The potential blinding of Advanced Driver Assistance Systems (ADAS) sensors due to contamination poses a notable threat to autonomous vehicles. These sensors' performance can be compromised by diverse sources such as dust, water, or snow. However, our investigation concentrates primarily on snow-related contamination, a frequent occurrence during winter. The accumulation of snow and ice significantly hampers the operational efficacy of autonomous vehicles. Over the years, a series of field tests and wind tunnel experiments have been conducted to analyze the mechanisms of snow interaction and soiling patterns on vehicles and bluff bodies. Notably distinctive patterns of soiling have been identified across multiple areas of these structures. The central challenge revolves around constructing an accurate model to predict snow buildup on vehicles. The precision in capturing the airflow dynamics, which substantially influences how the snow interacts with the specific body, is crucial for forecasting the resulting soiling patterns. This project's objective is to establish a digital simulation approach for depicting interactions between air, snow, and metal surfaces. This methodology is substantiated by comparisons with established wind tunnel findings.
The study utilizes simplified bluff bodies, specifically employing a Wedge geometry, to minimize unexpected errors. For aerodynamic simulations, STAR-CCM+ featuring the Detached Eddy Simulation (DES) turbulence model was adopted. This effectively captures transient aerodynamics, particularly in the wake region, which has been identified as a critical aspect influencing soiling patterns. The Smooth Particle Hydrodynamics (SPH) based tool, PreonLab, was employed to simulate snow particle accumulation. An extensive parametric analysis within the PreonLab snow solver was conducted, exploring its competency in simulating snow particles and their interactions with airflow and bluff bodies. A comprehensive framework for simulating snow accumulation was formulated, encompassing snow properties, drag models, adhesion, and roughness of surface. The simulation outputs demonstrate good alignment with the outcomes of wind tunnel tests.