Understanding PID tuning

Proportional-integral-derivative (PID) loop performance is often overlooked once the system is commissioned and seems to be functioning. Those loops are often not thought of again as a means of continuous improvement. They function “properly” and so they are often ignored until something goes so out of whack that it pops up to the top of the issue pile. Those incidents are few and far between and they don’t get much attentions. However, operational incidents are often responsible for critical components of product quality and production rate that can have a sizeable impact on operational efficiency.

There are some definitive engineering and mathematical steps to take when tuning a loop, but there is also a lot to be said for experience and understanding how to tweak things based on the loop makeup and response. It can certainly be a challenge those first couple times, but that hard-won experience on one application or process can often serve engineers well in other areas.

System integrators  typically interact only with the tuning parameters of a loop during the start-up. Once the control system commissioning is complete, what happens to those loop tuning parameters and overall loop performance is unknown and falls into the category of mysterious. We always wonder, do operators turn loops into manual (or even off) to control the system themselves?

Do different operators feel they have a better handle over the process than others and frequently tweak the loop to their “right” numbers? Do those loops responsiveness lessen over time due to degrading equipment and real-world conditions? Finally, does the initial system process design always capture all phases and modes of operation where alternate PID tuning parameters would be better than the originals?

Correcting sub-optimal PID performance

Control loops often are set up and tuned initially and then forgotten or ignored unless something “major” happens. If “nothing major happens,” the lack of attention to loop performance creates a breeding ground for quality issues, significant efficiency losses and operational inconsistency. This is often what helps to drive some of that manual intervention as things aren’t bad enough to cause obvious issues, but still sub-optimal enough for operators to try to make it a little better.

There are many reasons why a loop might be suffering from sub-optimal performance. One reason is mechanical wear or failure. These loops contain a variety of physical components subject to wear and breakdowns. It could be that proactive maintenance work is not being done on those components.

Regardless of whether users follow a time-based preventive maintenance schedule or a condition or analysis-based predictive maintenance schedule, proactive tasks and replacement of worn parts is important to optimal performance of control loops.

Six things to do when a PID loop underperforms

Whether it’s applying grease to prevent physical binding or replacing an actuator that isn’t moving as quickly, each element needs to be properly functioning to give the control logic its best chance at success. When a PID loop is under-performing:

1. Take a look at the physical situation first. It is easier to re-tune than rebuild a control valve, but the underlying issue remains and will get worse.
2. The logic supporting the PID block also should be considered. There are many programming approaches and techniques that can be coupled with the PID algorithm to improve the loop’s consistency and performance. For example, users might have a loop that finds itself a good distance away from the setpoint relative to the potential process variable change due to changes in recipe or loop dynamics. In this case it can be beneficial to have the code drive the loop in manual at a high output until they get into a defined range. When that happens, users switch the loop into automatic mode and allow the algorithm to take control.
3. Another beneficial coding component when a PID loop under performs is alarming. Setting alarms around taking too long to reach setpoint or too much variance while “at” setpoint can help flag issues in real time and/or trigger additional programmatic interventions. This is useful in minimizing response time when we do need to do some of that reactive maintenance.
4. Signal filtering also is an important consideration when PID loop underperforms. During startup, users are concerned about getting the loop to control and respond correctly. It might not be obvious that noise or other minor fluctuations on your process variable are causing the loop response to be jittery. A knee-jerk reaction would be to de-tune the loop to provide a damped response, but you can better deal with it by filtering the process variable (PV) signal.
5. There also is the PID algorithm itself to consider. It is important to make sure that the control doesn’t stretch the capabilities of the loop. If the tuning parameters are set too aggressively, users run the risk of causing physical issues in the process. For instance, users might cause water hammering by slamming valves fully closed or open too quickly. This can lead to damage of other elements in the system.
6. There also might be different scenarios the loop has to operate under. While the setpoint might not change, there might be other things going on in the process that affect the loop differently at separate times, such as seasonal or product-based changes in viscosity. This might lend itself to having different sets of tuning parameters that can be loaded in depending on the surrounding circumstances.

Given the current advancements in data and analytics technology, we are in luck and have options to solve this issue. With data capture and storage available, it is easy to gather loop information to provide better insight into how they are performing and where improvement is available. By aggregating this data, it is possible to identify changes and dips in performance that aren’t as apparent monitoring the loop in real time.

More manufacturers looking to use machine learning (ML) to evaluate their process-critical loops performance over time. This evaluation yields great insight to improve operations and know what is happening with the system when no one is watching. This approach provides a more structured and scientific “defense” to combating control loop performance than just the tinkering “art” of old.

Brian Fenn is COO at Avanceon, a CFE Media content partner. Edited by Chris Vavra, web content manager, Control Engineering, CFE Media and Technology, cvavra@cfemedia.com.

MORE ANSWERS

Keywords: PID, system integration

CONSIDER THIS

What benefits can your plant gain from constant PID monitoring?

Do you have experience and expertise with the topics mentioned in this content? You should consider contributing to our CFE Media editorial team and getting the recognition you and your company deserve. Click here to start this process.