1. Home
  2. ⧑ Industrial Control Group
  3. System Identification with Slicot, problem

System Identification with Slicot, problem

I have a MIMO thermal system where I am doing System Id. via. Scilab using the Slicot subprograms. I found that if I don't offset the temperature results to a reasonable approximation of zero the id is completely done in. I tried adding an auxillary input that had a constant drive value, hoping that the id would pick it up as the extra input causing the output offset; that attempt failed. My question is; is it logically necessary for me manually preprocess the output vector to eliminate static DC components?

Ray

Reply to
RRogers
Loading thread data ...

I don't have experience with the Scilab id programs. My experience is with LSQ identification and it was absolutely necessary to cancel any DC components on the input and output to get good results.

Ronny

Reply to
Ronny Rooman

Ronny, Thanks for the response. I got in touch with a person at Slicot and his conclusion was that it would be neccesary. Personally I think that it is not logical. In fact I added another static input (column of 1's) hoping the algorithm would pick it up and expand the D matrix. Didn't work; I am sure it's a failure internally but I don't have time to fix it now. I might dig in later. Aside from that the SciLab/Slicot process worked very nicely. If you want to try it out let me know and I will send a simple interface script; and data if you want. I don't watch this newsgroup too regularly so here is my email; remove the - . r-e-r-o-g-e-r-s@p-l-a-i-d-h-e-r-o-n-.-c-o-m

Ray

R> I don't have experience with the Scilab id programs. My experience is with

Reply to
RRogers

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.