本文利用人工神经网络自组织学习和神经元连接的类型(兴奋或抑制)与征兆和故障之间的特征映射,建立了旋转机械故障诊断的自组织特征映射模型.利用这一模型使人工神经网络无指导学习用于旋转机械故障诊断,通过旋转机械中常见故障进行诊断提供了进一步解决故障诊断专家系统中知识库建立及维护的新方法. In this paper, a self-organizing feature mapping model of rotating machinery fault diagnosis is established by using the neural network self-organizing learning and the type of neural connections (excitement or suppression) and the feature mapping between symptoms and faults. Using this model, artificial neural network Non-directed learning is used to diagnose rotating machinery faults and to diagnose common faults in rotating machinery. A new method to further solve the problem of establishing and maintaining a knowledge base in a fault diagnosis expert system is provided.