PID Loop and Automation Direct PLC

Some important information that I forgot to add:

we dont have the cleanest power in this area of the plant. there was a power spike or something that took out one of the inputs from one of the pump controllers. This also damaged the original plc. there is a new plc now, but the I/O card is the original one.

Direct acting/reverse acting. Is this feature correctly set?

sorry. it's a cooling loop, so i set it to reverse acting.

Ken

What happens if it is set to direct acting (the forum doesn't know at this time what the loop is driving)?

I'm sorry PWSlack, I don't understand what your confusion is. The loop is driving 2 VFD's that are controlling fans on the roof in two coolant towers. The coolant is ammonia. Am I missing something that is relavant to this?

The input that is driving the loop is pressure. When the pressure goes up, we turn on the coolant to drive it down.

This is a pre-existing setup.

I did try direct acting, but I believe that it was before I realized that the PLC was bad.

It looks like you are suffering from Reset Windup, the Reset component of the PID settings is slowly ramping in either direction whilst not in use and causing either a 0% or 100% output to be present when the controller is required to be bought into action. Modern controllers have a feature called "Anti Reset Windup" function, you may have this option in your PLC PID Controller, it may be an optional extra (software) or it may not be available for your hardware. If not available the only option is to force the controller into Manual when not in use, to "freeze" the output, and then force it into Auto when it is needed. This should be able to be done in logic in PLC.

Rok

Hi Rok,

there is a setting called Bias Freeze. I've tried this setting without much success.

I am wondering if there is an issue with my analog I/O card.

at this moment, I have the Gain set at max, and the Integration set to 99.99 (which forces it to 0) and the output should be swinging wildly, but i'm not getting higher then an output equivalent to 15 hertz and it never quite makes it back down to 0 hertz.

Ken

Hi Kandanews;

• For your application you don't need R/S enabled (Ramp/Soak) because it will cause exactly the behavior you describe, go to that tab and de-select it.

• If it is already disabled, try changing the algorithm to velocity, or the PID will continue to increase the output until some expected "position" is met (which is never, you don't have a position controller but a speed controller).

Do one thing at the time, if the first step fixes it, don't mess with it any further and make a back-up of your config.

I hope this helps.

Yahlasit

Thank you Yahlasit for your reply.

The R/S is not enabled. I switched to the velocity like you suggested, but absolutely nothing happened with the pid. Both the bias value and the Output value stayed on 0.

Question? Could there be an issue with my analog I/O card causing problems with the pid loop.

before this the loop worked just fine. It cycled all the way up to 60 when called, then back down to 0 (appeared to be in increments of 15 hertz). Now the loop is very sluggish incrementing by 2 or 3 hertz and it only makes it up to about 15 hertz.

Thanks for everyones answers and support

Have you checked the winding current of each pump phase? Plus the fan current? Check them in the Am vs Pm temps. You setup looks ok for most apps. Could be an I/O board from any vfd too. Oooh ammonia. Very corrosive. Do's wonders for electrical connections. We use a lot of clorine and peroxide. Sometimes a unit just disintegrates.

:) I agree with the corrosive portion of your comment.

I have not gone outside of the controller to do any checks. Could an output card from the vfd cause my loop to be sluggish?

there is no feedback from the vfd to the controller. it all runs off the system pressure input. Only 1 input to the control. the other inputs are only to show that the pumps and fans are on.

Ken

1) You've got the wrong control action. It needs to be direct action, not reverse action.

The industry standard for direct action is when the PV exceeds the setpoint, the output has to INCREASE to deal with the error. If the fans have to speed up to move more air to remove more heat the hotter it gets, then that is a definition of direct action (ISA standards)

20 years ago, Texas Instruments inverted the industry standard definition of reverse and direct acting, and since Koyo makes AD, there might have been a continuing historical anomaly. But according to AD's manual, that's not the case.

AD's definition of control action is on page 14 of the CL05/06 chapter on PID (http://www.automationdirect.com/static/manuals/d0user/ch8.pdf)

pg 14 Reverse Acting Loop
Although the PID algorithm is used in a direct, or forward, acting loop controller, there are times when a reverse acting control output is needed. The DL05 loop controller allows a loop to operate as reverse acting. With a reverse acting loop, the output is driven in the opposite direction of the error. For example, if SP > PV, then a reverse acting controller will decrease the output to increase the PV.

Your setting for 'reverse' action is incorrect. When the SP>PV, the PID output will decrease (slow down, lower VFD Hz output) as it gets hotter. You need the output to INCREASE, not decrease.

2) According to the way I read the CL05/06 chapter on PID (http://www.automationdirect.com/static/manuals/d0user/ch8.pdf)
you should use position algorithm.

pg 9:
The DL05 uses two types of PID controls: "position" and "velocity". These terms usually refer to motion control situations, but here we use them in a different sense:
• PID Position Algorithm – The control output is calculated so it responds to the displacement (position) of the PV from the SP (error term).
• PID Velocity Algorithm – The control output is calculated to represent the rate of change (velocity) for the PV to become equal to the SP.

3) > I have the Gain set at max, and the Integration set to 99.99 (which forces it to 0)

The manual says:
Eliminating Integral Action:
The effect of integral action on the output may be eliminated by setting Ti = 9999. When this is done, the user may then manually control the bias term (Mx) to eliminate any steady-state offset.

It isn't clear from the manual whether 99.99 (presumably seconds) turns the integral factor off. I doubt it. I think that's an integral factor of ~ 100 seconds. A value of 9999 minutes produces so little reset action that it effectively disables reset/integral. But that's a guess.

I doubt your process warrants disabling reset/integral action, but to do so I suspect you'd have to change the integral units to minutes and the setting to 9999.

What I find confusing is that my first reaction was 'reset windup', especially with a 30 second reset term. But you say it takes a day to max out the output. Of course, if the control action is wrong, all bets are off.

4) I am confused about AD's bias deal. I think they mean what was historically called "manual reset", but it isn't clear on my first read. If I were messing with it, I'd turn Bias off, not 'freeze', but set bias = 0 and see what happens, disable bias freeze and see what happens.

Thank you Carl. I will look into this today. Sorry it has taken so long to reply. The coolant has been working, although, no where near any resembalance to any PID loop I have seen. Probably due to the cooler weather.

So that project was effectively put on the back burner.

I stand corrected. When your post popped up, I re-read what the AD manual said and realized I had misread it the first time.

AD quote, "if SP > PV, then a reverse acting controller will decrease the output to increase the PV."

That does indeed invert the ISA standard.

The ISA standard for reverse action is:
PV > SP, output decreases
SP < PV, output decreases
PV < SP, output increases
SP > PV, output increases

The ISA standard for direct action is:
PV > SP, output increases
SP < PV, output increases
PV < SP, output decreases
SP > PV, output decreases

AD definition of reverse action, when SP > PV, output decreases; is the opposite of the ISA standard.

So your decision to opt for reverse action in the AD does fit the action you need.

Don't stand corrected. I appreciate the help that knowledgable persons like yourself take the time to help up and coming knowledgable persons like me.

and I went back and changed to a forward acting loop and the output went the wrong way. but It still does not work properly. I have some help in the form of control engineers coming to visit my plant tomorrow. I will let everyone know what we come up with. there is something wrong in there and i don't believe it is "setup". i believe the problem is a faulty component. and I am betting on the I/O card.