Problem with neoclassical conductivity when Ip=0

[hoppe93]

A problem arises with the neoclassical conductivity when starting DREAM with Ip = 0. In RunawayFluid::evaluateNeoclassicalConductivityCorrection(), the mu0Ip parameter is then 0, which causes qR0, and subsequently nuEStar, to become infinite. When X is then evaluated, a floating-point error is raised (in debug mode) which causes the code to crash.

The appropriate behaviour would probably be to throw an error if trying to run the code with zero Ip, as this should be incompatible with the equilibrium? (Not even STREAM can start from zero Ip, but is usually started from a small non-zero plasma current)

[hoppe93]

Apparently the floating-point error I encountered was caused by something else, in having Ip=0 is not actually a problem.