller manually reversed until it was broken free. The pump seemed to suffer no damage, but also did heat up more than under co nstant operation. Manual control of the pump speed is done by adjusting the front cap on the pump. While this could be done, a method to allow accurate control of the pump while still operating within the tanks was not fabricated. PWM of the fans was also found to be an issue. The use of a MOSFET would work as electrical switch but due to the increase power draw from multiple fans and the 12V requirement of the 120mm fans, a different technique was required. While the PC power supply was introduced to the design, solid state relays where not able to be procured within the time frame of the project. Solid state relays that met the specifications required in the set up (3 5V input signal, 12 VDC output) where found and ordered, but a shipping date was never g iven and the order was cancelled two days before the final project presentation. When programing the Arduino code it was found that the number of digital IO pins on the Arduino Uno R3 was a limiting factor. Research into the concept lead to the determina tion that while it was possible, it would be beneficial to go to the next size up Arduino board. The code became quite convoluted and a mistake was made during the programming process. In an attempt to fix the code, major sections of code were accidently c ut into a second file. While trying to determine what was now causing an error, both files were cut and pasted to a non functioning state. Attempting to save, nd i t was determined that using the code from the original lab, while only supporting one probe, was more time efficient and allowed for testing of the set up. After construction, programing, and the final changes were completed, the set up was ready to be tes ted. The three tanks were filled to a suitable, repeatable level. This level was found to be 1/ 2 below the top of the aluminum cans and just at the point to where the water level of the control tank touched the fill PVC pipe. The test set up was turned on with pumps and fans running and the heater off. This allowed for a room temperature stead y state to be reached. A 21 hour test was than run.

PAGE 19

Page 19 of 23 Figure 10 Full 21 hour test results. Many conclusions were derived from the initial test runs of the experiment. First, steady state temperatures were achievable with the constructed set up. Differences in temperatures between all the tanks and positions within the control tank were stark an d evident. The cooling system was capable of maintaining a lower temperature, yet was not able to maintain room temperature. Slow heating of the entire system was occurring, leading to the conclusion that heat input into the system was greater than the hea t removed. As the heater was the sole source of heat input, which will be denoted as generated from here out, 300W times the fraction of time the heater was on calculated a known heat input. The convection of the hot water at the top of the control tank an d the air, along with the total heat transfer between the cold tank and the air was less than this heat generated. Eventually, the cold tank would heat up enough that the interface would be lost and then the entire system would rise in temperature exponent ially. By carrying out an energy balance, an increased to use a more efficient heat transfer method, such as running the cold water through a fan cooled radi ator or using PC water cooling radiators. In terms of use to this project, controlling the rate of coolant flow through a PC water cooling system would be the most beneficial as it also provides for direct control of on the manipulated inputs with direct a ccess to the data. Due to only one probe functioning, the probe had to be moved throughout the system to measure the multiple required temperatures. To do so, an assumed steady state temperature within the hot tank was reached for a specified percent heat er on. Once a 15 minute period passed with no temperature change (+/ 0.06 o C) the probe was moved from the hot tank to the top (hot region) of the control tank and allowed to sit until a constant temperature was being read. This usually required 2 to 3 mi nutes. This process was repeated for the middle, bottom, and cold tank. The order was completed to reduce the temperature jump between readings as much as possible. 0 10 20 30 40 50 60 70 80 90 20 25 30 35 40 45 50 55 60 65 70 0 120 240 360 480 600 720 840 960 1080 1200 % Heater On Temperature (C) Time (min) Complete 21 Hour Run T(C) % On

PAGE 20

Page 20 of 23 Figure 11 Test run showing a step change from 10% to 20% heater on. Table 3 Steady state temperatures at multiple % heater on's. TEMPERATURE ( O C) % HEATER ON Hot Control Top Control Middle Control Bottom Cold 0 23.25 22.62 22.62 22.56 22.62 5 27.19 25.87 23.37 22.62 23.31 10 30.75 29 24.69 23.37 23.62 20 38.13 35.96 25.75 24.25 24.56 30 44.5 40.19 26.31 24.62 24.12 50 54.44 49.44 30.81 24.69 24.87 80 67.31 59.19 34.69 24.7 25.5

PAGE 21

Page 21 of 23 Figure 12 Comparison of steady state temperatures over test run. While it is found that the desired effect of a temperature difference throughout the control tank is evident, a major issue arises. With the heater running at 80%, the interfaces steady state temperature reaches 36 o C. The issue then becomes time. The heat er cannot be increased significantly and the coolant system is already running at max capacity. Any change in the set point that would allow observable results would take an extended period of time. Secondary runs were thus completed letting the system run as long as it needed to reach a complete and undeniable steady state. Below, it is shown that this took upwards of 4 hours. In the original experiment, this time was approximately 60 90 minutes, but the system was running at 10% heater capacity, allowing much greater control of the system. The results of the constructed set up allow no wiggle for the control to operate under. 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 90 Temperature (oC) % Heater On Steady State Temperatures of all Tanks Hot Control Top Control Middle Control Bottom Cold

PAGE 22

Page 22 of 23 Figure 13 Testing to find time to reach a true steady state.

PAGE 23

Page 23 of 23 Conclusion While the results section m ight seem short, there is good reason behind this. With the results shown above, it is not reasonable to continue with the current set up. Proper controls could not be designed under such tight constrictions. Also, the entirety of the multivariable problem still needs to be worked out. Quicker heating with lower usage of the heater to maintain a usable steady state temperature of the interface is required. A more efficient, and controllable, cooling system is required. Control of the flowrates for both hot and cold heat exchangers needs to be developed. When these factors have been implemented and flushed ou t, further progress can be made. Acknowledgements and Sources Undertaking this project was an amazing experience. This project allowed me to roll up my sleeves and use my engineering skill sets. Problems were faced at every corner, but I felt like nothing was impossible. I can only attribute this to my great professors in the Chemical Engineering department. These passion for learning and pass ing on their knowledge to students upholds the highest standards of the University of Florida and positively impact the world, Special thanks to my advisor, Dr. Svoronos. His undying dedication to his students and mentorship have bettered me as a person. Incorporating material that allows engineering students to work with their hands again is invaluable. The only source used are his notes, provided as a textbook for our controls course. Large portions of this experiment were recre ating using his original Portable Control Lab as a model.