[Getdp] Heat of fusion

[David Felsenthal]

I am doing a non-linear thermal model. The heat of fusion is of course a
large non-linearity. When doing a dynamic model I can do a crude
approximation with a jump in the sensible heat over a small temperature
range. It's kind of ugly though. To get reasonable results I need to 
have the rho*c vary by about 300:1 over a range of 1 degree for a 
problem whose range is 20 degrees.

I can make it work but I have convergence problems due (I think) to the 
large variations in the value of rhoc over temperature.

Is there a better way of dealing with this issue, or should I just 
accept long computation times?

Thanks again,
David