Direct Simulation Monte Carlo(DSMC) corner boundary problem

Noh-hoon Lee

Hello, I'm developing the DSMC code of 3d box(for test, no velocity, just applying the thermal boundary condition of 273 K).

If there are diffuse boundary only up and down and others are periodic, simulation result is good.

But if diffuse boundary is at all plane, wrong result has come.

So how it can be fixed? If there anyone know about corner boundary problem of Particle method, I need advice or information where I can get the algorithm or concept of corner boundary condition. Please help