False Results SIMSCAPE Pipe Block for cooling air stream

Hello,

I am modelling a system cooled by an air stream. Therefore I want to include 3 pipe blocks representing the main cooling areas. The flow velocity is below 7m/s. I get weird results, like an increase in the air temperature, when increasing the mass flow. Also my temperature is several Kelvin to high compared to test results. It can happen, that my cooling air is heating my system because it gets hotter then the surrounding walls. Also for some ambient temperatures I get an error at Block Pipe (MA) 5: Temperature must be less or equal to maximum valid temperature. Block Pipe 5 and four are used to represent some heat sources not connected to any wall only to the air stream. So the Pipe and Pipe 5 Block represent kind of the same pipe also Pipe 4 and Pipe 1. Combing both together gives also false results.

Thank you very much for any suggestions or input.

[Yifeng Tang]

It's usually issues with parametrization and/or boundary conditions. If you could share the model and include instructions to reproduce the results, I can help take a look.

[Bastian Kraemer]

Hi, thank you for your reply. I have further run some tests in a thermal chamber with a bread board. For now i can confirm, that the results off the simulation are several degrees off. So I think my implementation of the air stream has general issues. I would appreciate your suggestions to improve my model. To reproduce the results:

1. Run the Base_model.m
2. Run Simulation_Setup.m
3. open top_model_MA.slx
4. run simulation
5. vary "massFlowRate" in "VC/Flow Rate Source(MA)"

The main difference between simulation and test is the equilibrium temperature. The Test showed, that for the nominal case the equilibirums temperature is 13K higher than T_0 (Temperature of cooling air and system temperature at t=0) in the Simulation it is 19K.

Answer 1

[Yifeng Tang]

Hi Bastian,

It sounds like in the simulation, the air stream is taking away less heat than in the test, resulting in a higher equilibrium temperature. I hope I understand your observation correctly. I suggest you look in the following areas:

1. the temperature difference between the H port of the pipe and the temperature of the air inside. Between the pipe wall (H port) and the air inside, the heat flow relies on the correlation for the convective heat transfer calculation: https://www.mathworks.com/help/simscape/ref/pipema.html#mw_fd6e15c3-04d0-4b85-896c-c0fc5ac1b3b2 If you observe a large temperature difference, it's an indicator of large thermal resistance from the convective heat transfer calculation, and you may review and see if the paramters of the pipe make sense. I sometimes use the hydraulic diameter as a tuning nob to adjust the heat transfer area.
2. the temperature difference through out the thermal network. I notice you have many thermal resistance and conductive elements in the model. It'll be a good idea to identify where the largest temperature drop is happening among those components, and compare that to the amount of temperature drop you expect from the network. This may help you identify whether some of the parameters there were not set properly.
3. the driving conditions, including the input heat flux and air flow rate. In your test, does the energy change in the air flow matches the heat flow input? Any chance there is another passage of heat loss in your test? Then verify those driving conditions are properly represented in the model.

There is also a chance that the thermal network isn't correctly connected, but I won't be able to tell since I have no knowledge of the physical system being modeled.

[Bastian Kraemer]

Thank you for your suggestions. I will look into detail into this after the completion of the test campaign.

I am wondering if the behaviour can also be influenced by the 0.001 "smoothing"-factor in the pipe block i read in the forum about this: https://de.mathworks.com/matlabcentral/answers/1911900-why-is-the-pipe-g-block-giving-wrong-results-at-low-flow-conditions

Can the block be imported directly in a custom one to manipulate the factor?

[Yifeng Tang]

I doubt that's the case. You flow rate isn't that trivial. You may, however, estimate the local velocity and compare to the threshold. The source code of the pipe block is visible so you may change the correlations if needed. But first, make sure that's indeed the issue. Locate the biggest source of temperature difference first.