【摘要】 将叶片简化为悬臂梁模型,根据振动力学理论推导出了叶片损伤度与固有频率变化量以及损伤位置两个参数的内在关系,将其抽象为一种从损伤的位置及固有频率的变化量到损伤度的映射.在此基础上,将具有自学习能力和逼近非线性映射能力的人工神经网络引入到损伤度的预测中,构建了一种基于BP(back-propagation)神经网络的叶片损伤度评定方法,并以具体实验数据作为训练和测试样本验证了该方法的有效性.研究结果可应用于损伤叶片的工程处理以及维修规范的制定.更多还原

【Abstract】 The internal relationship among damage degree,natural frequency and the location of damage was deduced by simplifying the blade as a cantilever.In view of the problem in the maintenance of damaged blade,the relationship was Abstracted into a mapping about damage degree,natural frequency and the location of damage.Then the neural network,which has the ability of self-study and can approach the nonlinear mapping,was used in the prediction of the damage.The damage degree assessment method of blade was built up based on the back-propagation(BP)neural network.Furthermore,this method was validated by using the experimental data of a certain aero-engine blade as the sample and test data.The results indicate that this method is useful to disposal of damaged blade in the project applications and can be applied in the maintenance criterion of blade.更多还原