How Often Should PID Control Loops be Tuned?

 

The Age of Just In Time Tuning Has Arrived and It Answers The Question: When Should I Tune My PID Control Loops?

When it comes to PID control loops and the right time for tuning practitioners generally fall into one of two camps. There is the group that adheres to a philosophy whereby PID control loops should be tuned more or less every year. That’s that.  No debate. And there is another group that believes the answer for control loop tuning isn’t so cut and dried. From their perspective, it depends on the loop. It varies based on conditions. There is wisdom in both of these viewpoints. And, odd as it may seem, there is quite a bit of overlap between them as well. Still a third perspective is one that is increasingly coming into focus across the process industries.

There’s consensus that the goal of tuning PIDs is to maintain safe and efficient control of a production facility’s many, often interacting loops.  Staying on top of loop performance through regularly scheduled tuning – say every twelve months – offers a degree of assurance that each PID loop is properly calibrated.  Such a schedule would account for other changes affecting production processes. It’s understood that vessels corrode and foul, motors get replaced, valves are repacked, feedstock is replenished, etc.  These occur with such frequency that tuning every year is often just about right.  In terms of the other camp – the one that suggests tuning “depends” – they would agree unanimously that events impacting a process’ dynamics would necessitate retuning.  So if they’re largely in agreement, then what is the perspective of that third camp?

More and more manufacturers are taking an alternative approach to tuning – just in time tuning (JITT). Using existing process data they’re monitoring the performance PID controllers on both a plant-wide and near real-time basis. Unlike the scheduled approach, JITT allows engineering resources to be applied only as-needed.  In contrast to the dependent method, JITT weeds out those events that have no material impact on process dynamics.  Other aspects leading to the growth of this viewpoint include the following:

  • Size Matters

The average plant has 100s of PIDs and keeping them all optimized isn’t easy.  JITT solutions increase efficiency by pinpointing individual controllers that require attention.

  • An Ounce of Prevention

It’s often easier and more profitable to make minor adjustments than wholesale changes.  JITT solutions utilize KPIs to identify troublesome issues before they become costly problems.

  • Cause and Effect

Due to loop interaction practitioners often miss the root-cause and end up correcting symptoms.  Some JITT solutions include advanced forensic tools for isolating the root-cause.

  • Choices, Choices

Tuning a PID loop still requires some form of test.  One JITT solution solves that by actively capturing each change in Controller Output and modeling the dynamics regardless of the loop type.

  • Baby Steps

Advanced control solutions like MPC fail to deliver value when process dynamics change and change they do.  JITT solutions recognize when dynamics shift and models require adjustment.

Control loop monitoring products that support JITT have become widely adopted because they offer significant value.  Similar to Just In Time Manufacturing practices, JITT monitoring solutions allows practitioners to remain one step ahead of control loop performance issues while responding appropriately to changing process dynamics.  As the use of loop monitoring technologies spreads the adherents of this JITT philosophy are becoming a force to be reckoned with.

 

These resources offer related content:

Functional_levels_of_a_Distributed_Control_System.svg

What is a Distributed Control System?

In a recent blog post we paid tribute to Dick Morley and his pivotal contribution to the process automation industry: the Programmable Logic Controller (PLC). Since the PLC and Distributed Control System (DCS) are both instrumental in controlling complex production processes, people occasionally use the...
Iconic-Mark_Inverted-Color

What’s the Difference between FactoryTalk Historian and PI?

The market for data historians continues to expand rapidly.  Credit that trend to the growing appetite among process manufacturers for advanced analytics.  As the market grows so too does the number of questions about the different products that are available.  While Control Station’s base of...
Iconic-Mark_Inverted-Color

Where Can I Get Control Station 3.7? What Happened to Control Station v3.7?

While not all things get better with time some things do actually improve as they age. Anyone over 40 might argue with that statement, and they might have every reason to do so from a personal and physical standpoint. Even so, just a casual glance...

Still looking for more?

Now that you’ve gotten the basics, connect with our team to learn how our people, processes and technologies can help you optimize.