Nonlinear curve fitting using labview

[hariharan1992]

Hello all,

I want to create a VI for nonlinear curve fitting for a plot with datas of x and y values.

Function to be use for fit is attached. In that i want to fit the plot by only controlling the value of beta (using dial knobs in control panel). other parameters are constant. 

can anyone please give me idea to start.

Thanks in advance

[Bob_Schor]

If you want to hire someone with LabVIEW expertise to do your work for you, you are in the wrong Forum (see LabVIEW Developers Seeking Employment.  If you want help for something you are trying to do yourself, attach the VI that is giving you problems and ask for specific help.

If you are not sure where to look for Non-linear Curve Fitting functions, they are listed in the Mathematics Sub-Palette under "Fitting".  One, in fact, is called "Non-linear Curve Fit.vi" -- this might be a good place to start.

Bob Schor

[hariharan1992]

Thanks for your reply.

I will try and contact you further for clarifications.

[altenbach]

Your formula "picture" is completely useless, because it contains neither x nor y and it does not say what symbols are constants and what are the fitting parameters.

• What are the independent variable(s)?
• What is the dependent variable?
• What are the fitting parameters and what would be reasonable guesses for their values?
• What are the values of the constants. Some are recognizable (e.g. pi), but others are ambiguous.

What is the theory behind it all? Is there a web site that explain the theory and experiment?

This is definitely not a nice function (square roots, logarithms, divisions, integrals) so you might need to protect from negative or zero terms and divergent integrals, etc.

See how far you get.... 😄

[hariharan1992]

Dear altenbach,

Thank you for your reply.

This is the theory for z-scan experiment (non-linear optics).

here, beta is the fitting parameter. It may be in the order of e-12.

lambda (532 e-9), omega0 (1e-6), E (20e-6), A , t (10e-9), l (1e-3) are constants.

I'm new to labview. please suggest me how to start with this and to fit the experiment data (z vs T) using the above mentioned equations. 

[altenbach]

Do you have a set of real data?

[hariharan1992]

Yes sir, I have experimental data of x and y axis values ( z vs T)

[altenbach]

Well, I was wondering if you could attach some real data so we get an idea how it looks.

In any case, all you need to do is place a "nonlinear curve fit" on the diagram, create a model VI according to the template mentioned in the help so it calculates an array of y as a function of an array of x, given the parameters. Then feed it your data and rough parameter estimates and it will return the parameter set that best describes the data and the function calculated from it.

Have a look here for some examples.

[hariharan1992]

Thank you sir for your suggestions

As you asked and for your reference, the experimental data to be fit is attached herewith. I have to fit the obtained experiment data. Please see the attachment.

[altenbach]

That's just a picture of data and pretty useless for us. Don't you have a data file?