IMC Tuning For Integrating Processes


Thread Starter


I'm going to use Drum Level as the reference process here as this is what most are probably familiar with.

Also, for clarification, I have the FW Pressure constant, the Drum Level flatlining at 0mm, and the Steam Flow constant, therefore all is in equilibrium.

IMC Equation is:
Kc = (1/Kp)x(Tp/(Td + Tc))
Tn = Tp
Lambda Equation is:
Kc = Tn/(dL(Lambda + Td)^2)
Tn = Td + (2 x Lambda)
Lambda = 3 x Td

I do a manual step test on the FW Control Valve, and as expected the Drum Level starts to Integrate (rise) up. I collect the neccessary process data etc, etc.

I've now directly got:
* Td - Dead Time (sec)
* dL - Level Rate Of Change (mm/sec)
* dCV - Change in Control Valve Position (%)

But I don't have:
* Tp - Process Time Constant (sec)
* KP - Process Gain
* Tn - Controller Integral (sec)
* Kc - Controller Gain

So how can I implement IMC if these are unknown.
I know you can implement Lambda Tuning, but from what I've seen with this, you end up with a very sluggish system that responds quite poorly due to the low value of Kc (please don't comment here on 3 element control, as this is not apart of the discussion).

Am I missing something here, or have other people used different methods?

Personally I've used a mixture of ZN, common sense (CV must be full open/closed due to certain criteria, etc.) and the ol' tune by feel. At the end of the day I achieve the desired result of acceptable and stable drum level with good recovery (mainly due to correctly tuned Feedforwards).

Comments in all shapes and forms are graciously invited...