← Back to Your opinion category

How to Tune PID Controllers…Easily!

A Simple Method for Modeling Processes and Tuning Loops

A Simple Method for Modelling Processes and Tuning Loops

Has a colleague ever commented: “How’d you manage that?” When the question is spoken like a statement, that’s when you know you’ve accomplished something special. Ironically, when it comes to tuning PID controllers there are just a few tricks that can transform you from the guy making such a comment into the guy receiving all the praise.

I’ve tuned lots of control loops during my career but one experience when I was a young apprentice clearly stands out. With my time in the military and at university in the rearview mirror I started in industry as a Turbine Specialist with an engineering services company.

For sure – I was as confident as I was green. When I took my first assignment in support of a Senior Field Services Engineer I was certain that I’d deliver the type of service that would have my colleague singing my praise and the customer begging for more. It’d be textbook!!

The customer needed help tuning several PIDs and I’d scored pretty high in Process Control. But as the saying goes: A funny thing happened on the way to the forum.

Damien Munroe

 Damien Munroe 

 European Integration and Sales, Control Station 


That “recipe” has been my go-to resource ever since…although now I don’t have to rely on graph paper and a stopwatch!

The customer was an Oil & Gas producer located in Vietnam. Like most processes on an oil platform theirs relied extensively on Pressure, Level and Flow control. What immediately struck me was the highly dynamic nature of the plant’s controls – and how different they were from textbook examples.

The customer struggled to meet design capacity due to excessive process variability in its separation vessels. Recall that data collection wasn’t a standard in the 1990s so data trending was limited and software wasn’t yet an option. This job would require us to tune the loops manually, and the methods I’d learned in school were ill-suited for the task.

While there is a lot of knowledge to be extracted from textbooks I quickly learned that there’s even more insight to be gained from first-hand experience. As my concern rose minute by minute my colleague coolly engaged the plant manager with a handful of questions.

Equipped with an understanding of the process’ control objective and knowledge of changes that precipitated the increase in variability, he outlined a very simple procedure that we’d follow to tune the customer’s loops. That “recipe” has been my go-to resource ever since…although now I don’t have to rely on graph paper and a stopwatch!

While my last post offered tips for collecting good data for tuning, this one covers a simple method for calculating the three model parameters that are used to derive a PLC’s tuning coefficients. 

1. How Far: Process Gain, Kp

Process Gain is a model parameter that describes how much the Measured Process Variable changes in response to Controller Output changes. Essentially it describes “How Far” the process moves in response to a given change.

When manually tuning PIDs be sure to use trended data that shows the Process Variable moving from a steady-state in response to a change in the Controller Output. The value for Process Gain can be computed as the steady state change in the Measured Variable divided by the corresponding change in Controller Output.

Be aware that Process Gain should be based on the same engineering unit values that are used in the process. Also, know that the controller will not use those engineering units. Rather, the controller will use the percent span of signal - it will be necessary to convert the Process Gain into units that reflect the “percent span” associated with your controller.

Tuning PID Controllers: Process Gain

Calculate the Process Gain by determining the change to the Measured Process Variable (PV) and then dividing it by the associated change to the Controller Output (CO).

2. How Fast: Process Time Constant, Ƭp

The Process Time Constant is like a clock that further explains the Process Variable-Controller Output relationship. Specifically, the Process Time Constant describes “How Fast” the Measured Process Variable responds to changes in Controller Output.

To calculate the Process Time Constant review the same trended data. It’s important to identify the time at which the Measured Process Variable first shows a clear and visible response to the bump test.

Next, determine the time at which the Process Variable settles at its new steady-state value (i.e. the finish). The value for the Process Time Constant is equal to 63.2% of that total change.

The smaller the value of the Process Time Constant, the faster your process reacts to changes. The ability to control a process is a function of 1) the speed of the Process Time Constant relative to 2) the speed of the last model parameter.

Tuning PID Controllers: Process time constant

The Process Time Constant is found by identifying the time at which the measured PV reaches 63.2% of its change then subtracting the time at which the PV change visibly starts.

3. How Much Delay: Process Dead-Time, θp

The time that passes from the moment a change to the Controller Output is initiated until the Measured Variable shows a clear initial response is called the Process Dead-Time. This model parameter indicates “How Much Delay” exists in a given process.

Process Dead-Time arises because of transportation, sample or instrumentation lag. Calculating a value for Process Dead-Time is relatively straightforward. Begin by identifying the time at which the controller output is changed then establish the time at which the Measured Process Variable responds.

Process Dead-Time undermines the PID controller’s ability to maintain stability as it limits the speed by which a controller can react to upsets. It is for this reason that Process Dead-Time is widely referred to as the “killer of control”.

Tuning PID Controllers: Process Dead Time

Calculate the Process Dead-Time value by determining when the measured PV clearly responds to the change in CO, then subtract the time at which the CO was initially changed.

Once we calculated the three model parameters the hard part was done. It was easy from there to convert each set of values to tuning coefficients that worked with the platform’s PLCs.

During that first experience tuning PID controllers I learned two important lessons: first, textbook knowledge has its limitations and, second, methods passed down from experienced professionals are like gold.

At the end of our visit the customer was indeed requesting additional support services, and I was grinning at my colleague while both questioning and stating: “How’d you manage that?”

Most industrial control loops are tuned infrequently, affording practitioners few opportunities to apply their knowledge and often resulting in skills to slip over time. 

Fortunately workshops that focus on proven best-practices are widely available. These workshops offer a practical means for keeping skills sharp.

If attending one isn’t an option whether due to budget and/or time constraints, then there are guides like this one that can prove helpful

Damien Munroe


Damien lends a view of process control and automation that’s born of his experiences in military aviation, offshore oil and gas, precision pharma, and semiconductor manufacture. His education at the

Read Full Bio

Related news

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Don’t miss a single Viewpoint!

Process Industry Update Enewsletter

- Subscribe to our weekly newsletter & keep up-to-date with latest news & events, products, technical articles & Viewpoint posts

Send me great stuff

Add your Viewpoint

Become a Thought Leader

- Enhance your credibility, increase brand recognation and boost your reputation as a key influencer in the industry

Find out more

wolfson short courses 2019