DNS of Turbulent Combustion in HCCI Engine by GPU-Cloud Supercomputer

为了解决环境和能源问题，均质充量压缩点火（HCCI）燃烧技术有望成为低排放、高效率的燃烧技术之一。本研究对GPU云超级计算机上的湍流燃烧直接数值模拟（DNS）程序进行了优化，开发了适用于实际动壁燃烧室的计算方案。对具有浓度和温度不均匀性的预混混合气的自燃和火焰传播进行了数值模拟，以揭示HCCI发动机中湍流燃烧机理和火焰-壁面相互作用。

To resolve the environmental and energy issues, homogeneous charge compression ignition(HCCI) is expected to be one of the promising combustion technologies in terms of low emission and high efficiency. In this research project, Direct Numerical Simulation(DNS) code of turbulent combustion has been optimized for GPU-cloud supercomputer and the computational scheme which can be applied to practical combustors with a moving wall has been developed. DNS on autoignition and flame propagation of premixed mixture with inhomogeneity of fuel concentration and temperature has been conducted to reveal turbulent combustion mechanisms and flame-wall interaction in HCCI engines.