Modern cars are becoming more advanced, with several automated features offloading, or completely eliminating, certain tasks from the driver. Systems for active safety, driver support and self-driving promise to reduce risk and impact of traffic accidents, thereby increasing safety. These computerized systems rely on sensors for understanding the environment and conditions they are operating in. However, these sensors can be negatively impacted if they are covered by natural contaminants found in the environment such as snow, dust or rain. It is crucial that these sensors are kept as unobstructed as possible for the correct operation of these safety systems. An important design aspect for car manufacturers is then how these sensors should be kept clean. Sensor placement and exterior design features are crucial factors that influence the exposure of the sensors to soiling. These aspects can in principle be studied using theoretical tools such as computational fluid dynamics, where soiling particles can be tracked in simulated air flow over the vehicle. The computational tools give much information about the flow in question, but they suffer from extremely large costs in terms of computer processing power. For a full car, these requirements are so significant that it would take years to simulate the deposition process on a single vehicle. As a way to mitigate these issues, several data-assisted simulation methods have been proposed that use precomputed flow-field data to approximate the computationally intractable complete solution. One such method is recurrence computational fluid dynamics (rCFD). This method is able to approximately reproduce flow fields, as long as there is some degree of periodicity in time in the flow. When using this method, the required CPU-hour resources are significantly reduced by instead generating a database of precomputed flow fields and then replaying them multiple times in a certain pattern. The prohibitively large amount of required CPU-hours for the full solution is then exchanged for increased storage requirements to get an approximated solution. In order to evaluate and validate the rCFD method and the developed computational tools, this project starts with investigating simplified geometries and then increases geometric complexity in steps until a full car has been simulated using the method. This strategy provides quick development cycles and minimizes computational resource requirements in the development process. The results of this research are applicable in fields where particle dispersion and deposition are important, such as soiling in the car industry, icing of aircraft and ash build-up in boilers.