Single-loop Method for Reliability-based Design Optimization of Turbine Blades with Poor Information

Turbine blade design optimization is a typical reliability-based multidisciplinary design optimization (RBMDO) problem. Due to the high experimental cost and cycle, the experimental data is limited and the design task will face the situation of poor information. In this paper, the poor information is described with interval variable and the RBMDO model based on non-probabilistic reliability is established. A single-loop method for RBMDO is proposed to improve the efficiency of turbine blade design, in which a two-stage phased construction method is presented to establish the response surface model to approximate the reliability constraint. The results of turbine blade design show that the proposed method has higher calculation efficiency with acceptable calculation accuracy and the calculation accuracy is affected by the accuracy of the response surface.