Reliability Analysis of Blade Fatigue Life Based on Fuzzy Intelligent Multiple Extremum Response SurfaceMethod
Chun-Yi Zhang,Tian Sun,Ai-Hua Wang,Hui-Zhe Jing,Bao-Sheng Liu,Cheng-Wei Li
DOI: https://doi.org/10.2298/fil1805897z
2018-01-01
Filomat
Abstract:In order to more reasonable analyze the dynamic reliability of aero-engine blade with coupling failure mode. A fuzzy intelligent multiple extremum response surface method (FIMERSM) was proposed. Considering the coupling effect of temperature load and centrifugal load, the maximum stress point, the maximum strain point and the minimum life point on blade were found by deterministic analysis. Then, the density of blade, rotor speed, elastic modulus, blade-tip temperature, blade-root temperature, fatigue strength coefficient, fatigue strength exponent, fatigue ductility coefficient, fatigue ductility exponent, blade width, blade thickness, blade torsion angle, and blade height as input random variables. By using Latin hypercube sampling technique, the sample values of the input random variables were acquired and finite element basic equation was calculated for each samples which obtained the corresponding dynamic output response of their stress, strain, and low cycle fatigue life within the analysis time domain. By taking the entire maximum values of the dynamic output response in the analysis time domain as new output response, the fuzzy intelligent multiple extremum response surface function (FIMERSF) was established. Finally, the dynamic reliability of the blade structure were obtained by using the Monte Carlo method (MCM) large amount linkage sampling of the input random variables and take it into the FIMERSF to calculate the output response. The results imply that the comprehensive reliability of blade is 99.46%. Through the comparison of MCM, Multiple extremum response surface method (MERSM) and FIMERSM, the computational results show that the FIMERSM has high computational precision and computational efficiency.