A 2-D CFD model of oscillatory flow with jets impinging on a random wire regenerator matrix

This work describes the CFD portion of an overall plan to examine a modern type of Stirling engine regenerator matrix both computationally and experimentally. An experimental program has already started at the University of Minnesota (UMN) to examine a stacked-screen regenerator. A 2-D axisymmetric computational model was developed to simulate the UMN rig encompassing a piston/cylinder, cooler tube, plenum, regenerator matrix and heater. The CFD-ACE+, a commercial CFD-code, has been utilized for this study. A systematic approach was taken where CFD results were obtained for the following: 1) Turbulence modeling (low Re k-/spl epsiv/) of a unidirectional pipe flow with sudden expansion (expansion ratio D/d=1.95). Good agreement was achieved with the experimental data; 2) Laminar flow with porous media for a flow over and through a stack of rectangular parallel plate fins (porosity=0.579 & permeability=2.28E-08 m/sup -2/). Good agreement was achieved with published CFD data; 3) Good agreement was achieved for prediction of the oscillatory velocity field, in the plenum, with the UMN data and without the regenerator matrix where a low Re k-/spl epsiv/ turbulence model was used; 4) Results for the oscillatory flow data with the regenerator matrix in place showed that the laminar flow model overpredicted the centerline velocity (as expected), since the flow is turbulent. The laminar flow model with permeability of 1.65E-08 m/sup -2/ seems to show velocity uniformity in the matrix as early as 15 screen-layers deep into the regenerator matrix (from the cooler side). This is consistent with the UMN temperature data. A combined study of a turbulence model and a porous media model is underway to provide accurate data both in the plenum and the regenerator matrix. Our next step is to apply this model and include the energy equation to predict not only velocity but also temperature profiles as well.