航空发动机磨损趋势预测模型研究

文章讨论了神经网络的BP算法和遗传算法,提出用遗传算法来优化BP神经网络,应用遗传算法训练神经网络权重,实现网络结构的优化,用优化后的BP人工神经网络建立了航空发动机磨损故障趋势预测模型,利用发动机的光谱监测数据作为预测磨损趋势的特征参数,进行了模型的训练和预测试验,并将该模型预测结果与BP算法和多元线性回归法的预测结果进行了比较,证明了基于遗传算法的人工神经网络是航空发动机磨损故障趋势预测的一种理想方法。

Research on Aircraft Engine Wear Trend Prediction Model

In the paper, the BP algorithm of neural network and genetic algorithm were discussed, the BP neural network was optimized by genetic algorithm, neural network weight was used to get the weight of genetic algorithm in order to optimize the network structure, then the optimized BP artificial neural network was used to build the aircraft engine wear fault trend prediction model. The spectrum monitoring data of the engine was used as a characteristic parameter to predict the tendency of the wear and tear, then the training and prediction of the model experiment were carried out, and the model prediction results was compared with the BP algorithm and the predictive results of the multiple linear regression method, which proved that the artificial neural network based on genetic algorithm was an ideal aircraft engine wear fault trend prediction method.