Error using griddedInterpolant. The coordinates of the input points must be finite values; Inf and NaN are not permitted.

I am trying to solve a System of differential equations for a torsional pendulum with euler forward finite difference method. (lines 37 to 51) After every step I need to find an specific value of time (tn) by interpolation with the angular position y(1,:) and the time t (lines 59 to 65). It works most of the time but for some values of final time tf it gives me the following error:

Error using griddedInterpolant. The coordinates of the input points must be finite values; Inf and NaN are not permitted.

I can see that the value PHIt_tn= (y(1,i+1))-(2*pi/N) is resulting to be NaN but I don't understand why or how to fix it. Some values of final time (tf) that give me error are: 59,48,64 and up. I can also see that in these cases the values ydot turn into NaN (really don't know why because all the vectors that operate to create ydot are real numbers).

[Satyajeet Sasmal]

Looks like you might have either/some/all of the conditions given below:

1.Any arithmetic operation on a NaN, such as sqrt(NaN)

2.Addition or subtraction, such as magnitude subtraction of infinities as (+Inf)+(-Inf)

3.Multiplication, such as 0*Inf

4.Division, such as 0/0 and Inf/Inf

5.Remainder, such as rem(x,y) where y is zero or x is infinity

All of the above scenarios can give a NaN.

You could give conditions in your code to test against the above conditions and only proceed with further calculations after being sure that there are no NaNs.

[Walter Roberson]

At the MATLAB command line, give the command

dbstop if naninf

and run your code. It will stop in the debugger as soon as it finds a NaN or infinity being generated in your code. You can then examine the commands and the data to figure out why they are happening. Infinities often quickly lead to NaN.