Control Views - Question of the week

Question #8 - Reset windup - what is it and what harm does it cause?
January 2, 2004

Question:

I have often heard about reset windup and methods of preventing it. What is it and what harm does it cause?

Answer:

Reset windup describes several situations in which the reset element of the controller continues to increase (or decrease) the output of the controller even when the change in output does not cause any change in the process measurement (controlled variable). With no resulting decrease in error, the output will continue to increase until it reaches its limit.

The problem with reset windup is that, when the condition causing the windup is eliminated, the output must “wind down” for a period of time before the decreasing output has any effect on the process.

The original form of reset windup occurred with pneumatic controllers. The standard pneumatic signal was from 3 to 15 psig. However, if the controller was supplied by, for example, a 25 psig instrument air source, the reset of the controller could cause the output to continue to increase above 15 psig, the point at which the valve was fully open, up to almost the pressure of the instrument air supply. Then, when conditions changed and the controller should begin closing the valve, the output would have to fall from 25 psig to 15 psig before having any effect on the position of the valve. Even though the controller should have been closing the valve, the valve was still wide open. This would usually result in an overshoot in the process above its set point.

This form of reset windup was long ago corrected by placing pressure limits on the controller output to keep it within the 3-15 psig signal range. This same technique is used in electronic and digital controllers.

A typical cause of reset windup is where the output from the controller to a valve is limited, perhaps by an override controller. The signal from the override controller controls the valve, perhaps holding it at a limit lower than 100%. Meanwhile, the controller’s output continues to increase, in a vain attempt to open the valve. Eventually, the controller output will reach 100% even though the valve is far from fully open. If conditions change and the valve position should be decreased, the controller will have to ramp the output down from 100% to the actual valve position before having any affect on the valve.

extfeedback.gif (2981 bytes)

A solution to reset windup is external feedback, also known as External reset or reset feedback. This method uses a positive feedback loop to form the reset section of the control algorithm. The actual signal going to the valve (or the valve position, if accurately measured) is fed back into the lag unit that is part of the positive feedback equation. The error multiplied by the gain is, with zero and 100% limits applied, the new output of the controller. This output will be greater than the external feedback if the error is positive and less than the external feedback if the error is negative. If there is no error (process measurement equals set point) the output is the same as the feedback. MORE INFORMATION

Another form of reset windup can occur with cascade control. It is described in more detail HERE.

Other Questions of the week:
12/26/03: Best scan time for loops.
12/19/03: Velocity and full value forms of the PID algorithmn.
12/12/03: Interacting and Non-Interacting Integral and Derivative.
12/05/03: Level switch connections, failsafety.
11/28/03: What is the "standard" PID equation? Gain affects all terms?
11/21/03: What are the benefits of realistic 3-D graphics for operator interface?
11/14/03: Should derivative be used on temperature loops?

Back to Control Views

Page Updated 03/21/04 20:26