Imagine a heat exchanger cooling down city water to make ice water and send it all around the factory. On the loop are multiple loads that may or may not consume ice water and basically heat it up on the return.
I have an on/off pump on the glycol side with a control valve, and a VSD pump on the ice water side. Temperature is controlled with the control valve choking the glycol supply if the loop supply temperature falls below setpoint, which is close to freezing. Tuning is done via PID, using only P and I.
Now I get that I have to be miserly with my constants since by the time the controller sees any reaction, my valve will have modulated too much and might bring temperature to point that is too hot or cold, leading to oscillations and the risk that "too cold" might be "heat exchanger froze" and "too hot" might mean "this product is useless". So the valve has to react slowly so it does not overshoot too much the point where the glycol supplied would be enough to keep water flowing at optimal temperature for any given charge. But they can't be too mellow either else the system will never cool down enough.
With the length of the loop, the number of intermittent loads, and my cooler/hotter ice water coming back after a few minutes and further affecting the supply temperature, I'm finding it difficult to appropriately tune the thing. There is apparently also a load that dumps some water, lowering the amount of water in the loop. In school we often tuned level or flow loops and they were artificially quick (small "workbench" installations). Any tips and tricks for these kind of slow going loops?
I have an on/off pump on the glycol side with a control valve, and a VSD pump on the ice water side. Temperature is controlled with the control valve choking the glycol supply if the loop supply temperature falls below setpoint, which is close to freezing. Tuning is done via PID, using only P and I.
Now I get that I have to be miserly with my constants since by the time the controller sees any reaction, my valve will have modulated too much and might bring temperature to point that is too hot or cold, leading to oscillations and the risk that "too cold" might be "heat exchanger froze" and "too hot" might mean "this product is useless". So the valve has to react slowly so it does not overshoot too much the point where the glycol supplied would be enough to keep water flowing at optimal temperature for any given charge. But they can't be too mellow either else the system will never cool down enough.
With the length of the loop, the number of intermittent loads, and my cooler/hotter ice water coming back after a few minutes and further affecting the supply temperature, I'm finding it difficult to appropriately tune the thing. There is apparently also a load that dumps some water, lowering the amount of water in the loop. In school we often tuned level or flow loops and they were artificially quick (small "workbench" installations). Any tips and tricks for these kind of slow going loops?
