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. The blade is simplified as a cantilever beam model. According to the theory of vibration mechanics, the relationship between the blade damage degree, the change of natural frequency and the damage position is deduced. It is abstracted from the position of damage and the variation of natural frequency Based on which the artificial neural network with self-learning capability and approximation ability of nonlinear mapping is introduced into the prediction of the damage degree, a damage degree of blade based on BP (back-propagation) neural network The method is validated by using the experimental data as the training and testing samples.The results of the study can be applied to the engineering treatment of damage vanes and the development of maintenance codes.