Virtual real-time prediction of sensor soiling

Purpose and goal:

Sensors are absolutely crucial for active safety, driver support systems and self-driving vehicles and must be placed so that any contamination is as mild and slow as possible. In order to be able to do this at an early stage, efficient simulation models are needed. In this project, we will use the recurrence computational fluid dynamics (rCFD) technique to construct such models.

We will develop an experimentally validated rCFD solver for virtual real-time prediction of sensor soiling for vehicle applications. Expected results and effects:

The project results can be summarized as: 1) A validated rCFD methodology (solver and associated working methods) for virtual real-time prediction of sensor soiling; 2) Knowledge of differences and similarities between different contaminants in terms of soiling of realistic vehicle bodies; 3) General guidelines for sensor positioning based on rCFD simulations. The developed methodology will be ready to be applied in advanced technical development projects in the near future after the project is finalized.

Approach and implementation:

A state-of-the-art rCFD solver is implemented and its quality and performance are verified against literature data. New experimental data are generated through tunnel testing on simplified geometries and are used to further develop, verify and validate the rCFD solution for pollution problems relevant to vehicle aerodynamics. In the next step, new experimental data is generated for a real vehicle.

The rCFD solver is validated for this actual case and then the behavior of different types of contaminants is examined and guidelines are compiled.