主要研究进化神经网络在旋转机械故障诊断中的应用 ,提出了一种基于递阶结构的遗传算法与进化规划相结合的神经网络学习新算法 ,利用该算法可以同时对网络进行结构优化和权重求解。通过旋转机械故障分类应用实例 ,与传统的 BP训练算法作了比较 ,证明基于递阶结构的进化神经网络算法不仅在权重训练方面比传统 BP训练算法更加快速稳定 ,避免陷入局部极小点 ,而且同时对网络结构进行了优化 ,得到了结构更为简捷的旋转机械故障分类网络 This paper mainly studies the application of evolutionary neural network in the fault diagnosis of rotating machinery, and proposes a new neural network learning algorithm based on hierarchical structure genetic algorithm and evolutionary programming, which can be used to optimize the structure and weight of the network simultaneously . Compared with the traditional BP training algorithm, the hierarchical ANN evolutionary algorithm is proved to be faster and more stable than the traditional BP training algorithm in terms of weight training and avoid falling into local minima At the same time, the network structure has been optimized, and a more simple structure of the rotating machinery fault classification network