A question about Monte Carlo model

Forrest

Hi all,

I am using CFX5.7.1 to model convection and radiation in an open ended tall cavity. One side of the cavity wall has a high temperature, and all the others are adiabatic. They are all opaque walls with emissivity of 0.9.

After the simulation I plotted the temperature profile along the central line of the wall which is in front of the heated wall. The profile looks sawtoothed rather than smooth. I tried to increase the 'number of history' from 1,000,000 to 2,000,000, it helped a bit, the swing is reduced a bit, but still far from smooth; also tried to reduce the 'coarsening rate', but it didn't help on this.

Can anyone give me any clue for this?

Thanks a lot, Forrest

p.s.: All other RMS residuals were managed down to 10^-5 except the RMS H-Energy, it was settled about 5x10^-4 and not going any further no matter what timesteps used. One thing I noticed was after reduced the 'coarsening rate' the RMS H-energy is slightly increased. I used Discrete Transfer radiation model, the corresponding temperature profile is smooth and all the residuals were down to 10^-6.