Is there a way to curve fit a non-linear second order differential equation in DIAdem?

[myra]

Hi Brad,

   Thanks for your help!  I'm having trouble opening the example that you gave me - in Navigator I tried to drag the files into the data portal, but it gives me an error about not having the correct data plugin for it. I didn't see any appropriate data plugin on the download page, but I doubt that is the problem anyway.  Would you walk me through the steps?  (Sorry!)

    I was also concerned about where to find the coefficients in the output.  I also tried making an equation with three independent terms - sine, cosine, and x, and only two of the coefficients seemed to show up... As I mentioned before, I'm still confused about where the coefficients for each term should be found in the output files, and what the 'minimum' and 'maximum' terms correspond to.  For example, I would think that one possible arrangement would be to have an output file corresponding to each term, following the order of the terms as given in the matrix, and min. and max. would give the possible range that the coefficient could take on.  Apparently this is not correct however, since when I put in three terms, there were still only two output files!  I feel like the answers should be really obvious - but so far they're not 😞

Thanks, Myra

[Brad_Turpin]

Hi myra,

All you have to do is open the VBScript file (*.VBS) in DIAdem Script and press the "Run Script" icon near the top right of the SCRIPT screen.  The VBScript will automatically create the data channels, run the calculation, and load the VIEW template (*.TDV) to view the results, including the fitted coefficients.  The General LS Fitting function creates 2 output channels in the Data Portal.  One of these output channels contains N coefficient results for the N channels you added in the Matrix parameter-- in the order you added them to the Matrix parameter.  The other output channel contains an array of generated values using these fitting coefficients so that you can easily plot the resulting fit against the original f(x) function to see visually how close you came.

Brad Turpin

DIAdem Product Support Engineer
National Instruments

[myra]

Hi Brad,

Thanks for the instructions.  I was able to open the example, and I see where I went wrong with finding the coefficients. I should be able to get all of the error information in the covariance matrix, so I believe I should be set now.  Thanks very much for your help!

Myra