1. Interdisciplinary Project with South Texas Project Funded by The Nuclear Power Institute 1

2.  Team Members: Matt Langston – CHEN: Senior Kyle Bowzer – MEEN: Junior Ryan Bigelow – MEEN: Junior Matthew King – MEEN: Junior Richard Colunga – ELEN: Sophomore Jennifer Banegas – CVEN: Freshman George Campa – CHEN: Freshman Brent Mayorga – AERO: Freshman  Mentors Graduate Mentor: Andron Creary TAMU Mentor: Mr. Cable Kurwitz STP Mentor: Mr Rick Grantom 2

3. Agenda  Motivation & Project Components  Project Objectives  CFD Analysis  Lumped Parameter Simulations  Experimental Results  Summary  Future Work

4. Motivation & Project Components  Motivation Provide STP’s Probabilistic Risk Assessment (PRA) Team with air temperature profile data after a hypothetical loss of HVAC. Provide STP’s Probabilistic Risk Assessment (PRA) Team with air temperature profile data after a hypothetical loss of HVAC.  Project Components CFD Analysis CFD Analysis ○ Using a SolidWorks created model of room and electrical cabinets Analytical Calculations Analytical Calculations ○ Perform calculations in Matlab using a Lumped Parameter Method Laboratory Experiments Laboratory Experiments ○ Run experiments investigating the heat transfer and energy storage within a solid material

5. Simulation Objectives  Determine air and metal heat up rates during various HVAC failure scenarios  Gain information on when and where the air temperature reaches manufacturer’s critical temperature (104 ° F)  Investigate the effect of energy storage within metal in the transformers using a cabinet CFD model

6. CFD - SolidWorks Model Inlets Inlets Outlets Heater Rods 6

7. Distribution of Heat Loss 29935 watts 3131 watts 1966 watts 12596 watts 200 watts 2234 watts 1759 watts 2519 watts 1523 watts 7 29935 watts

8. CFD Simulations  Computer Simulations Case 1: Steady State ○ Simulates the EAB room’s “Normal Operating Conditions” (50 ° F inlet air temp and 19870 cfm) Case 2: Transient ○ Simulates the loss of one of the HVAC trains (50% air flow) Case 3: Transient ○ Simulates the HVAC chiller failure (73°F inlet air temp instead of 50 ° F) Case 4: Transient ○ Simulates the total loss of HVAC 8

9. Case 1 - Temperature Profile for “Normal Operating Conditions” 9 64°F average

10. Case 1 - Maximum Temperature “Normal Operating Conditions” 10 Max air temp ~78 °F (above main cabinet) Max air temp between cabinets ~64°F (at 5ft) 5 ft

11. Case 2 - Temperature Profile “Half-flow Simulation” 11 68°F average SS after ~ 21min

12. Case 3 - Temperature Profile “HVAC Chiller Failure” 12 83°F average SS after ~ 19 min

13. Temperature Profile “Total HVAC Failure” When Critical Temperature (104 ° F) is Reached 13 Critical temperature (104F) location After ~19 minutes

14. Final Results Plot 14 Case 1 Case 3 Case 2 Case 4 104 ° F

15. Energy Storage in Transformers  The HVAC failure problem is more complicated because it is a transient problem Stored thermal energy flow is important in the temperature history In particular, the heat up of the transformer’s copper windings and iron cores due to the high specific heat capacity.  Bounding the Specific Heat  Based on manufacturer’s specifications of transformer cabinets in the EAB room, metal mass composition values were assumed: Stainless steel: 15-20% Aluminum: 5-20% Iron: 20-60% Copper: 40-60%

16. Bounded Values  Using Matlab, all possible mass combinations were computed Used to determine max, min, mean of lumped specific heat Min CpAvg CpMax Cp 453.3 J/kg*K 504.1 J/kg*K 554.9 J/kg*K

17. Transformer Cabinet Model

18. Cabinet Simulations  Steady state conditions with a uniform air flow across the cabinet’s outer surface  Transient simulation with no forced flow using: Maximum specific heat Minimum specific heat

19. Cabinet Temperature Profile: Steady State

20. Front View of Air Velocity Profile: Steady State

21. Side View of Air Velocity Profile: Steady State

22. Cabinet Simulation Results

23. Lumped Parameter Simulations (1/5)  Objective: Create a theoretical model of the EAB room’s thermal activity Provide an alternative solution method that will predict air heat up rate. ○ Provide confidence in computational model. Allow an additional means of connecting the simulation results with the experimental results. 23

24. In our current analytical approach, the room is reduced to two heat-storing masses, the cabinets and the air. From the basic equation for heat storage, two differential equations can be derived for the air temperature and cabinet surface temperature: Lumped Parameter Simulations (2/5) 24 and

25. Lumped Parameter Simulations (3/5) The two equations on the previous slide can be arranged in a heterogeneous linear system of equations, which can be solved simultaneously through matrix methods to yield: 25 and Where:is the second eigenvalue. (λ 1 = 0)

26. Lumped Parameter Simulations (4/5) To confirm simulation validity, geometric parameters were taken from SolidWorks model P g = heat generation = 85700 W = 292400 Btu/hr M air = mass of air in room = 5657 lbs M m = mass of cabinets = 1638000 lbs Cp a = air heat capacity = 0.241 Btu/lb °F Cp m = metal heat capacity =.117 Btu/lb °F 26

27. Lumped Parameter Simulations (5/5) Once all parameters are known, the constants C 1 and C 2 can be determined from initial conditions (t = 0). Initial conditions used: T air (0) = 63.4 °F and T m (0) = 181.3 °F Once constants are known, equation for T air = 104°F can be solved for t, which may be used to determine T m at that time 27 From SS simulation under normal operating conditions

28. Analytical Solution Assumptions  Uniform heat generation.  The convection coefficient does not vary spatially.  The convection coefficient is fairly constant over the temperature range. 28

29. Overall Approaches  Three approaches: Perform calculations by hand/in Excel spreadsheet. Model simplified version in FloWorks with cabinets lumped together. Use differential equation solver ODE45 in MATLAB 29

30. Analytical Solution Results  Hand Calculations/Excel file (with h = 6 W/m 2 °C = 122.4 Btu/hr ft 2 °F: 30

31. Analytical Solution Results 31 Simplified FloWorks Simulation (h calculated automatically by FloWorks a CFD package):

32. Experiment Overview 1. Goal 2. Approach 3. Experimental setup 4. Tests 5. Results

33. Experiment Goals  Determine thermal conductivity (k)  Benchmark the FloWorks CFD package using experimental results

34. 3.7in 2.5in Experimental Setup ( 1 /2 ) Fourier’s Law

35. Experiment Setup (2/2 )  Aluminum & steel blocks 2.5x2.5x3.7 in  200 W cartridge heater Approximately 95% Efficiency  Block system: Cartridge heater and thermocouples are covered with silicone grease to remove insulating effects of air

36. 4. Tests  Test 1: Insulated aluminum block Power remains constant Determine the thermal heat generation and conductivity (k)

37. 5. Results Test 1: Temp Deviation at 373.15 (deg C) k avg (W/m*K) k_standard (W/m*K)% Error in k 0.28 190 200 5

38.  Convection experiment using same setup  Conduct testing with different materials  Create FloWorks model with the same material and conditions to benchmark simulation results Experiment Future Work

39. Project Accomplishments Used computer simulation results to predict the heat-up rate of the EAB room. Normal Operating Conditions : ~63°F Half flow single train failure: ~79°F HVAC chiller failure: ~ 82.6°F Total HVAC failure: ~19 minutes after total failure (104°F) Derived equations to analytically calculate the heat-up rate using lumped parameter model. Heat-up rate: ~7 minutes after total failure (104°F) Designed an experimental setup that can be easily compared with a Cosmos FloWorks CFD package. 39

40. Andron Creary, Kyle Bowzer, Brent Mayorga, Matthew King, Ryan Bigelow, George Campa, Jennifer Banegas, Matt Langston, Richard Colunga QUESTIONS?