Issue2884 pid autotuning

[mwetter]

This merges the PID autotuning to the master.

[mwetter]

@SenHuang19: This requires considerable more work. Please also see the email I sent.

[JayHuLBL]

What will happen if the setpoint is changed during the tuning process? I see that you have a assert blockassMes3. The assertion will give a warning not the error, so the tuning/simulation will not stop. So, should the tuning stop?

If the intention is to use the setpoint value at the moment when the tuning request becomes true, you may use the sampling value (sam_u_s.y) as input for the tuning process. If during the process the setpoint changes, it would give warning, but the tuning still use the old setpoint value. However, the question still is, should we allow the setpoint change?

[SenHuang19]

Yes, we could stop the tuning once the setpoint changes. However, I am not sure if changing setpoint is the only reason for the square root negative number. Thus, the current implementation stops the tuning once the square root negative number occurs.

[JayHuLBL]

So I am assuming that when there is setpoint change during the tuning process, current implementation will stop the tuning process (no matter if or not the setpoint change causes the negative square root number)?

However, when check the result of Buildings.Fluid.HeatExchangers.Examples.WetCoilCounterFlowPControlAutoTuning, the variable con.inTunPro.y has following profile:

It shows that although the setpoint con.u_s is changed at 2400 seconds, but the con.inTunPro.y keeps being true till to 2493.84 seconds, meaning that the tuning process does not stop at 2400 seconds. Am I missing something?

[SenHuang19]

Sorry for the confusion.
In the current implementation,

• The tuning doesn't stop if the setpoint changes. A warning will be triggered.
• When the tuning stops, we will check if the tuning has been completed successfully or not. If not, we will use the control gains before the tuning.
• The flag that determines whether the tuning completes successfully or not will be false when the negative square root number occurs.

In my previous reply, I was saying we could change the current implementation into:

• The tuning stops if the setpoint changes. In this case, we will use the control gains before the tuning.

But my concern is that we cannot handle the case where the negative square root number is caused by other reasons.

Let me know if we need to have a quick call to discuss it.

[SenHuang19]

08/26/2024 meeting note:

• Need to stop the tuning process once the setpoint changes. In this case, we will use the control gains before the tuning.

[SenHuang19]

Changes have been made in #3987 to stop the tuning process if the setpoint changes. In such a case, control gains will not be changed. If the setpoint is constant during the tuning, the tuning will be processed without interruption. We will check if the negative square root number occurs once the tuning is completed. If so, the tuning is considered to fail and control gains will not be changed.