文摘
To evaluate the risk of complex structures such as a 2.5D/SiC composites turbine blade, Response Surface Methodology is applied to investigate the reliability due to its high efficiency. In this paper, the Response Surface is based on Support Vector Machines (SVM), and new sample strategies are proposed to assess the failure domain and probability as well as reliability index with less computational cost and higher accuracy. 2.5D/SiC composites are defined by 5 geometric parameters to represent their architecture. Using the finite element method, we establish the structure of 2.5D/SiC composites and predict the mechanical properties with double scale models. The stochastic behaviors of load, material strength and microstructure of 2.5D/SiC composites are considered to analyze the reliability of a turbine blade in an aircraft engine. The methodology in this paper provides an accurate and effective way to value the risk of turbine blade design.