Statistical error analysis for the direct simulation Monte Carlo technique (Q1924706)

The ways of analyzing and the minimization of statistical errors of the direct simulation Monte Carlo method are studied. The dependence of the statistical error on the execution time, the particle number per cell, and the memory usage is discussed. Simulations for three kinds of flow (low Mach number supersonic flow, hypersonic flow, and expansion nozzle flow) under rarefied conditions are performed. Plots are shown representing the dependence of the root mean square error on particle number per cell and translational temperature. The number of particles in the simulation must be greater than a certain minimum. Recommendations are given for the maximum sampling time steps.