Analysis of discrepancies between 3-D coupled and uncoupled schemes based on CFD in full engine simulation

Abstract

Full engine simulations based on three-dimensional computational fluid dynamics (3-D CFD) allow the visualisation of flow details accompanied by total performances for aeroengine design. In this study, 3-D coupled and uncoupled schemes are performed to simulate the full engine performance of a micro turbojet engine. For the 3-D uncoupled scheme, the general performance map of each core component based on CFD simulation is used as a component model for solving the equations of thermodynamic equilibrium without aerodynamic data crossing the boundary of each component. For the 3-D coupled scheme, a full engine simulation is performed using a Reynolds-averaged Navier-Stokes CFD solver coupled with a power balance iteration, and all components are assembled in one computational domain. The calculation processes of these two schemes are introduced, and the feasibility of using the Regula-Falsi method in the coupled scheme for power balance iteration is verified. Despite some discrepancies, the overall performance results yielded by these two schemes are consistent with the experimental results. A detailed comparison of these two schemes is presented to elucidate the reasons contributing to the discrepancies. The results show that the uncertainty of the boundary conditions limited by the 3-D uncoupled simulation method and the low-dimensional forced mixing at the component's interfaces can result in significant calculation errors. Interactions among core components are demonstrated methodically. Finally, based on a comparative analysis with the 3-D coupled scheme, correction strategies for the component data-based model are proposed to improve the accuracy of uncoupled simulations. With these coupling corrections, the maximum relative error of the thrust between these two schemes reduces by 50% and the result is more close to the experiment.