A new adaptive mesh refinement to model water flow around fishing nets

This paper presents a new adaptive mesh refinement strategy which offers an alternative to the classic AMR methods. The automatic local refinement needs only few additional storage and has good cpu-time results which makes it a fine candidate for some specific studies. It is here applied to a fluid-structure interaction problem: the prediction of the behavior of a submerged net for fisheries applications from fishing companies and scientific questioning. Indeed, the structure code modeling the net already uses a lot of memory space so we introduce a scheme which does not require data structure. The original algorithm employs a local numbering of the cells to identify the neighbor of every cell. The auto-adaptive mesh is applied to the fluid model for a test case describing the hydrodynamic behavior of a netting panel.