1. Home
  2. ⧑ Industrial Control Group
  3. Observers types and differences

Observers types and differences

Hello,

I'm just starting in observers field and I'm a beginner.

I've to study model-based observers, closed loop observer and adaptive obsever for my phd work. Where can I find a clear explanation of their structures and differences between them? In particular model-based observers, aren't they the "classical" observers (lumberger - with a copy of the system)?

Thank you very much for your help... excuse me for my english.

MIMMO

Reply to
Mimmo
Loading thread data ...

I got most of my information from "Digital Control Systems Analysis and Design" by Charles L Philips and H. Troy Nagle. It covers full, reduced and closed loop observers.

Peter Nachtwey

Reply to
pnachtwey

Thank you very much, I'll go immediately to look for in university library.

Reply to
Mimmo

"Mimmo" schrieb im Newsbeitrag news: snipped-for-privacy@e9g2000prf.googlegroups.com...

A observer is a math models for a process (= process transfer function). A Luenberger observer is a 'selfcorrecting' observer. The real process corrects exclusively the math model of the process.

On my website process errors are PID feedback controlled [1]. The math model is inclusively loop-corrected/controlled.

The PID feedback control is also necessary using a Luenberger observer.

So far I can't see an advantage using a Luenberger observer.

[1] Built-in State Observer:
formatting link
See also State Observer:
formatting link
Nonlinearities can be corrected by adaptation of coefficients c (e.g. c1*A1, c2*A2, ...).

Example:

Approximation of 3 process transfer functions on 25%, 50% and 75% w. Then you have c = f(w).

Reply to
JCH

Newsbeitragnews: snipped-for-privacy@e9g2000prf.googlegroups.com...

??? I don't see the letters PID on the wikipedia web page. The observer will work no matter what the source of the control signal is.

This is because:

  1. Your instructors failed to tell you why it is important. Maybe they don't know.
  2. You do all simulations assuming the feedback is perfect instead of quantized.

Try using one of your PID 2D controllers when the feedback is realistically quantized. The control output will be jumping all over the place and will be unusable. So far we have got a free pass for using perfect numbers for feedback. Some one should have called us on this point in during all the thread over the summer.

Peter Nachtwey

Reply to
pnachtwey

Sorry, I won't mess-up the discussion, and think that what I've said before is ok. Maybe you find better definitions.

Reply to
JCH

"JCH" schrieb im Newsbeitrag news:4721ad66$0$4371$ snipped-for-privacy@newsspool4.arcor-online.net...

Observer Comparisons:

formatting link
Simulating Luenberger with example data
formatting link
't show differences.

v1 error ~ 0

Reply to
JCH

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.