The current paper investigates new collision-pair selection procedures in Direct Simulation Monte Carlo (DSMC) method. Collision partner selection based on the random procedure from near neighbor particles and deterministic selection of nearest neighbor particle have already been introduced as schemes that provide accurate results in a wide range of problems. In the current research, new collision-pair selections based on the time spacing and direction of relative movement of particles are introduced and evaluated. Comparisons between the new and existing algorithms are made considering appropriate test cases including fluctuations in homogeneous gas, 2D equilibrium flow, and Fourier flow problem. Fluctuations in different correlational parameters and the values for collisional parameters are reported. Distribution functions for numbers of particles and collisions in cell, velocity components, and collisional parameters (collision separation, time spacing, relative velocity, and the angle between relative movements of particles) are investigated and compared with existing analytical relations for each model. The capability of each model in the prediction of heat flux in the Fourier problem at different cell numbers, numbers of particles, and time steps is examined. For new and existing collision-pair selection schemes, the effect of an alternative formula for number of collision-pair selections and avoiding from repetitive collisions are investigated in the prediction of the Fourier heat flux. The simulation results demonstrated the advantages and weaknesses of each model in different test cases.

Keywords