NX Nastran random analysis for response dynamics SOL 103


Dear NX Nastran Community,


I am running random analysis via response dynamics SOL 103. I tried to extract the RMS-Value which is displayed for each plot in the information as shown in the picture below.



I tried to verify the RMS value with another tool, however, the values are different. Does any of you know how Nastran is exactly calculating the RMS Value? And how NX is interpolating the random PSD input curves? Because the error might be due to a different interpolation method as well.

I used these formulas https://femci.gsfc.nasa.gov/random/randomgrms.html to calculate the RMS-value.


Best regards,



Re: NX Nastran random analysis for response dynamics SOL 103

Siemens Genius Siemens Genius
Siemens Genius

When you create your Random Event for Response Dynamics, you specify if you want to use Even or Uneven functions for the excitations. If you use an Even Event, evaluations are only performed at the frequency points specified in your ecitation function.  If you use UnEven, you also specify the additional spectral lines desired, and the interpolation method to be used to add the additional spectral lines. So, the results points will be at any frequency specified in the excitation function, along with frequency points at the normal mode frequencies. Additiona Spectral lines will be added between these points using the interpolation method you have specified (eg: LogLog,LinearLinear,etc).  During response evaluation, all of these points will be used to generate the response PSD Functions. The RMS is then calculated from the square root integral of the PSD function.

Response Dynamics also uses a Log/Log method when doing the integration of the PSD function to calculate the RMS.

If you are using Nastran, it defaults to using LinearLinear interpolation to determine the RMS from the PSD function. For Nastran, you can change this interpolation by using PARAM,RMSINT,LOGLOG.

So, if you are comparing the Response Dynmamics results against Nastran results, you need to make sure that you are using approximately equivalent excitation points, and that the interpolation method used to obtain the RMS from the PSD function uses the LogLog method of interpolation.

Re: NX Nastran random analysis for response dynamics SOL 103

Thank you for your explanation of the interpolation methods! I only used UnEven until now. I will try to compare the Even and UnEven response with the Nastran results then.