Tuning cascade controller


Thread Starter

Rahul singh

I am facing problem in tuning cascade controller, i.e basically temperature control for De-superheater. After rigorous try its maintaining 10-15 Degree difference from set point. Can any one help me in this !!
The two most likely causes are:

1. The output of one of your control loops (i.e. inner or outer loop) is constrained. For example, a valve may be 100% open, in which case you may need a bigger valve. Or it may be constrained to a smaller opening (if so, find out why). Is the plant new or upgraded, or has the mode of operation changed recently (e.g. increased throughput)?

2. You need integral action in the outer (temperature control) loop, but don't have it. Do you know what type of 'basic' (between all the interlock logic, etc.) control loops you have, e.g. PID, PI or P?

Could be one of many things:

- lack of integral action

- controller using one PV and screens showing a different PV (offset due to calibration errors, or differences in location)

- saturation of actuator meaning temp can't be attained

- incorrect offsets / scaling in translation from outer controller OP (usually in %) to SP of inner controller (usually in engineering units)

- additional control logic you may not be aware of

- the inner loop controller mode may not be correct (e.g. just auto, rather than auto with SP from outer)

- the outer loop in manual

I would check all of these things.

Is there an offset between PV and SP on both inner and outer controllers, or is it only on one of them ?

Are you sure it is a permanent offset, and not just one that is moving back to the SP slowly ?

Andy Clegg
Industrial Systems and Control Ltd., Glasgow

James Fountas


Can you forward me a the P&ID of the de-suoperheater section along with steam and water data for the streams?

James Fountas, james [at] fountas.net
From what you describe it sounds as though the intergral action is not working, what settings do you have on the main loop?

What is the output percent on both controllers?

I assume the other loop is water flow control.


I really would like to help you but with such a brief description like this it would be more a guess than a concrete solution. Because you write about a "cascade controller" I'm assuming that there are at least two controllers working in series. I would like to know more about these controllers:

Are these controllers hard- or software-controllers?

How many controllers are involved?

Do we talk about continuous or discontinuous (step-) controllers?

Are these controllers P, PI or PID-Controllers?

To what values did you tune their parameters?

Do these controllers have a kind of threshold parameter on the deviation signal?

What have you done yet to optimize the control response or to minimize the remaining deviation?

What can you tell me about your system, which needs to be controlled ( I really do not know what a "De-Superheater" is)?

My first advice would be to check if you have at least one Integrator (or "I-Controller") in your control loop. You will never reach zero deviation if you do not have a little portion of "I"-control. With a simple P-control a deviation of zero is more or less is pure coincidence...

If there is an I-control within your control loop, my next question would be how do you know that your deviation actually is about 10-15 Degrees? How did you quantify/measure this? Is it a measured or a calculated/derived value? If it is a measured value, what is your measuring-device?What is your measuring range?Did you measure it directly or indirectly (for example by simply reading a value on a HMI-System)?

Looking forward to get some details!

Best regards,