故障诊断模型是开展旋转机械设备状态检修的核心环节之一。采用弹性反馈(RPROP)神经网络算法建立旋转机械的神经网络故障诊断模型,通过与标准BP、带动量因子的反向传播(BP)算法相比较,表明了RPROP算法在故障模式识别中具有良好的学习效率与泛化能力,可以更好的描述频率特征和诊断结果之间的关系。 Fault diagnosis model is to carry out the rotating machinery and equipment state maintenance one of the core links. The neural network fault diagnosis model of rotating machinery is established by using RPROP neural network algorithm. Compared with standard BP and back propagation (BP) algorithm with momentum factor, it is shown that RPROP algorithm has good performance in fault pattern recognition Learning efficiency and generalization ability, can better describe the relationship between frequency characteristics and diagnostic results.