home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.engr.control
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!eagle!mars.lerc.nasa.gov!mattern
- From: mattern@mars.lerc.nasa.gov (Duane Mattern (Sverdrup))
- Subject: Re: Simulink problem!
- Message-ID: <17NOV199216571543@mars.lerc.nasa.gov>
- News-Software: VAX/VMS VNEWS 1.41
- Sender: news@eagle.lerc.nasa.gov
- Nntp-Posting-Host: mars.lerc.nasa.gov
- Organization: NASA Lewis Research Center
- References: <1e9hpgINNc8j@iskut.ucs.ubc.ca> <1992Nov17.183057.18351@den.mmc.com>
- Distribution: usa
- Date: 17 Nov 1992 16:57 EDT
- Lines: 34
-
- >In article <1e9hpgINNc8j@iskut.ucs.ubc.ca>,
- >lug@balsam.ppc.ubc.ca (Guoqiano Lu) writes:
- >I am using Simulink to simulate continuous-time self-tuning algorithms.
- >Generally it works well except sometimes there are some crazy peaks in
- >the output which is supposed to be smooth. It seems to me that these are
- >due to the numerical intrgration algorithms that SimuLink used. Any one
- >has the same experiences or any suggestions to overcome this problem?
-
- Numerical integration schemes can cause funny problems, particularly when
- used with something like a self-tuning algorithm. Variable step integration
- schemes in particular tend to "hunt" in an attempt to converge, particularly
- in simulation that include random noise. But typically they will only run
- slowly and not create "peaks". To check to see if it is the integration
- scheme and not the self-tuning system, switch to a fixed time step
- integration scheme (like Euler), with a small time step. (I hope that
- you can switch integration schemes in Simulink. You can do this in ISI's
- SystemBuild, of which I am familiar). If this clears up the problem,
- then you know its the integration scheme and you can deal with that issue
- by changing your integration technique.
-
- If you still have the "crazy peaks", all that I can suggest (with the
- given information) is to check to make sure that the self-tuning algorithm
- is "sufficiently excited". Astrom had published some papers in the early
- 1980's about the "bursting" phenomen where a self-tuning or adaptive
- scheme can blow up the varying gains if not receiving the proper system
- excitation. I've experienced this with self-tuning simulation in the
- past. Typically this might happen in a simulation during steady
- state regulation (non-transient), when the characteristics of the
- system build identified or being tuned to, are not available in the
- measurements.
-
- Duane Mattern (mattern@mars.lerc.nasa.gov) (216)433-8186
- Sverdrup @ NASA Lewis Research Center, M/S 77-1, 21000 Brookpark Road
- Cleveland, OH 44135
-