Virtual real-time prediction of sensor soiling

Important results from the project

The project’s main objectives are considered to have been achieved. An rCFD-based methodology and an industrial workflow for the rapid prediction of contaminant deposition were developed, verified against conventional CFD and experimental data, and applied to a full-scale vehicle. The project also generated new knowledge about how turbulence, wake structures, and particle size affect deposition, as well as principles for sensor positioning and several scientific publications.

Expected long term effects

The methodology is expected to be integrated into industrial development processes and used to compare sensor positions and local geometric solutions early in vehicle development. This could contribute to shorter development cycles, less need for late-stage physical testing, and more robust sensor performance. The results may thereby support safer driver-assistance systems and automated driving while strengthening the competitiveness of the Swedish automotive industry.

Approach and implementation

The project was conducted as a doctoral research project in close collaboration between Chalmers, Volvo Cars, and Geely Technology Europe. The numerical development of rCFD was combined with conventional CFD, Euler–Lagrangian particle tracking, and wind-tunnel testing using water and artificial snow. The work progressed stepwise from reference cases and simplified geometries to a full-scale vehicle. The workflow was transferred to the industrial partners through close collaboration.