8 Replies Latest reply on Aug 18, 2014 11:11 PM by Jared Conway

    NEBS vibration simulaton

    Randy Palmer

      I am frequently tasked with running earthquake simulations for GR63 CORE (NEBS).  I have obtained the data file that generates all of the g amplitudes, however, when I input the curve into SolidWorks simulation, it truncates it to the first 5000 data points.  I have no idea how to a) get it to read all the data points b) account for this in my results c) make an approximation that would allow me to use 5000 of the 6500+ points in the curve.

       

      Any suggestions?

        • Re: NEBS vibration simulaton
          Shaun Densberger

          Which GR-63 data set are you talking about, the time-history data set or Response Spectrum Curve data set? I'm guessing you're talking about the time-history data set based high number of data points you're talking about, but I just want to confirm.

           

          If you are using the time-history data set, then I'd like to point out that this isn't a realistic way of doing a seismic analysis, and you should instead be doing a Response Spectrum Analysis. The Response Spectrum Curves provided by NEBS have 5 or 6 data points (depending on what zone you're looking at) with a logarithmic interpolation between data points.

           

          Let me know if you'd like more information regarding a Response Spectrum Analysis.

            • Re: NEBS vibration simulaton
              Randy Palmer

              It is the time history data, which is the data that is used as inputs to test equipment.  From my understanding the response spectrum curve is given such that the response spectrum of the test equipment can be compared to the ideal and adjusted using a fourier transform (if necessary) so that the test equipment (e.g. shaker table) output meets the required output.

                • Re: NEBS vibration simulaton
                  Shaun Densberger

                  "From my understanding the response spectrum curve is given such that the response spectrum of the test equipment can be compared to the ideal and adjusted using a fourier transform (if necessary) so that the output meets the required output."

                   

                  True, but that means that the RSC is defining the output of the table (since the transfer function for the table is tuned such that the output of the table matches it), which is the base excitation for your structure. You can therefore use this RSC curve as a base excitation input in a RSA in SolidWorks.

                    • Re: NEBS vibration simulaton
                      Randy Palmer

                      No, it doesn't because the specification indicates that "the shaker table's analyzed acceleration,known as the Test Response Spectrum (TRS), must meet or exceed the Required Response Spectrum (RRS)".  To me that means that it could be somewhat different, however slight.

                        • Re: NEBS vibration simulaton
                          Shaun Densberger

                          Are you referring to bullet point #4 under Shaker Table in Section 5.4.1.2? If so, then my original point is correct. What bullet point #4 means is that the table's response spectrum output must meet or exceed what is shown in Figure 5-18 (and thereby what is shown in Figure 5-17). I diagram might help to illustrate this.

                           

                          Shaker Table System.png

                           

                          In other words, the input to the shaker table and its unique characteristics must result in a response spectrum curve that meets or exceeds Figure 5-18. If the n number of accelerometers responded such that they produce the response spectrum curve in Figure 5-18 (or exceed it), then the table's output meets GR-63's requirements.

                           

                          Since the table must have a RSC that meets or exceeds Figure 5-18, that means you can use Figure 5-18 as in input in a base excitation analysis within SW. There is no reason to do a time domain analysis unless you're interested in some stress of deflection values other than the statistical maximum. For GR-63, it's merely a matter of not exceeding 75 mm of peak deflection at the top, having a fundamental mode greater than 2 Hz, and not have permanent structural damage.

                            • Re: NEBS vibration simulaton
                              Randy Palmer

                              Yes, I understand all that.  I guess I am still unclear what is meant by "exceed".  Does that mean the PSD curve is higher than the desired PSD, perhaps at certain points?  I have done the analysis both ways, and interestingly, get similar results.  But I would still like to know how to get Simulation to recognize all the data points in the time history curve.

                                • Re: NEBS vibration simulaton
                                  Shaun Densberger

                                  "I guess I am still unclear what is meant by "exceed".  Does that mean the PSD curve is higher than the desired PSD, perhaps at certain points?"

                                   

                                  NEBS knows that no one will be able to exactly reproduce Figure 5-18 on a shaker table, so they've added the caveat that the table's output can exceed the required input. If your table's response exceeds Figure 5-18 and the structure still passes, then we know that it would pass if the table were to exactly reproduce Figure 5-18. For example, if I determine that a certain component needs to withstand 100 lbf to survive operational requirements, then in the test procedure I'll specify that the applied testing load must meet or exceed 100 lbf.

                                   

                                  I also want to point out (for completeness) that a RSC and a PSD are not the same thing. A PSD is the power associated with each frequency, whereas a RSC is the response of a structure at each frequency.

                                   

                                  "I have done the analysis both ways, and interestingly, get similar results."

                                   

                                  Make sense. Both curves (Figure 5-17 and Figure 5-18) define the same "event"; they just do it from different "points of view". A RSC defines how a structures responses to a given dynamic input (hence them having damping incorporated into them). You can read more about it here.

                                  • Re: NEBS vibration simulaton
                                    Jared Conway

                                    can't be done i believe. please check the solidworks KB. i think you have to split the curve into multiple segments.