Controller tuning


Thread Starter

Dave Nelson

Can anyone suggest any good software (preferably cheap) for tuning process controllers e.g. flow and pressure loops?


Chris Jennings

I would probably suggest you should invest in some training first. Magic tuning software isn't going to solve anything unless you understand what you are doing. I have seen a number of situations where someone's idea of tuning
might look good on paper or in theory but in practise has caused more problems than it solved.

Find a training provider that deals with your industry on a regular basis that can provide a good introduction on control loop theory and process control optimisation. This will help you understand what good loop control tuning is and how to identify problems with control loops.

One of the most important things you will learn is that tuning doesn't go bad, the process or process control elements change and that causes the tuning to no longer be optimal for control. In most cases it is best to try and reverse the change that is causing the problem, rather than retuning the loop.

Chris Jennings

Martin Dufford

I agree completely that tuning packages can't produce magic. It all depends on the skill of the person doing the tuning. I use tuning software to give a ball-park figure and take it from there.

Packages I've used are:

Protuner - not cheap but reasonable results.

Expertune - again not cheap.

I've taken a look at several other packages such as Bestune and Intune. These are generally in the £300-400 range. I bought a copy of U-Tune ( recently, partly cos I know the guy and mainly cos its only about £100. It worked well with a turbine diesel fuel system last week and its easy to use.

All the best
Thanks for the comments Martin. Although I'm probably rather biased, I have a lot of faith in U-Tune's results, in particular when there's lots of process noise. That was actually the reason for designing the package in the first place, since the other packages I've used were really thrown by disturbances in the time delay between making a step change and the PV moving.

One point to bear in mind is that any tuning exercise has to consider not only the process dynamics (which is the bit the tuning packages do) but also the condition of the instrumentation and the overall objective of the controller. For example, a slugcatcher generally uses a level controller to 'smooth' the outlet flow: tight tuning parameters will prevent the vessel working properly and will introduce instability throughout the downstream process. In this case, I'd still use U-Tune or similar to calculate the process dynamics, but I'd probably use non-linear techniques such as error squared in place of the PID controller.

John Greene
Process Control Consultant
Contek Systems Ltd
Aberdeen, UK