1. I tried to break the equation into muliple parts for z. Since my z range is -50 to 700. These are parts I have gets where 6th order polynomial fits.
-50 to 61
61 to 192
192 to 278
278 to 346
346 to 401
401 to 446
446 to 477
477 to 599
599 to 668
668 to 700
2. "test RI_2Dpolynomial-set2_break.vi" is used where small sets of data is taken to see in which regions it shows polynonimal fit. One at time. First I tried on dataset of -50 to 61 ,then ran VI again for next data set and so on. Max error allowed is +-0.0999.
3. Then I ran test on all available data-points for -50 to 61 range and saw the results. It shows a poly fit of 6th order. Need to check for rest of changes though.
4. Still is there any better method which you can suggest? It may not poly fit always, may be any other non-linear fit ?
Yes, x and y (2nd and 3rd column) also form a reasonable grid and can be fit. However the fit does not look that great and might need higher terms, at least in one of the dimension..Of course a 2D polynomial is just a generic model and you could of course implement any other theory and use nonlinear fitting instead.