double precision and residuals

Mark Austin

Hi,

I am running a simulation of a high temperature channel flow over a solid with the following characteristics:

+ Re = 37 + Energy Equation (without radiation) - heat generation in the solid and a convective boundary condition on the underside of the solid volume. + Species sources and sinks in the channel (CO2, CO, H2O, H2) + Temperature dependent gas properties like thermal conductivity, specific heat, etc. + The simulation is run using the double precision solver.

Everything seems to work well until the residuals seem to take an unexpected jump. With regards to residual asymptotes, the simulation looks like it is probably 75% complete, but then, these residual values spike and hold at levels that I believe were common for the same simulation run as a single precision.

Any ideas would help.

Thanks,

Mark