Auto Tune SP

[YINGD]

Per the Visilogic manuals, when Auto Tuning a PID loop, thefollowing values are needed:

· Set Point

· Input Range:PV Low Limit &PV High Limit

· Output Range:CV Low Limit &CV High Limit

What does Auto-Tune use the SP for? Does it tune the wholerange based on the range given for the high and low PV and CV values or does itjust aim for part of the range centered around the SP?

[Emil]

Hi Joe,

Autotune preforms several (1-3 - stated in Autotune module) ON-OFF cycles and then "investigates" over and undershut waves. Based on parameters of these waves, it calculate the basic PID parameters - Pb, Ti, Td, ST and additional parameters in 32 MI vector.

[YINGD]

Hi Joe,

Autotune preforms several (1-3 - stated in Autotune module) ON-OFF cycles and then "investigates" over and undershut waves. Based on parameters of these waves, it calculate the basic PID parameters - Pb, Ti, Td, ST and additional parameters in 32 MI vector.

Thanks for your reply. I'm not sure that I completely understand the process yet though. So when the Auto-Tune runs, does it have a specific time interval that it runs for each cycle or does it run until it finds some consistent values?

Based on your previous statement, it sounds like the Auto tune set point is not necessarily used for the auto tune cycle. Is this true? If not, how does the set point relate to the 1-3 cycles that the auto-tune goes through?

The reason I ask is because when my auto-tune loop runs, it seems to be running inconsistently and I'm trying to narrow down the source. Also, my loop needs to be very accurate so I need to find the very best method to tune.

[Emil]

Hi Joe,

Attached is a picture - capture from PID server (VisiLogic > Toolks menu). It shows Autotune process of an electric kattle at 60C.

Blue line is Setpoint (60.0C); Red line is PV and Green line is CV.

You can see clearly 3 stages. PV = 100.0% (1000). PV rises to SP. Then CV is switched off (0%). PV makes overshut and then fails below 60 degrees. CV is set ot 100% again.

At the end of third cycle PID parameters are set and the system staqrts PID control.

Autotune time depends on the proces. In this case it took about half an hour. In some cases it can take seconds, in oter - hours.

[Aric]

How long will auto-tune attempt to "hit" the high and/or low range?  I had a customer that was trying to run the auto-tune with the system hot.  The auto-tune ran for hours without completing. 

[Kikis]

17 hours ago, Aric said:

How long will auto-tune attempt to "hit" the high and/or low range?  I had a customer that was trying to run the auto-tune with the system hot.  The auto-tune ran for hours without completing. 

If PID is set to reverse mode (heating), Auto-tune works like this:

Sets control output to 100% until target temperature is reached, then resets the output  so that the temperature goes lower than the set point.

This process is repeated several times until the system is finely tuned.

 

Therefore A/T process may take some minutes or even hours to be completed, depending on how the system behaves.

 

 

[Joe Tauser]

On 3/10/2018 at 1:59 AM, Kikis said:

Therefore A/T process may take some minutes or even hours to be completed

+1

If the system has a lot of thermal inertia (won't cool off quickly) you'll be waiting a long while for the AutoTune to Complete.  The number of times it runs the algorithm is set by the Stages parameter.

So it could be hours.  Or days.

I often resort to my Days Of Old manual tuning methods on a system that is overpowered like this.  There's no rule that says AutoTune has to complete before the PID block will run and I don't write my PID code the exact way Unitronics examples show.

  

As you can see, I allow myself to turn AutoTune off and just run the PID control if I get impatient.

 

Joe T.