PID Tuning Step by Sep (Part3)

Continuing with PID controller basics .. But better to refresh with previous 2 parts (1) & (2)

We discussed the PID main equations, Controller types (direct & reverse), physical meaning of each controller part (P, I & D) and how to achieve aggressive/non-aggressive control action. Also discussed PID controller main features and addressed the "Integral wind-up" issue.

Will complete here with the solutions provided for such an issue ..

Hope to be useful ..

Q12. How to avoid "Wind-Up" issue?

"Wind-Up" issue could be avoided by utilizing "Output Tracking" & "Reset Feedback" controller features .. they are designed to overcome this problem.

Usually we are facing "Wind-Up" issue in the following 2 cases (Needs to review part-2):

1 - In the case of "Wind-Up" due to manual/auto-switching. This way "Output Tracking" feature is to be utilized. It will allow the initial value of "I" part contribution (“b” in Figure below) to track a proper value whenever the controller is in manual and cancel "I" part accumulations:

So when controller is back to Auto it picks up from last controller O/P value (which is manually entered by the operator during manual mode) and begins to control from this new starting point .. instead of jumping to some previously-calculate value from error accumulation.

2 - Another case of "Wind-Up" is happening in Auto mode .. while PV stays deviated from SP for a long time. This could happen for any reason. "Reset Feedback" feature will be helpful here.

In such a case .. the downstream Block (AO in the below example) will send a "Back Calculation" feedback signal to the controller, informing it with the status of the final control element (Valve is good and responding or it stucked!) .. And based on this signal the controller will take the decision if it is efficient to keep winding up the output or ceasing the integral part.

Below is an example for PID block configuration (within DCS) for "Reset Feedback" feature (Example for illustration only):

In general .. for PID best performance and results, the vendor manual should be consulted for blocks interconnection and parameters configuration (This is for soft or physical PID) .. this will shorten a lot of efforts and guarantee the best utilization of the system features.

So far we discussed "Output Tracking" & "Reset Feedback" features .. last PID feature to discuss is "Setpoint Tracking", which is our next question ..

Q13. What is the benefit of using "Setpoint Tracking" feature?

"SP tracking" feature is simply equalizing SP and PV (SP = PV) while controller in manual mode (So SP will follow/"track" the PV value whatever that value happens to be). Finally when back to Auto .. SP freezes at last PV value, and controller continues to control the loop at that SP value. Of course, the operator is free now (in Auto mode) to adjust the SP value as needed.

Without this initial state of (SP = PV) .. Controller will transfer from manual to auto with existing error (difference between PV & SP), while transfer at (SP = PV) is better to have what is called “bumpless transfer” .. So the controller transfers without sudden change in O/P value to the final control element (Valve/Pump).

Well .. For some applications setpoint value should remain fixed at all times, or it should be set remotely from another part of the process loop. This way "Setpoint Tracking" feature shouldn't be enabled. “Bumpless Transfer” here could be achieved using "Output Tracking" feature only .. which will work alone this time.

So when transfer from manual to auto there will be a deviation between SP and PV .. but the controller will react to this error by gradually moving the valve to a new position and "I" control action will not stop integrating till the error is eliminated.

Ending of this part .. we finished discussing the important features of PID .. which are mandatory to be clear for the PID tuner. Next part .. will start inspecting the process and testing it, to be able to classify its dynamics and characteristics .. before start tuning.