论文部分内容阅读
针对反映转子系统工作状态的特征参数时间序列具有不确定性的、差异较大的分段函数变化规律的特点 ,提出了一种组合式神经网络转子系统状态预测模型。该模型将故障诊断和状态预测有机地结合起来 ,利用转子系统当前状态特征参数样本 ,通过故障诊断系统判断预测时的转子系统工作状态模式 ;从多种神经网络预测模型组合而成的预测模型库中调用同该工作状态模式相应的神经网络预测模型 ,对多种特征参数时间序列进行预测 ;依据预测出的未来某一时刻的多种特征参数 ,利用故障诊断系统判断转子系统的未来工作状态模式。仿真试验结果表明 ,该模型可以对转子系统状态进行可靠的预测。文中详细讨论了模型的建立和仿真实验结果。
Aiming at the characteristics of the subsection function changing law with uncertainties and large differences in the time series of the characteristic parameters reflecting the working state of the rotor system, a state prediction model of rotor system with combined neural network is proposed. The model combines fault diagnosis and state prediction organically. By using the current state characteristic parameters of the rotor system, the fault diagnosis system can be used to determine the operating mode of the rotor system. The prediction model library is composed of a combination of neural network prediction models , A neural network prediction model corresponding to the working mode is invoked to predict a plurality of time series of characteristic parameters. Based on the predicted characteristic parameters of a certain moment in the future, the fault diagnosis system is used to determine the future working mode of the rotor system . Simulation results show that the model can predict the state of the rotor system reliably. In this paper, the establishment of model and simulation results are discussed in detail.