Investigation on Multi-Row Film Cooling Calculation Based on the Source Term Model

Film cooling technology is an effective method for reducing the wall temperature of high-temperature components in aero-engines and is one of the primary thermal protection measures. Conventional film cooling calculation methods heavily rely on highly refined mesh generation, which complicates the modeling of film cooling hole structures. To simplify the complex mesh requirements, this study proposes a source term model-based film cooling calculation method. Based on flat-plate film cooling experimental data, a neural network-based deep learning approach was employed to establish the relationship between the discharge coefficient and operating conditions as well as film cooling hole parameters. The effects of different operating conditions and film cooling hole parameters on the discharge coefficient were analyzed. Furthermore, the wall temperature distributions, pressure distributions along the flow path, and temperature values at multiple points obtained from the source term method were compared with experimental results and full-size numerical simulation results. The findings demonstrate that the discharge coefficients derived from neural network training facilitate accurate calculation of source term parameters. The source term model can effectively replace the real film cooling hole model for numerical simulations of film cooling structures, achieving high computational accuracy.