A Nonlinear Reduced Order Method with Overset Adaptive Cartesian/Unstructured Grid for Moving Body Simulations
Navy SBIR FY2008.2

Sol No.: Navy SBIR FY2008.2
Topic No.: N08-118
Topic Title: A Nonlinear Reduced Order Method with Overset Adaptive Cartesian/Unstructured Grid for Moving Body Simulations
Proposal No.: N082-118-0405
Firm: CFD Research Corporation
215 Wynn Dr., 5th Floor
Huntsville, Alabama 35805
Contact: H. Yang
Phone: (256) 726-4824
Web Site: www.cfdrc.com
Abstract: An innovative hybrid nonlinear reduced order method (ROM) for deforming unstructured mesh and overset Cartesian grid for moving body problems is proposed. Utilizing the properties of a maximum 6 degrees of freedom (DOF) of a moving body, the proposed method computes the grid deformation using nonlinear large deformation theory to preserve the original grid quality under each DOF, so that no extra computations are required during the unsteady motion, just matrix-vector multiplications. The introduction of the overset Cartesian method will provide the flexibility when dealing with extremely large deformations and deterioration of grid quality. The present method has the merits of reduced grid distortion, minimum modification to an existing code, efficient moving grid methodology with simple matrix multiplication, and minimized needs of hole-cutting and donor cell searching of overset grid. The Phase I effort is on the testing of each individual components of the method, including nonlinear ROM for unstructured deformation; overset Cartesian grid for coupling with an unstructured grid; a controlling strategy and criteria to switch between the two methods; and the successful demonstration. In Phase II, an API will be developed so that the proposed methodology is generic enough to be utilized by a variety of solvers and flow functions. The technology will be demonstrated on USM3D solver.
Benefits: The proposed nonlinear reduced order method is an innovative approach to handle the dynamic meshes in moving body applications. This concept will have a wide-ranging implication to simulation of complex fluid-structure interactions in aerospace and automotive applications. The immediate relevance is to the JSF program in accurate prediction of store separation. The commercial outcome of this development will be phenomenal as even some of the sophisticated CFD codes lack an elegant framework for mesh handling in moving body applications.