System Identification

Hey!

I am designing a control system for a bearing system.

Here I have 3 inputs and one output. Can anybody tell me about the system identification of such system.

Regards, mac.

Answer 1

[Rajiv Singh]

Try estimation functions such as OE, ARX, N4SID and PEM in releases R2011b and older. In R2012a you can also try TFEST, SSEST, PROCEST and POLYEST.

http://www.mathworks.in/help/toolbox/ident/ref/tfest.html http://www.mathworks.in/help/toolbox/ident/ref/oe.html

etc...

You will have to indicate what you tried if you want more concrete help.

[Markus]

Hi Rajv! Im using version R2010b. i don't understand the parameters of nb nf and nk! Now i have 9 Inputs and 1 Output. Can u give me an example how to estimate this case with 9 Inputs? Have you get my PM with the problem?

[Rajiv Singh]

 y(t) = B/F u(t-nk)

Number of coefficients of B is nb and number of coefficients of F minus 1 is nf. If you have 9 inputs nb, nf and nk are row vectors of length 9, one entry for each transfer function.

Example:

 nf = 2*ones(1,9); nb = nf; nk = ones(1,9); 
 data = iddata(output, input_matrix_with_9_columns, Ts); % Ts = sample time
 model = oe(data, [nb nf nk])

Answer 2

[Markus]

i have a similar problem. I have no clue. Do you have any answer?