论文部分内容阅读
针对航空发动机滑油系统状态监测问题,提出了递归过程神经网络模型。其隐层和输出层为过程神经元,该网络的输入信号为时变函数或过程,并且含有一个特别的关联层,在建模过程中能储存系统过去更多时刻的状态信息,使得网络结构适于预测时间序列问题。文中给出了相应的学习算法,并且分别利用人工神经网络和递归过程神经网络对航空发动机滑油系统状态进行预测。结果表明,递归过程神经网络预测精度高,优于传统人工神经网络的预测能力。为航空发动机滑油系统状态监测问题提供了一种有效的方法。
Aeroengine oil system condition monitoring problem, put forward a recursive process neural network model. The hidden layer and the output layer are process neurons. The input signal of the network is a time-varying function or process, and contains a special association layer. During the modeling process, state information of the system at more moments in the past can be stored, so that the network structure Suitable for predicting time series problems. The corresponding learning algorithm is given in this paper, and the status of aeroengine oil system is predicted by using artificial neural network and recursive process neural network respectively. The results show that the recursive neural network prediction accuracy is higher than the traditional artificial neural network prediction ability. It provides an effective method for the condition monitoring of aeroengine oil system.