摘要
在同时考虑材料性能、几何参数和载荷随机性基础上,以遗传算法优化后的三层BP神经网络模拟非线性接触部位应变幅分布。而后建立应变-寿命分布模型,用蒙特卡洛法模拟随机化的Manson-Conffin公式,得出相应可靠性下的疲劳寿命。对航空发动机低压一级涡轮盘榫接触部位进行了疲劳寿命可靠性仿真,结果与实际吻合较好。
We take into simultaneous consideration the material performance, geometrical parameters and load randomness of a turbine, and simulate the nonlinear distribution of strain broadness in its contact region using the three-layered back propagation （BP） neural network optimized by genetic algorithm. Then we establish its strainlife distribution model, simulate the random Monson-Conffin equation with the Monte-Carlo method and thus obtain the fatigue life under the corresponding reliability. Finally we simulate the fatigue life and reliability of the contact region of an aeroengine＇s first-level low-pressure turbine, and the simulation results agree well with the actuality.
出处
《机械科学与技术》
CSCD
北大核心
2007年第5期585-588,共4页
Mechanical Science and Technology for Aerospace Engineering
关键词
应变幅
神经网络
遗传算法
疲劳寿命
可靠性
strain broadness
neural network
genetic algorithm
fatigue life
reliability
