Issue2884 pid autotuning #3609
Issue2884 pid autotuning #3609
Conversation
| then <code>y = yHig</code>, <code>yOn = true</code>, | ||
| </li> | ||
| <li> | ||
| else if <code>yErr > deaBan</code> and <code>trigger</code> is <code>true</code>, | ||
| else if <code>yDif > deaBan</code> and <code>trigger</code> is <code>true</code>, |
There was a problem hiding this comment.
Note that in current implementation, if yDiff (the input to instance hys) > deaBan, then, hys.y = true, yOn=true and y=yHig. This is opposite to the documentation here.
Which one is right?
There was a problem hiding this comment.
Sorry for confusion and actually I confused myself with the signs. I have updated the model doc.
| @@ -50,23 +50,23 @@ protected | |||
| "Lower value for the output" | |||
| annotation (Placement(transformation(extent={{-80,20},{-60,40}}))); | |||
| Buildings.Controls.OBC.CDL.Reals.Subtract conErr | |||
| "Control error (set point - measurement)" | |||
| "Control error (measurement - set point)" | |||
There was a problem hiding this comment.
In CDL.Reals.PID and other PID controller variants, the control error is (set point - measurement). See here.
I see that the output yErr from this block becomes input to the block PIDWithAutotuning.SystemIdentification.FirstOrderTimedelayed.ControlProcessModel, so for this block, is it expecting set point - measurement or measurement - setpoint?
There was a problem hiding this comment.
I removed the yErr and define a new variable called yDiff. yDiff = yErr if reverseActing is true; otherwise yDiff = -yErr.
@JayHuLBL Please take a look at the PR based on your comments.