ABSTRACT:We develop a robust and accurate three dimensional Eulerian method to compute droplet impingement on an aircraft body, wings, and engines. One of the principal challenges in developing a robust Eulerian method is the existence of singularities in some analytic solutions. We have found that, particularly for higher order methods, that the non-linearity of the equations facilitates the creation of spurious singularities. These result in an unbounded growth of particle density in localized regions. We propose an automatic, physics based adaptive numerical diffusion model whereby the existence of singularities is detected and used to derive a numerical diffusion stabilization term. Presence of the numerical diffusion appears to significantly improve the stability of higher order simulations. The model of droplet collection efficiency is implemented in CHEM solver in which a traditional finite volume scheme is applied to discretize both droplet and gas flows. The governing equations for droplets are presented, along with the analysis and design of the singularity diffusion model. Test results on three dimensional bodies are presented in comparison with NASA experimental impingement results. The numerical diffusion model is shown to be effective and simple.