A quasi-one-dimensional model for an outwardly opening poppet-type direct gas injector for internal combustion engines

摘要

Direct injection of compressed natural gas in internal combustion engines is a promising technology to achieve high indicated thermal efficiency and, at the same time, reduce harmful exhaust gas emissions using relatively low-cost fuel. However, the design and analysis of direct injection-compressed natural gas systems are challenging due to small injector geometries and high-speed gas flows including shocks and discontinuities. The injector design typically involves either a multi-hole configuration with inwardly opening needle or an outwardly opening poppet-type valve with small geometries, which make accessing the near-nozzle-flow field difficult in experiments. Therefore, predictive simulations can be helpful in the design and development processes. Simulations of the gas injection process are, however, computationally very expensive, as gas passages of the order of micrometers combined with a high Mach number compressible gas flow result in very small simulation time steps of the order of nanoseconds, increasing the overall computational wall time. With substantial differences between in-nozzle and in-cylinder length and velocity scales, simultaneous simulation of both regions becomes computationally expensive. Therefore, in this work, a quasi-one-dimensional nozzle-flow model for an outwardly opening poppet-type injector is developed. The model is validated by comparison with high-fidelity large-eddy simulation results for different nozzle pressure ratios. The quasi-one-dimensional nozzle-flow model is dynamically coupled to a three-dimensional flow solver through source terms in the governing equations, named as dynamically coupled source model. The dynamically coupled source model is then applied to a temporal gas jet evolution case and a cold flow engine case. The results show that the dynamically coupled source model can reasonably predict the gas jet behavior in both cases. All simulations using the new model led to reductions of computational wall time by a factor of 5 or higher.