Lunar dust particles are a severe threat to different sensitive surfaces of landers and rovers. The success of future missions require tools for modelling such environments. Technical challenges are linked to accurately model major factors influencing the contamination transport: reconstruction of the plume flow field, regolith erosion due to plume impingement, drag forces acting on particles, electrostatics, particle-surface and particle-particle interactions.

The plume-originated supersonic flow field is in the continuum flow range (Knudsen number Kn<0.01) near the rocket engine bell and later changes to free molecular. Most literature examples use coupled FVM (in continuum range) and DSMC approach (in transitional and molecular range) [1], [2], [3], [4] requiring thousands of HPC processor hours (core-hours) to obtain the result for a single snapshot. In this study, the coupled flow simulation is compared with pure FVM to check if obtaining correct particle contamination results is possible without time-consuming molecular simulations.

One of the most computationally challenging tasks is to simulate particle movement and interactions due to the enormous number of transported particles. The Apollo lander's number of moved particles can reach up to 1e11, assuming an erosion rate of the regolith is at 50 kg/s and a specific size distribution [5]. Storage of such data, assuming 6 DOF (degrees-of-freedom), would require 1e4 TB, and the computational expense of such a simulation would also be a significant obstacle to obtaining valuable results. To reduce the number of particles in simulation, a combined Eulerian-Lagrangian simulation approach is proposed. Larger particles will be simulated as 6 DOF, and an advected scalar field is considered for lower diameters. Figure 2 shows the possible division between Eulerian and Lagrangian approaches for diameter at the level of 0.2 mm.

The presentation will cover the software design [6] and selected contamination simulation results for the full-scale test cases.