The design of exhaust hood is a typical high dimensional,expensive computational and black box problem.Multi-Point Search based Efficient Global Optimization(MSEGO)is proposed to solve this problem.MSEGO is used for t...The design of exhaust hood is a typical high dimensional,expensive computational and black box problem.Multi-Point Search based Efficient Global Optimization(MSEGO)is proposed to solve this problem.MSEGO is used for the aerodynamic performance optimization of a low exhaust hood with non-axisymmetric outer flow guider.After optimization,the static pressure coefficient of the exhaust hood increases by 284.54%,and the aerodynamic performance analysis explains the reason of the improvement.Further,the analysis of variance(ANOVA)as the data mining technique is used to extract information of design space and analyze the influence of variables on the performance.Though aerodynamic performance analysis and data mining,it indicates that non-axisymmetric outer flow guider and the width of outer hood has a significant effect on the aerodynamic performance.Thereby,design lessons are derived and accumulated for the optimization of similar designs.展开更多
文摘The design of exhaust hood is a typical high dimensional,expensive computational and black box problem.Multi-Point Search based Efficient Global Optimization(MSEGO)is proposed to solve this problem.MSEGO is used for the aerodynamic performance optimization of a low exhaust hood with non-axisymmetric outer flow guider.After optimization,the static pressure coefficient of the exhaust hood increases by 284.54%,and the aerodynamic performance analysis explains the reason of the improvement.Further,the analysis of variance(ANOVA)as the data mining technique is used to extract information of design space and analyze the influence of variables on the performance.Though aerodynamic performance analysis and data mining,it indicates that non-axisymmetric outer flow guider and the width of outer hood has a significant effect on the aerodynamic performance.Thereby,design lessons are derived and accumulated for the optimization of similar designs.
