Re: [foxboro] PIDE & PIDA
- From: "Cossitt, Howard" <howard.cossitt@xxxxxxxxxxxxxxxx>
- To: foxboro@xxxxxxxxxxxxx
- Date: Thu, 30 Mar 2006 12:29:26 -0500
Excellent summary!
Please allow me to add the following info:
1. One of the reasons that "neither the PIDE nor the PIDA can be trusted in
the self tune mode without keeping an eye on it" is that it is tuning for
robustness not minimal error. Users often expect the PIDA block in
self-tune mode to retune itself continually to "optimize" the control by
minimizing summation of error. This is not the case, the selftune monitors
"peaks", essentially looking for some form of instability therefore
re-tuning to eliminate these peaks and achieve robustness. This means that
as time passes, your control will be robust for all conditions encountered
but will not necessarily be optimal depending on your actual goal.
2. One of the less known features of the FBTUNE block is the blocks ability
to memorize tuning parameters based on zones and direction of measurement
for a given signal that may affect the tuning. Generally this can be the
actual final control element position. As the element crosses these zones
in a specific direction, the controller changes the tuning parameters to the
memorized values. No calc block needed...
Howard C. Cossitt
National Training Coordinator, Lifetime Learning Center
Invensys Systems Canada
Phone: +1-514-421-8095
Fax: +1-514-421-8059
Email: howard.cossitt@xxxxxxxxxxxxxxxx
-----Original Message-----
From: foxboro-bounce@xxxxxxxxxxxxx [mailto:foxboro-bounce@xxxxxxxxxxxxx] On
Behalf Of Rick Rys
Sent: Wednesday, March 29, 2006 4:44 PM
To: foxboro@xxxxxxxxxxxxx
Subject: Re: [foxboro] PIDE & PIDA
Nurul,
PIDA was developed after some operating experience with the PIDE and is
generally superior. The PIDE is self contained but the PIDA uses the =
FBTUNE
block. Both have a Pretune and Self Tune mode. The self tune for PIDE =
is
using the BMP parameter with a step upset and waits for WMAX time for a
response. The PIDA also has a BMP parameter (actually in the FBTUNE) =
but
uses a bi-polar pulse (better) and only uses the THRESH (noise =
threshold).
You must have enough BMP to exceed the THRESH before the bi-polar pulse =
will
invert. The Self-Tune mode in the E or the A is a unique feature that to =
my
knowledge no other vendor provides. It tunes with the controller in
automatic by detecting peaks. No peaks =3D no tuning changes. It is =
not
uncommon for the controller to tune itself to sleep, but both PIDE and =
PIDA
can tune themselves back from sluggish when an upset comes along. The =
PIDA
is much better but not perfect at sorting out upsets entering the loop =
at
the natural frequency of the loop. PIDE will futilely attempt to tune =
such
upsets out. PIDE criteria is set with damping and overshoot as the =
tuning
objective. PIDA has no settable objective but typically results in a =
nearly
critically damped response. =20
Both algorithms were designed for load upsets rather than so called =
"servo"
or setpoint response. PIDA also has more mode options and the self =
tuning
function is adapted to each including the PIDTAU. The PIDA tuner also
computes the SPLLAG and the KD derivative filter time with the Pretune. =
The
PIDA can use an algebraic method or a fuzzy logic method to determine =
the
tuning. Don't ask me to explain the difference, but the default =
algebraic
method is generally successful. As I recall the algebraic method does
create an internal transfer function process model that is used in the
computation of the tuning parameters.=20
Generally, I like to use the PIDA Pretune function for most tuning. =
Pretune
usually finds excellent tuning settings in one attempt. It tunes =
shortly
after completing it's bi-polar pulse, although sometimes it will fail to
tune and it is not clear why. You can use a Pretune request of 2 to help
force it to tune, but it can still fail to tune in some cases. The =
Pretune
will also compute the PR_TYP (process type) parameter automatically. In =
my
opinion neither the PIDE nor the PIDA can be trusted in the self tune =
mode
without keeping an eye on it. You can however restrict the range of
adjustment on the tuning parameters. PIDA will detect if it really =
screwed
up and will turn itself off and use the fallback tuning settings. In
Self-tune you can also see it identifying peaks/quiet/settling etc in =
the
block displays.
You can make a simple simulation with PIDA, FBTUNE, DTIME, and LLAG =
blocks
and get a little practice before using it in the plant.
Best regards
Rick Rys P.E.
www.R2Controls.com
508-369-5186 Cell
302-836-6540 Delaware Office Days
R2 E-Mail rys@xxxxxxxxxxxxxx
Valero E-Mail: Rick.Rys@xxxxxxxxxx
=20
_______________________________________________________________________
This mailing list is neither sponsored nor endorsed by Invensys Process
Systems (formerly The Foxboro Company). Use the info you obtain here at
your own risks. Read http://www.thecassandraproject.org/disclaimer.html
foxboro mailing list: //www.freelists.org/list/foxboro
to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave
_______________________________________________________________________
This mailing list is neither sponsored nor endorsed by Invensys Process
Systems (formerly The Foxboro Company). Use the info you obtain here at
your own risks. Read http://www.thecassandraproject.org/disclaimer.html
foxboro mailing list: //www.freelists.org/list/foxboro
to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave
Other related posts:
