PID with Do-more

[PLCwannabe]

I'd like to use a BRX plc as the PID control for a high-pressure water pump system (2500 psi) utilizing 2 vfds with 2- 20hp motors. This application is going to require very fast reaction times to keep the system from overpressuring. The vfds have a 0-10vdc input to control the speed. MY main concern is, will the plc be able to react as fast to changes in pressure as the built-in PID controller on the vfds? (SQD ATV320D15S6C).

[BobO]

You're gonna have to be more specific than "very fast".

[PLCwannabe]

I've seen these systems go from 1500psi  to 3500 psi (line ruptured) in less than a second when a valve is shut off. I would say it needs to go from 10v to 0 v within half a second.

[BobO]

I've seen these systems go from 1500psi  to 3500 psi (line ruptured) in less than a second when a valve is shut off. I would say it needs to go from 10v to 0 v within half a second.

I would say that BRX can easily do that, as long as the analog is fast enough. Onboard channels are plenty fast, current modules are a bit slower. We have new modules coming in March that will match the speed of the internal channels.

[Controls Guy]

And the control loop shouldn't be your primary overpressure safety anyway.  Every high pressure pump I've seen has a pressure-activated bypass valve.

[PLCwannabe]

That's right. There will actually be a bypass valve set at 2500psi and also a popoff valve set at 3000psi.Though I would like to get it to respond quick enough so it hardly ever goes into bypass.

[Controls Guy]

Good. Seems doable, in that case.  Wire it up and give it a shot.

[PLCwannabe]

For the system described above, I'm also trying to set up  a PID loop to control the temp of the supply water. Setup is a belimo 3-way mixing valve 0-10vdc control . However the runtime of the belimo valve is 90 seconds. this causes the pv to  zigzag sharply above and below setpoint. The belimo also has a 0-10vdc feed back signal to indicate the actual position of the valve. Is there some way to use this position feedback signal within the PID loop so that the loop will not assume valve is at whatever position the control output is at.

[Controls Guy]

If the valve is fast enough to suit your process response criteria but your loop is too aggressive for the valve, then your loop is more aggressive than the process needs.   Use a less aggressive integral gain.  If this only happens following on startup or a step change in SP or load, you could freeze or limit bias when you're far from SP to prevent integral windup and depend on proportional to get you in the neighborhood, only then allowing integral to run.

If you know the hot and cold supply water temperatures, you can estimate the required valve position to achieve the SP mathemagically and discard the PID, or just use it as a trim to the calculated setpoint.

[PLCwannabe]

I've got it working well now. Just lowered the proportional gain to 0.1   . The output now doesn't increase any faster than the valve can open. PV stays within 1 degree of SP consistently.
    Playing with the PID viewing screen I noticed a little glitch. When the PV. gets really close to the setpoint, the scale of the numbers on the vertical axis is too small, making it look like the PV is fluctuating wildly. Is there some way to make those numbers stay within reason,say 115-120 degrees, as opposed to 118.923-119.99  ??

[Controls Guy]

Yeah, I think you can turn off the auto scale and assign upper and lower limits.  I know you can in Do-More trends, and I always do.  I hate the auto scale.

[pbw]

You can disable Auto Scale in the PID Trend.