Integral won't converge to solution

[swahlgren]

I have an integral that won't converge to a solution for a specific set of variables (the solutions over a range of variables are sought). I've attached the whole worksheet and thus a lot in it isn't relevant to the solution so just skip to the bottom the find the error-riddled integral.

Thanks in advance to anyone taking the time to help out.

[AlanStevens]

You need to upload the .txt file input at the start as well or the rest of the file won't calculate.

Alan

[swahlgren]

Opps, sorry. Changed the worksheet so the txt isn't needed.

[MikeArmstrong]

Because you are passing an range variable to the integral. Firstly pass a integer to the integral and change the IF statements to integers.

[swahlgren]

That can't be it because it works just fine except for i=j.

[AlanStevens]

Yes, it's not the range variables that are the problem.

See my reply below.

Alan

[MikeArmstrong]

Yes sorry maybe I was a little hasty with my response.

[AlanStevens]

You have some singularities under the square root sign in the h function.  One possible solution attached (you'll need to decide if this approach is acceptable for your scenario).

Alan

[swahlgren]

I'm not sure that's it cause it works just fine except for i=j. I can't look at you suggestion right now but I'll take a look when I get back to work tomorrow.

[AlanStevens · Accepted solution]

Ok.  It's a numerical sizing problem.  The solution, strange though it may seem, is simply to put the 2/E term inside the integral!!

See attached.

Alan

[swahlgren]

It works! Much obliged. Mathcad sometimes works in mysterious ways.