PID values during autotuning

[Blore]

To start the autotuning process using labview PID tool kit what should be the initial value of P,I and D? And also does the tuning duration required depends on initial P,I and D values?

[Kristi_P]

The initial PID values needed for autotuning must produce a stable system and be fairly close to what is needed for the end result. The time required to complete the autotuning will depend upon how good the initial approximation is.

You might be interested in looking at an example program that ships with the PID toolset called General Auto PID Simulator.vi. You can find this in Help->Find Examples->Toolsets and Modules->PID->General PID Simulator.vi.

If you have any other questions, please post a reply.

Regards,

Kristi H
Applications Engineer
National Instruments

[xtalgrwr]

Krist H:

Could you give me an example of just how long that might take to tune? I ran a stable system and invoked autotune. But ten hours later it still hadn't givem me an ok or next to save the values. Suggestions?

[xtalgrwr]

Hello? Anyone on this thread? Planet Earth?...

[MattP]

Hello,

The autotuning wizard typically only takes a few minutes to complete, so I suspect that there may be a deeper issue occurring here.  It would be helpful to know on what stage of the wizard you experience the hang, as well as whether the chart continues to update during this time.  Additionally, what versions of LabVIEW and the PID Toolkit are you using?

As a good test, try running the General Auto PID Simulator.vi example (available in the Example Finder) and invoking autotune with the default parameters.  How long does this take to complete, and do you see the same hang behavior?

[xtalgrwr]

MattP,

Thank you for that reply.

As always it is best if we first read the book. When running the autotune wizard it opens a window with PV, SP and Output, if I am not mistaken. But the limits of the graph, in my case went from +100 to -100. My temperature measurements are triple that. So of course I could not see the cycling. What happen in that 10 hr run is anybody's guess. Today I was able to tune in a bit better. It does so much faster. It was getting stuck when it was beginning the relay test. Which I hadn't understood until I changed the limits on the graph. Are those the default values?

Thnx.

[MattP]

One thing to keep in mind with the Autotuning Wizard is that by default, it expects SP, PV, and output to be expressed as a percentage, which is why it displays only the +-100 range by default.  Are you still running into difficulties with this VI?  It sounded like you've already made good progress on it, but if I can assist you further please let me know.

[xtalgrwr]

Matt P.

Thank you. Actually I'm ecstatic.

Even though I can still improve the PID tunning at least I am moving forward. The relay test is just a perturbation to help with the tunning. I wasn't sure if it was something else -cutoffs or something to that effect. Once I saw it on the graph I inmediately recognized it. Initially I was using smaller values for the relay test when I should have been using a bit bigger values. e.g. instead of +/- 1 to 2 degress I turned to doing +/- 5 to 10.

Many many thanks.

[MattP]

It's my pleasure!  Good luck with the rest of your application, and feel free to post a new thread if you run into any more questions.