PID with multiple Sources

I was wondering about what would happen if I had a PID loop and I changed SP and PV to drastically different values. Example, I have a burner with temp feedback, air feedback, and gas feedback. When certain conditions are met I want to use Temp feedback and setpoint on the gas valve. When those conditions are not met I want to use gas feedback and setpoint.

If all feedback and setpoints were scaled to the same range what effect would a sudden PV and SP change have on a PID loop. Such as PV=50 SP=50 and a sudden PV=70 SP=70 have on PID

Depending on the PID Gain factor, such a drastic change could cause an upset condition. The usual solution to go from one situation to a drastically different one is to set up a ramp that gradually makes the Setpoint change over some time period.

So if the rate changes were equal the SP and PV (all scaled to set range) would ramp up to the new points equally and not upset the CV. The transitions I am looking at are at equilibrium points so all transitions should be to where SP=PV within a few points.

I would also think that the three different controls will have very different tuning parameters. You might want to have three separate PID loops and just select the one you want then ramp the CV to the new value.

Such as PV=50 SP=50 and a sudden PV=70 SP=70 have on PID.

Click to expand...

I do not understand you. Normally you make a change in the SetPoint, wait for the Control Variable to change and cause the Process Variable to reach the new Setpoint. It will not happen immediately, but after some time period. Of course you can switch your PID instruction to Manual Mode and set the Contorl Variable directly, but normally you cannot manually change the Process Variable. The PV is not directly under your control. If it was, you would not need a PID loop and a controlling device, and we would have to change your name to God.

Example, I have a burner with temp feedback, air feedback, and gas feedback. When certain conditions are met I want to use Temp feedback and setpoint on the gas valve.

Click to expand...

Are you taling about the Startup Phase as opposed to the Running Phase? Usually most people do not try to control the burner startup, but use a pre-certified Burner Relay to get the thing ignited and operating at Low Fire. Only then is the PID instruction allowed to take over and control using the temperature input.

I am thinking about using cross limiting control on my burners which, by design uses 3 different loops for the temp gas and air. I am trying to do it in 2 loops as the controller the boss is looking at is only a 2 loop controller and cause im curious. Right now the 2 loops use temp feedback to control a gas valve and a ratio of gas feedback to control the air. I wanted to try cross limiting cause I am new and am expanding my experience. Since gas and air have to remain in ratio my thinking was this.

Scale everything to a set PV SP range. Use a selector bit to control Gains, and what setpoints are feed into the SP and PV registers.
So on temp lower than setpoint i would send Temp SP, PV, Gains to the air loop, and a ratio of Air PV to the gas SP(loop register) and Gas PV to Gas PV(loop register).
On temp Higher than SP switch so Temp settings would load to gas loop and the Air loop would have a ratio of Gas PV to air SP and have air PV be air PV

So the temp would go to appropriate loop to have air lead gas follow on heat and gas lead air follow on cool and the other loop would remain in ratio.

Here is a good read.

http://www.google.com/url?sa=t&rct=...vMDpCg&usg=AFQjCNEGkZi4M_JuyTGHDj2d90E4mUcERQ

Mickey said:

Here is a good read.

http://www.google.com/url?sa=t&rct=...vMDpCg&usg=AFQjCNEGkZi4M_JuyTGHDj2d90E4mUcERQ

Click to expand...

That was where I was pulling my information about the cross limiting technique. I have a 2 loop ratio system that works decently except for one or two times air linkage breaks and had overtemp. So I looked at cross limiting but the controller I am allowed to upgrade to only has 2 loops and its the model my boss tells me the safety specs allow for without causing a complete safety system upgrade(which we cant afford) So I thought why not try to do it in 2 loops.

Which lead me to consider how a PID loop would respond to that kind of change

You will have to use a cascade control, so 2 PID loops.

This is easy in a fuzzy controll system.

The PID does balk at the large change in both PV and SP causing the CV to bottom out or max out and takes too long to recover. Testing with logix 5000 pide and setting the integral sum to 0 and setting CVprevious to current CV, during the SP and PV change, does allow the control to work but loses all the benefits of the integral response, and these settings are not available in the controller i will have to use.

Darkzadow, what are you really trying to do? You can't control to 3 set points with 1 PID output.

I am trying, to see if i can, implement a cross limiting burner control system using 2 pid loops as opposed to 3. I was trying to switch the SP between Temp setpoint and a ratio of the PV of the other loop. So i was trying to control a single CV using 2 different SP/PV sources.
While I was able to switch between the sources I encountered some expected problems. When a new SP/PV is used it throws CV to the limits. Correcting this by clearing integral and setting CV to keep valve position would allow me to resume control after switching but would induce poor performance as pid . Ramping to the new values essentially froze the pid until the new values were acquired

I think you really need to consider using 3 PIDs. Don't mess with the scaling. Tune each one as if it was stand alone. Use the PID with the lowest output. You must make sure that the PIDs that aren't the lowest do wind up their integrators. You will get smooth transitions between the 3 PIDs then.