Using Percent Time in Normal to Uncover Tuning Issues

Even Simple Metrics Such as Percent Time in Normal Reveal Issues that Negatively Affect Performance and Signal the Need for Controller Tuning

It’s easy to get tripped up if a shoelace is untied. Just ask any child. While that’s a lesson most learn early in their lives each of us can occasionally slip up and find ourselves at risk. A helpful remark from a family member, a colleague, or even a stranger is all we need to remedy the situation. It’s awareness that allows us to make the correction and carry on without incident. The same applies to process manufacturing. As complex as most production environments can be it’s the benefit of awareness that allows staff to make the necessary corrections.

Percent Time in Normal tracks the relative amount of time that a PID controller spends in its designated or ‘normal’ mode of operation.  Pretty simple, right?  What’s most interesting is how much it reveals about a production facility’s day-to-day operations.  More often than not PID control loops are operated in a ‘non-normal’ mode due to ineffective control which frequently stems from poor tuning.  Consider the following:

  • Product Changes

Different products often involve different process dynamics.  Unless a PID’s Gain is designed to accommodate the dynamics of the new product, the process can struggle to effectively maintain control. Increases in variability and corresponding decreases in stability routinely result in a shift from automatic to manual operation.  A simple adjustment of tuning parameters can provide the necessary correction.

  • Mechanical Adjustments

Repacking a valve or other final control element (FCE) almost always impacts the associated loop’s performance as it can induce Stiction or other bad behavior.  If replaced with a different sized FCE, then even more profound deviations in performance can ensue.  Again, an increase in volatility often drives staff to shift the loop to a more manageable mode (e.g. Manual).  While tuning can’t correct for Stiction, it can improve control of an under- or over-sized FCE.

  • Process Interaction

A loop that exhibits increased variability isn’t always the culprit.  Quite often the erratic behavior is a symptom of something else in the production environment that’s changed, and adjusting the mode is one way of taking matters into your own hands.  Most loops are highly interactive and they respond to changes in upstream processes so it’s key to isolate the root-cause of that uptick in variability.  Addressing the root-cause and/or tuning the controller for disturbance rejection are both options to consider.

While monitoring Percent Time in Normal can help process manufacturers to avoid obvious issues, this and other key performance indices are not tracked by the DCS.  Technologies like Control Loop Performance Monitoring (CLPM) focus specifically on the health of a given facility’s regulatory control, enabling them to understand performance from a different – and equally important – perspective.

 

These resources offer related content:

info overload

Fewer Metrics Leads to Better Results

Big Data and IIoT have captured the automation industry’s imagination. A key take-away from all that’s been written is that plant data holds the key to making many production processes safer, more efficient, and more profitable. While there’s growing evidence that Big Data and IIoT...
Iconic-Mark_Inverted-Color

Accelerating Start-Up, Optimizing PID Control

Control Station has been fortunate to partner with leading, global OEMs like Yokogawa.  Award-winning products like csTuner are helping process manufacturers at all stages of operation.  As GranBio in Brazil recently shared, csTuner is especially valuable during start-up when highly variable process conditions are the...
Industrial,Technology,Concept.,Communication,Network.,Industry,4.0.,Factory,Automation.

PI Asset Framework: Analytical Pros and Cons

OSIsoft’s PI System dominates the data historian market. A reason for the PI System’s success is its use of the Asset Framework (AF) feature. First introduced in 2007, AF provides PI users with an intuitive structure for organizing and using process data. With better organization...

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.