PID Nightmare

Peter Nachtwey said:

...a big valve that had a 12 inch stroke....

Click to expand...

Hooah! I haven't seen one of those up close in a few decades.

I think... MarkT is going to leave us hanging and we are never going to get our questions answered, I still think if this is a new system the designer of the system needs to be involved, there are way to many variables that will need to be addressed before you could even start on the PID tuning

geniusintraining said:

...if this is a new system...

Click to expand...

Good point, if this is a commissioning, by now the OP may have even forgotten they asked a question here.


And this is the first time they attempt to tune a PID? Oh dear.

drbitboy said:

And this is the first time they attempt to tune a PID? Oh dear.

Click to expand...

We all remember the 'first times'....

I was reheating a thermal oil system, because it was old and had so many additions it was a water boiler to heat exchanger, to oil, back to water, the hot water tank was just over 100k gallons and ran about 500 deg f and was about 16 bar pressure ran good until you had to cool the system down and work on it... I about cried when they told me to reheat the system (everytime), you would start at the boiler and work your way up, the banging of the pipes would scare the living **** out of you, all I could think of was that hot water tank blowing up, I did make one of the heat exchanges jump about 6 inches but the worst was.... when we got a leak in the oil system and the water leak in once the water would hit a hot side and steam off it would be like a small explosion but when you saw a pressure gage go from 2 bar to 20 (max) in less than a millisecond

Being from the heat treat industry, I hate tuning Heat/Cool PIDs. Tuning can be a big issue due to performance of the type of heating device and the cooling device.
There are temp controllers that handle Heat/Cool PIDs rather well.

1746-BTM... add whatever comments you like.


It could be interesting...

NetNathan said:

There are temp controllers that handle Heat/Cool PIDs rather well.

Click to expand...

yes! We have a thermal chamber for heat stressing boards. The control is made by Watlow. We need to test boards from 0 to 70 degrees C do the controller must heat or cool.


However, I don't think that is the main issue. The OPs system is much more complicated than a simple thermal chamber. I still haven't found out of there is feedback from the valve spool position.

I know you are using AB but here is link for an example from Siemens. You can examine the principles and maybe you can find useful info.

https://support.industry.siemens.co...cascade-control)-with-pid_temp?dti=0&lc=en-AR

Peter,
Yes, both valves are actuated with air.

Guys,
Sorry I have not been able to answer quickly. I attend college and have been extremely busy. With that said, the system is a new installation and works perfectly in manual mode, I can actuate my valves 0 to 100% in response to 4-20mA (both heating and cooling without any issues). The problem is strictly PID control. I did a step test a few days ago and using the Ziegler Nichols formula calculated the values for tuning which ended up giving me a very sluggish response when I tested today. The PID slowly creep up and down, thus overshooting. The system uses two jackets in series and is set for heating only (steam) or cooling only (chilled water) at a time, that is why I need the critical damping. The process is monitored by an RTD at the bottom of the tank. The PID is set to Dependent, not sure if switching to independent may help to achieve better control. I got with Rockwell for training, but due to COVID that class is on hold for now.

mark.t, l will re-ask Peters question, do you have feed back on your valves, in other words how do you know if the valves have gone to the desired settings?

PLCnovice61 said:

mark.t, l will re-ask Peters question, do you have feed back on your valves, in other words how do you know if the valves have gone to the desired settings?

Click to expand...

The valve has a digital display that shows the percent open at any given time and also a mechanical indicator at the stem. Have not been able to do any further tuning. I will give it a shot next week again with maybe P and D only settings to see how it goes and also do a few more bump tests for repeatability of dynamics.

mark.t, does the valve give feed back to the PLC, showing % on the display isn't going to help the PID

I feel like this

You can have the PLC change the PID settings when the temperature gets 15 degrees from the setpoint. Look at the Logix help section for the PID instruction for the name of the setting you want to change. For example, the Control Variable Minimum will be .MINCV and the Control Variable Maximum will be .MAXCV. You can use a MOV instruction to change the values back and forth to get the PID to do what you need it to do.