MATLAB: System identification procedure

identificationSimulink Control DesignSystem Identification Toolbox

Hello, anyone knows? How to make implementation or realization system identification procedure in matlab .Using simulink model(electric device model such as transfer function and integrator).From this model I can get experimental data(input and output).

Best Answer

Have Simulink data available in MATLAB workspace as double vectors. In the GUI (ident) import time signals to load the data. Then you work from there. If you are estimating models in the command line, package the data into an iddata object:

 data = iddata(y,u,Ts); % y = output; u = input; Ts = sample time

Then try various estimation commands, e.g:

 model = tfest(data, 2); % second order transfer function

The logged data (y,u) must be equi-sampled. If not, interpolate them to a uniform time grid.

Related Solutions

MATLAB: How to set a limit to the coefficient of the estimated transfer function in the System Identification app

The GUI for transfer function estimation does not support parameter constraints, but it's possible in command-line, without using grey-box modeling.

More details are at

https://www.mathworks.com/help/ident/ug/imposing-constraints-on-model-parameter-values.html

For example, given the data

u = randn(200,1);
y = randn(200,1);
ts = 0.1;
z = iddata(y,u,ts);
% Generate an initial model, perhaps by unconstrained initialization, prior knowledge, or random system. 
Gtf = tfest(z,2,0); % unconstrained initialization
Gtf.Structure.Numerator.Minimum = 0; % Constrain the numerator coefficient to be positive 
Gtf.Structure.Numerator.Value = 1e-4; % Give a + numerator guess
Gtf = tfest(z,Gtf); % Refine the initial model estimate

MATLAB: How to show the power spectral density of a time-domain arx model in System Identification Toolbox

For frequency response of the model (H(w) = B(w)/A(w)), use FREQRESP or BODE command, as in: H = freqresp(m, w); or [Magnitude, Phase, w] = bode(m);

For the disturbance spectra (PHI(w) = L*||A(w)||^2), first extract the noise model and then call the spectra computing command:

In releases R2011b and earlier, do mn = m(:,'noise'); DS = freqresp(mn)
In releases R2012a and later, do: mn = noise2meas(m); DS = spectrum(mn)

Frequency response for a nonlinear model is not well-defined. You can linearize the model about a given operating point (see idnlarx/linearize) and call BODE/FREQRESP on the resulting linear model.

Or, you can construct the response of the nonlinear model over a grid of frequencies by exciting the system with one sinusoid at a time. This is more easily done in Simulink if you import the identified model using the Nonlinear ARX block. Then, use FRESTIMATE (requires Simulink Control Design).

Or, if you are able to write the governing equations for the model, you can do a (painstaking) describing function/harmonic balance analysis by hand to compute the "frequency response".

Best Answer

Related Solutions

MATLAB: How to set a limit to the coefficient of the estimated transfer function in the System Identification app

MATLAB: How to show the power spectral density of a time-domain arx model in System Identification Toolbox

Related Question