Ziegler Nichols 2nd Method - Isn't it for Series Algorithm? (PID library uses ISA form)

[edgsc]

Hi Guys,

Isn't Ziegler-Nichols 2nd Method (Frequency Response Method or Oscillation Method or Ku/Pu Method etc) for Series PID Algorithm?

From https://www.pidtuning.net/pid-algorithm.php

From Autotuning of PID Controllers A Relay Feedback Approach_, 2nd edition, Cheng-Ching Yu (2007), page 30?

Bret's PID library uses ISA (Ideal) PID Algorithm, right?

So, shouldn't we transform Kp, Ki, Kd from Series to Parallel form?

from http://www.acsysteme.com/en/serial-or-parallel-pid

Or from ISA slides:

from ISA Effective Use of PID Controllers 3-7-2013, https://pt.slideshare.net/sarodp/isa-effective-use-of-pid-controllers-372013
@br3ttb

Thanks!

[br3ttb]

interesting! the library is a straight-up port of another library so it's possible that there's a mistake there that I wasn't aware of. The pid library uses parallel form that's listed on your table.

…

On Fri, Apr 12, 2019, 6:15 PM edgsc ***@***.***> wrote: Hi Guys, Isn't Ziegler-Nichols 2nd Method (Frequency Response Method or Oscillation Method or Ku/Pu Method etc) for *Series* PID Algorithm? [image: image] <https://user-images.githubusercontent.com/32095476/56068791-82878180-5d56-11e9-9b34-c32f38e3e806.png> [image: image] <https://user-images.githubusercontent.com/32095476/56068931-18bba780-5d57-11e9-99b5-5355df2bc5cf.png> From *Autotuning of PID Controllers* A Relay Feedback Approach_, 2nd edition, Cheng-Ching Yu (2007), page 30? Thanks! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#20>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAWL8lzWYsATnwGL7TO4GgmMewiTQuOHks5vgQWJgaJpZM4ctif5> .

[edgsc]

interesting! the library is a straight-up port of another library so it's possible that there's a mistake there that I wasn't aware of. The pid library uses parallel form that's listed on your table.
…
On Fri, Apr 12, 2019, 6:15 PM edgsc @.***> wrote: Hi Guys, Isn't Ziegler-Nichols 2nd Method (Frequency Response Method or Oscillation Method or Ku/Pu Method etc) for Series PID Algorithm? [image: image] https://user-images.githubusercontent.com/32095476/56068791-82878180-5d56-11e9-9b34-c32f38e3e806.png [image: image] https://user-images.githubusercontent.com/32095476/56068931-18bba780-5d57-11e9-99b5-5355df2bc5cf.png From Autotuning of PID Controllers A Relay Feedback Approach_, 2nd edition, Cheng-Ching Yu (2007), page 30? Thanks! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#20>, or mute the thread https://github.com/notifications/unsubscribe-auth/AAWL8lzWYsATnwGL7TO4GgmMewiTQuOHks5vgQWJgaJpZM4ctif5 .

Thanks for your reply, Bret!

Just as an example, here's what it looks like after conversion from Series Kp-Ki-Kd to ISA Ideal (or ISA Standard Form):

Ku: 155,839
Pu: 269,617
Series:
Kp: 91,67
Ki: 0,68
Kd: 3085,77

ISA Ideal:
kp: 114,56
Ki: 0,68
Kd: 3085,77

Only Kp changed, about +24%; making the parameters a bit more aggressive.

[xlla]

hi @edgsc, I have work out a set of parameters using autoTune lib for dc motor.
this is original PID and response curve.

this is Kp scale up to 1.24

its aggressive and take a little more time to convergence.