Issue with Basin Level Fluctuation After Parallelizing Cooling Towers in MATLAB Data Center Cooling Example

Problem Overview

I am using the "Data Center Cooling" example provided by MathWorks and have made some modifications to the cooling tower system. Specifically, I parallelized the cooling towers to simulate a more realistic configuration. However, after making this modification, I observed that the basin level of the cooling towers fluctuates abnormally during the simulation.

Modifications I Made

1. Parallelized the cooling towers (previously there was only one).
2. Adjusted related flow rates in cooling tower water pump to distribute water appropriately.
3. Maintained the same heat load and system parameters as the original example.

Questions

1. What could be causing the abnormal basin level fluctuations after parallelizing the cooling towers?
2. Are there any specific MATLAB/Simulink settings or solver configurations that might need to be adjusted when adding parallel cooling towers?
3. Are there any other better layout for parallel cooling tower modeling?

Water level of basin in one cooling tower:

simulink layout:

I also attach my simulation model. Thank you for your help!

Answer 1

[Yifeng Tang]

It looks like those two T-junctions are creating different flow resistance into the two towers. As a result, the inflow can fluctuates based on the instantaneous flowrates driven by the pumps. Removing those seems to make the level fluctuation within 0.1m. See picture below and attached model.

A few other things I noticed with the model you uploaded.

(1) the initial water level is 5m, while the partially filled pipe next to it is only 3m in elevation. In the original model, the initial water is 0.5m, so the pump needs to overcome this 2.5m of heat to pump water. In your case, the elevation is negative (3-5). Please check whether 3m is still a reasonable value.

(2) the max rpm in the pump drive is set to 1200, instead of the 2400 nominal value in the pump parameter. Is this intended? Looks like the flow is still sufficient at this point. If you intention was to reduce flow, change the pump parameter (nominal flow) may be more appropriate.