1

I am trying to design a PID control (or perhaps just PI), for a system which uses evaporative cooling. The product starts very hot and is cooled with water, which is what I am attempting to control. With evaporative cooling, if too much water is applied, it can actually insulate the product and slow the effect of cooling the product. I will be measuring both temperature and moisture, and want to use BOTH to feedback to my water control. At the end I want both a cool and dry product, but you can see these variables would be at odds with each other taken independently. How do you incorporate both variables into a single PID?

Below is a graph to help illustrate.enter image description here

Jonas
  • 121,568
  • 97
  • 310
  • 388
Beutler
  • 83
  • 1
  • 8
  • I've considered using either water usage or moisture as a stop gap for the temperature control, however, several other variables will come into play, such as ambient temperatures or humidity, which vary wildly with season. – Beutler Jun 18 '18 at 14:41
  • Probably you will use temperature to calculate the ideal moisture target, and then control the water supply with moisture feedback to hit that target. – Matt Timmermans Jun 18 '18 at 18:36
  • I would use 2 PI(D) regulators one for temperature one for moisture. Because even if you manage to create a formula to PID from 2 inputs you still have one output and what this output will control? Heater or moisturiser? Whichever you chose you never reach ideal conditions on 2 things with one output. – Sergey Romanov Jun 19 '18 at 11:06
  • @MattTimmermans The ideal moisture target will be zero (or minimal) all the time. Ideal temperature will also be approx. 100 deg. We can allow a small amount of moisture to make sure we hit the temperature target. – Beutler Jun 19 '18 at 12:51
  • @SergeyRomanov In this case, we are not controlling a heater, ONLY the amount of water, which obviously affects temperature and potentially moisture. The product comes out at approx. 250 degrees F, and we have a small span to decrease that temperature to 100 deg F. See previous comment, if that helps clarify. Thanks for the responses. – Beutler Jun 19 '18 at 12:52
  • I can't get it anyway. If you regulate moisture according to temperature, as on your chart, I would simple youse one input to PID but I would create a small function that change set moisture value for PID according to temperature. What confuses me is IDEAL condition on your chart. How you gonna get it if you do not control temperature? If you would say that blue line is ideal moisture relative to temperature that would make sense. – Sergey Romanov Jun 20 '18 at 11:40

1 Answers1

1

I have done what is referred to as Cascade loops several times. Sounds like this might work for your application. Basically, one PID's (orPI) Control Variable (Output) supplies the next PID's Set Point (Target). I use mostly Rockwell controllers, from your description, use the temperature as your first loop's (Master Loop) Process Variable, tune this loop so you can maintain Target Temperature. Now use the control variable from the Temperature PID as the SetPoint or Target for the second PID (Moisture). The Moisture PID's Control Variable will be whatever device (Pump drive, Control Valve) is controlling the Flow of water. Then comes the Tuning.

Cascade Loop

PaulMc
  • 109
  • 11