Comparison of the Results of Modeling a Dispersed Medium by Wave and Ray Methods

A large number of works on dispersed medium modeling use either pure ray optics or light transport equation in which propagation of light obeys geometric optics while scattering properties of the medium can be either calculated with wave optics or measured. In either case the distance between individual particles must be much greater than wavelength. At the same time current computer power allows to simulate paint layer with wave optics. We decided to compare paint simulation done by the scalar wave approach and by ray tracing with individual particles. One of the goals of this work is to verify the correctness of ray tracing results for various sizes of metal flakes used often in production of metallic or pearlescent paints. Ray tracing had been done in two variants. One assumes the flakes have perfectly mirror reflection, while in the other variant the reflection is slightly diffuse with the angular distribution taken from the Fraunhofer diffraction on thin disk. For not too large flakes results of these two approaches substantially differ. The second “hybrid” method is considerably closer to the wave optics results.

A Self-Consistent Transient Paint Simulation

Optimizing the painting process involves understanding the many nonlinear dynamics that effect paint transfer efficiency (PTE) such as bell applicator design (bell cup and shaping air ring), operating conditions (shaping air flow rate, paint flow rate, electric potential, downdraft, and bell rotation speed), target location, and paint type. PTE optimization results in less experimental testing and improved paint quality. We have modified a numerical simulation to include self-consistent momentum coupling between a continuum and discrete mass particles, as well as the inclusion of a self-consistent electrostatic field. These modifications enable the simulation tool to address several seminal questions concerning PTE. Several simulations have been performed and the influence of self-consistent modeling is addressed in this paper.