Hey Guys,

I still need your help. Anyone have an idea?

Best regards!

Hello Viatorus,

I can think of two approaches:

1) use the white noise source and filter it. In Modelica.Blocks.Continuous.Filter several filters, including bandpass are defined. This would give you the noise with the correct spectrum

2) define a PSD (power spectral density) in the Noise toolbox. This is more involved, but you have a better control on what you do (so you have to understand it better too.)  If you look in Noise.Example.ComparePSD, you find how you can redeclare the PSD. You need the PSD_Interpolation. For this, you have to determine the convolution terms yourself.(google: bandpass filter coefficients).

Hope this helps
Bas

Thank you for your answer.
I just used an PT1 element for band-limited noise.

But the current problem is, that I don´t know how to set the standard deviation and power spectral density?

Hello Viatorus,

With the filter you just have set the spectrum. Changing the standard deviation can be done by multiplying the output signal. In the Noise library there is an example on how to calculate the std.

Hope it helps
Bas

Thank very much you for your help!

With the filter I got my standard variance if I have the right spectral density:


My RNG:
Noise.PRNG MyNoise(redeclare function PSD = Noise.PSD.PSD_WhiteNoise, useSampleBasedMethods = true, redeclare function PDF = Noise.PDF.PDF_Uniform(interval = {-1, 1}, samplePerioad= 0.001)

So interval tell me the upper and lower noise value, but what how can I set a power spectral density like 0.5 N^2 / Hz?

So what I need is this in OpenModelica:

http://de.mathworks.com/help/simulink/s … noise.html