Auto tunning (air flow control)

[Ric]

I have an electronic pressure regulator (EPR) and a flowmeter and what I'm trying to accomplish is to have a user input a flow SP and the EPR open or close to achieve desired flowrate. Fluid is air with temp range between ( -80C and 22C), EPR  input is 0-10v and the flowmeter outputs a 4-20mA output (50-500 l/min). I clamped the EPR to only output a max of 20 psi otherwise I'm outside the measuring range of the flowmeter. Therefore, at 10v I get 500l/min. Because the user can enter an input above current setpoint or below I figured I needed two PID loops, one indirect acting and one direct acting. When I initiate auto tune for the direct acting it just sits there with a status 2. The indirect acting goes through it's 4 stages but the results produce a very unstable process (large oscillations, using full output range). the Ki values is 1000, Ki and Kd =2.  The setpoint when I initiate the auto tune is 225 LPM  but I have no flow when I initiate the auto tune since CV is 0 and there is no output to the EPR until I start the auto tune. Is this the correct approach?  Is this a process that needs to be manually tuned with one loop or is the two loop strategy, correct?  Any help is appreciated.

Thanks,

Ric

[Ausman]

Does this query include using a Vision PLC?  If so, please tell us the model in use and other relevant info. 

[Ric]

Samba 3.5 SM35-J-RA22

[Ausman]

Is there anything variable after the pressure regulator?  And how far apart are they?  If the answers are no, and fairly close, then there is essentially a direct relationship between the reg and the flow rate.  In this case I think that PID is not needed and you're over-thinking it. You could "roll your own" closed loop system that varies a rate of change to the reg according to how far from the target it is, and thus lets it settle at the perfect amount quite quickly, and then keeps it tweaked through the minute adjustments because it's in the "tiny adjustment" range.  Your optimal amount of adjustment rate needs to be finalised through trial and error, so it is best done with MIs as the basis for initial rate numbers, which you then make the power up values.  You can also use linearisations for this rate, but sometimes fixed amounts for how far apart things are works better.

PID works well for some things, but I have often had better life of equipment through doing what I suggested.  ie I've had PID wear control valves out in far less time than doing it my way, with no measurable difference in performance...in fact, mine was better and far more stable.  PID enthusiasts will no doubt chime in with different opinions.... 🙂

[Ric]

Ausman, the more I think about this I tend to agree with you (I am overthinking this). I'm going to try the closed loop approach because I'm not getting anywhere with the PID control.  To answer your question, there is just tubing connecting the output of the EPR to the flowmeter and the distance is only 2-3 feet.  Right now, I'm just bench testing but the system distances will be the same. Thanks for the input.

 

Ric