Enhanced view-dependent adaptive grid refinement for animating fluids
Rinchai Bunlutangtum, Pizzanu Kanongchaiyos · 2011
This paper presents a method for animating fluids on an octree grid. Our approach combines dynamic grid refinements with viewing information i.e., viewing angle, camera-to-fluid distance and resolution ratio of simulation grid over output image in order to optimize the grid for speed as well as preserve the animation quality. Our method reduces and removes fine details that are unnoticeable by means of level-of-detail. We determine whether to merge or divide the grid by using our proposed refinement conditions. In addition, particles, which are more flexible to conform to obstacle-fluid boundaries, are integrated in to our model to enhance animations and reduce the artefacts caused by dynamic refinements. Our method is fast and easily implemented, yet capable to be integrated as an extension of a current grid-based solver.