VIBRATION FATIGUE LIFE PREDICTION OF TURBINE DISK BASED ON BP NEURAL NETWORKS

摘要：

A method is proposed based on BP (back-propagation) neural networks to predict the fatigue life of the turbine disk of rocket engine under random vibration loading conditions.With the combination of finite element model (FEM),the random load spectrum and damping ratio are utilized as inputs and turbine disk fatigue life are outputs in the BP neural networks.The orthogonal array (OA) technique is used to produce the training samples.Some noises are injected into the input samples during the learning process to enhance the generalization capability of the neural networks.Simulation results show that the method can provide reasonable vibration fatigue life for practical turbine disk of rocket engine.

关键词： turbine disk BP neural networks fatigue life life prediction

作者: Z.X.He S.W.Song F.Li X.T.Luan Q.Sun

作者单位: College of Aeronautics,Northwestern Polytechnical University,Xian,Shaanxi 710072,China Xian Aerospace Propulsion Institute,Shaanxi 710100,China

会议类型: 国际会议

会议名称: 第二届国际非均质材料力学会议(The Second International Conference on Heterogeneous Material Mechanics)

会议地点: 安徽黄山

会议语种:英文

页码: 1233-1236

在线出版日期: 2008-06-03（万方平台首次上网日期，不代表论文的发表时间）