Pitfalls for Accurate Steady-State Port Flow Simulations

Abstract

Steady-state port flow simulations were carried out with a commercial three-dimensional (3D) computational fluid dynamics (CFD) code using Cartesian mesh with cut cells to study the prediction accuracy. The accuracy is assessed by comparing predicted and measured mass-flow rate and swirl and tumble torques at various valve lifts using different boundary condition setup and mesh topology relative to port orientation. The measured data are taken from standard steady-state flow bench tests of a production intake port. The predicted mass-flow rates agree to within 1% with the measured data between the intermediate and high valve lifts. At low valve lifts, slight overprediction in mass-flow rate can be observed. The predicted swirl and tumble torques are within 25% of the flow bench measurements. Several meshing parameters were examined in this study. These include: inlet plenum shape and outlet plenum/extension size, embedded sphere with varying minimum mesh size, finer meshes on port and valve surface, orientation of valve, and port centerline relative to the mesh lines. For all model orientations examined, only the mesh topology with the valve axis aligned closely with the mesh lines can capture the mass-flow rate drop for very high valve lifts due to flow separation. This study further demonstrated that it is possible to perform 3D CFD flow analyses to adequately simulate steady-state flow bench tests.