Surrogate-based optimization and experiment validation of a fan-shaped film cooling hole with a large lateral space

•A novel parametric shaping method is proposed to optimize diffusive film hole for large pitch space.•Variable-sensitivity oriented sampling improves the accuracy of Kriging surrogate model to accelerate optimization convergence.•Optimization result gets validated with experiment based on an infrareds thermography method.•The optimized hole shows 70 % higher overall film cooling effectiveness than the baseline hole. The exit shape of a fan-shaped hole plays an important role to improve the film cooling effectiveness. In this study, a fan-shaped cooling hole on a flat plate was optimized through Reynolds Averaged Navier-Stokes (RANS) analysis for a specific cooling application of unusually large lateral span. A novel hole was configured with four linear dimensions including the leading and trailing edge lateral widths, the exit longitude, and the round metering length. Two rounds of Latin Hypercube Samples were used to narrow the design space for an accurate surrogate model. The first samples were used to evaluate individual variable sensitivity. The second samplings of predominant variables were then combined with the first sampling to build a Kriging surrogate model to predict the film cooling effectiveness of new candidates. A multi-island algorithm was adopted to obtain the optimization solution. The enhanced cooling effectiveness of the final optimal design got verified in an experiment. Results show the improved prediction accuracy of Kriging model and the fast convergence rate of the optimization solution. For a cooling case with a relatively large lateral span, an optimal fan-shaped hole intends to increase the lateral expansion width and to reduce longitude opening width. The spatial averaged cooling effectiveness of the optimized hole achieved a 70% increment over the baseline one.