Flow or Pressure Controller or both?

PLC: Modicon Quantum 113-03
I/O:
PIT (located downstream of control valve)
FIT (located upstream of control valve)
FCV (in the middle)

Basically the customer wants a flow control loop using proportional and integral. but...

Wants to add a pressure setpoint that he doesn't want exceeded.

Please tell me if this would work or if there is a better way:

I was thinking of setting up two PI loops. One with a process variable of Flow and one with Presure. Whichever told the valve to go the most closed, its output would be written to the valve.

Note: I would point both PID function blocks' "reset feedback word" at the 0-4095 value going to the control valve. Is there any problem with the one loop being direct acting and the other loop being reverse acting? I don't fully understand the feedback word except that it is supposed to help thwart integral windup while in manual.

Also, I have noticed on earlier posts that you want to have the loop update time match the pid trigger time. I believe that the modicon works like the slc in this regard and needs to be energized every scan else the integral and other params will reset. Am I correct?

Additionally, any other tips/tricks that anyone has for the Modicon PID block would be helpful. I don't understand how to use the "track input" word neither.

Thanks in Advance!

I can't help with the PID in Modicaon - I've done flow and pressure control in the quantum, but I don't use PID for that.

Without knowing the process, I'm going to make an educated guess that the primary controlled variable is flow, and that the pressure is just there to avoid equipment damage or erratic behavior when line pressure exceeds a certain limit.

You cannot control both pressure and flow in the same line. From a fluid dynamics standpint, pressure downstream of your control valve is a result of resistance to flow. If the static lift or the restriction in the piping is unchanged the pressure will increase roughly as the square of the flow.

If the resistance to flow downstream of the valve decreases, the flow will increase. Your PID (or other algorithm) loop will close your valve to increase total restriction and restore flow to setpoint. It's like V=IR in electrical circuits - if you establish two parameters the third is determined.

The best you can do is use the PID loop for flow control. If pressure exceeds setpoint, then decrement the flow setpoint on the PID loop incrementally until the pressure is OK again. When pressure drops you can increment the flow setpoint for your PID until it equals the original operator established value.

I recently had two different suggestions:

1) The PID fuction has some "clamp" registers that can be dynamically modified to limit the output (CV or MV) of the PID instruction. Use the Pressure PID's output to control the clamp of the flow pid.

2) Create two modes. One is pressure control, the other is flow control. Create a deadband between the two loops so that they don't switch back and forth too frequently.

I don't really like number two. Is there anything wrong with #1? I like your idea it is just that number 1 would be a lot easier to implement.

I like Tom's idea

Except I would use the high pressure to increase the flow value. The PV is going to float around anyway. This lets you keep the SP constant.

I was thinking of setting up two PI loops. One with a process variable of Flow and one with Presure. Whichever told the valve to go the most closed, its output would be written to the valve.

Click to expand...

This won't work because one of the PIDs will always have its integrator winding up. If you can keep the integrators from winding up this works well.

Re: I like Tom's idea

Peter Nachtwey said:

Except I would use the high pressure to increase the flow value. The PV is going to float around anyway. This lets you keep the SP constant.

Click to expand...

"increase the flow value"? please explain

Peter Nachtwey said:

This won't work because one of the PIDs will always have its integrator winding up. If you can keep the integrators from winding up this works well.
.[/B]

Click to expand...

I was thinking that the feedback pointer could do this, but I wasn't sure.

The only problem with changing the sepoint automatically is that it confuses the operators. I would have to do something to pacify them.

What about the one pid modifying the clamp registers of the other pid?

Re: Re: I like Tom's idea

ganutenator said:

"increase the flow value"? please explain

Click to expand...

If the flow value is artificially increased then the control valve will shut. Tom's idea of decreasing the SP will do the same thing.
Either will make the error more negative and therefore cause the valve to close. Error = SP - PV.

I would do something like this:

flow = K*sqrt(FIT-PIT)

I am assuming that FIT and PIT are pressure transducers. K is the scale factor to convert and scale the result of the sqrt function to flow.

FlowModifier = min(max(FlowModifier+G*(PIT-PressureLimit),0),MaxFlowModifier)

PV = ModifiedFlow = flow + FlowModifier;

You can see that the FlowModifier can not go below 0 or above MaxFlowModifier. G is kind of an integrator gain or the error between the PIT pressure and the pressure limit. The ModifiedFlow is now used as the PV.

Re: Re: I like Tom's idea

ganutenator said:

The only problem with changing the sepoint automatically is that it confuses the operators. I would have to do something to pacify them.

Click to expand...

You can handle this with an "Impending High Pressure" indicator. If you have an opertor inerface you could add a message that says "Flow Limited by High Pressure" and display both the "Operator Setpoint" and "Setpoint Used" values.