On the Use of the Immersed Boundary Method for Engine Modeling

Historically, boundary fitted grids have been used to model complex geometries. The authors are investigating a novel grid generation concept, Immersed Boundary Method, as an improved methodology to boundary fitted grids. There are two significant disadvantages to using a boundary fitted grid. First, whether structured or unstructured, fitting a grid to a complex geometry prevents the use of simple orthogonal grids, thereby foregoing the benefits of numerical accuracy and computational efficiency associated with simple orthogonal grids. Second, generating a boundary fitted structured grid for a complex geometry can be extremely time consuming and difficult. Often, the grid generation difficulties and time can be a roadblock in simulating complex geometries such as an internal combustion engine. These problems can be overcome by using the immersed boundary method. The immersed boundary method models complex geometries by representing the necessary shape through forces within the grid. This concept was first introduced by Peskin (1972) for performing simulations of flow in a heart. The work of Peskin allows for a two-way coupled treatment of the immersed boundary. However, this method has severe drawbacks in terms of CFL requirements for stability. Other researchers (Mohd-Yusof, 1997, Verzicco et al., 2000, Fadlun et al., 2000 ) have improved on Peskin’s technique for the specific case of a solid wall where there is no need to couple the effect of the fluid to the wall treatment. By using the immersed boundary approach, the underlying grid need not coincide with the geometry being solved. Since the grid does not need to coincide with the surface geometry, the type of grid used is chosen for computational efficiency instead of geometry, which allows the use of simple orthogonal grids. Furthermore, the grid generation complexity and time is greatly reduced, as the complex geometry only needs to be mapped onto the underlying orthogonal grid. Currently, Convergent Thinking in cooperation with the University of Massachusetts—Amherst is developing the immersed boundary code MOSES (MOdifiable Source Engine Simulation) for use in engine simulations. In this paper, we present immersed boundary background, improvements to the immersed boundary method, progress in the development of the code, and MOSES simulation results.