论文部分内容阅读
转子轴心轨迹形状反映转子系统的工作状态，用神经网络方法可识别转子轴心轨迹形状，但神经网络的训练速度和稳定性与网络输入数据编码方式有关，提出一种使轴心轨迹图形编码得到较大压缩的平面图形可变等长度压缩编码方法，从而减少了轴心轨迹神经网络识别系统的输入变元数，使训练后的神经网络的联想能力得到较大提高，也加快了网络的训练速度及稳定性．将该图形形状识别系统加入到故障诊断专家系统中，提高了故障诊断专家系统的自动诊断水平．
The shape of the rotor axis trajectory reflects the working state of the rotor system, and the neural network method can be used to identify the trajectory shape of the rotor axis. However, the training speed and stability of the neural network are related to the encoding method of the network input data. Which can reduce the number of input variables of the neural network identification system of axis trajectory, greatly improve the associative ability of the trained neural network, and speed up the network’s Training speed and stability. The graphic shape recognition system is added to the fault diagnosis expert system to improve the automatic diagnosis level of the fault diagnosis expert system.