Analysis of Balance Between Modeling Accuracy and Computational Speed for a Hybrid Ray-Tracing and Radiosity Method Used in Lighting Simulation

A new hybrid ray-tracing and radiosity method was developed to adopt a balance between computational speed and accuracy when simulating building interior illuminance distributions. The method is able to model spaces with highly specular components, highly diffuse components (i.e., walls, floors, roller shades), as well as components with mixed properties that strongly affect the daylighting performance of the space (i.e., venetian blinds and light-shelves). The specular components are processed with a forward ray-tracing method, while the diffuse components are processed with a radiosity method. The concept of shining factors is used to separate luminous flux though components with mixed properties, while parameter settings affect both simulation speed and accuracy of results. In this paper, the amount of light rays (tracing resolution) used in the ray-tracing part and the grid size (space resolution) used in the radiosity part are investigated to analyze the impact of those parameters when predicting work plane illuminance and lighting energy consumption. Use of high numbers of generated rays and high grid density can provide detailed illuminance results. However, the penalties associated with computational time may exceed the accuracy requirements. Recommendations for best settings for different modeling cases are provided so as to balance accuracy versus computational speed. The outcomes of this study can provide useful information for modeling daylighting performance of spaces with specular, diffuse and mixed reflection components such as blinds and light-shelves.