[buckngnr11]

Probably a dumb question. I'm using a double precision solver, and using a UDF to make various calculations. When I display the output of the calculations in the console I notice that some of the variables appear to have a floating point error (values are on order E+200). The variable being calculated here involves no division, just multiplying two very small numbers (1e-8 and another of order 1e-9). I suspect that I'm having a precision error. I'm assigning all variables to be type real. Not sure what my options are to remedy this other than hardcoding any large numbers to zero. I'd rather have a more elegant solution since I'm not entirely positive this is necessarily the issue.

[vinerm]

As far as double precision is concerned, it is machine dependent. However, 1e-8 or 1e-9 is quite large as compared to machine precision if you are using double precision. So, these should be treated properly as finite numbers without any issue. However, it matters for what parameter in Fluent are you using these numbers. E.g., there are certain limits set in Fluent, which could be changed by user, if required, that do not allow numbers to be smaller than a certain limit. If you are using single precision, then certainly these numbers will be treated as 0. Also note that in UDF, use 1.e-9 and not 1e-9.

[AlexanderZ]

how do you display your variables?