Siemens PID for high inertia process.

[Semuel]

Hi,
Have problem with getting PID to work. Have been using 4x different programmers on siemens, not one could make PID to work as I want.

Problem is I need PID to control biofuel furnace master power. Idea is that such furnace has quite big inertia and irregularity on process.
Usually I use Allen Bradley and there it is easy (or programmer is great).
When PV gets close to SP AB already starts to drop out put and it is possible to fine tune that output temperature is in +/-1c to 2c.
With siemens it is always like maxing out and dropping power only after SP reach. And then I get overheated output on boiler...I can tune how fast PID dial up or down, but not how fast/far from setpoint it starts to do it.
Can anyone can give some directions so I can push plc programmer on right direction?

 

[chantecler]

Hi,

I don't know much about Siemens, but according to my experience with PIDs and temperature control over ovens, I recommend you to start by increasing the "D" part of your PID being careful not to increase it to the point the controller becomes unstable.

 

[parky]

Many years ago, I was taught to start with a reasonable Prop gain, in integral or derivative, decrease prop gain til it starts to become unstable, then increase it till stable, then introduce integral & derivative. May nor work on Siemens PID it's a bit unusual.

 

[kamenges]

You might want to look through some of the information here:

https://controlguru.com/table-of-contents/

Specifically, you want to look at section 2 pretty carefully because everything else builds off the model you will come up with based on the step response.

I don't know enough about the PID function you are working with. Ideally it would modify the integral sum so that the combined output of the proportional and derivative response and the integral sum are not greater than some limit. In your case it is likely only the proportional and integral you need to consider but without the step test data it is hard to know. If the integral sum and proportional output are limited in this way and the integral time is set correctly, as the process variable approaches the setpoint you will basically start swapping integral sum for proportional response at a rate determined by the rate of change of temperature. This is why the step response data is so important. Without knowing the rate of change of temperature for a given output it is hard to have any idea what the integral time should be.

Keith

 

[JRW]

Is it being called from a cycle OB like OB35 ....like it is supposed to be?

 

[parky]

Good one JRW, I forgot about that, just like I did some years ago (pullin my hair out)

 

[lfe]

I worked a lot in the past with the regulation of large furnaces with gas burners but not using Siemens.

As chantecler said you have to increase the D part which is the one that reacts to the changes of the error (Measure - Setpoint).