基于ARIMA⁃LSTM的飞机液压泵性能趋势预测方法

Aircraft Hydraulic Pump Performance Trend Prediction Method Based on ARIMA⁃LSTM

针对飞机液压泵工作强度高、工作环境复杂而导致传统性能预测方法对飞机液压泵性能变化趋势预测精度不高的问题,提出了一种基于自回归积分滑动平均⁃长短期记忆(autoregressive integrated moving average⁃long short⁃term memory,简称ARIMA⁃LSTM)网络的飞机液压泵性能趋势预测方法。首先,将获取的飞机液压泵性能表征参数回油流量数据进行线性分解,得到趋势项数据和细节项数据;其次,采用自回归积分滑动平均(autoregressive integrated moving average,简称ARIMA)方法对趋势项数据进行预测,同时采用长短期记忆(long short term memory,简称LSTM)网络方法对归一化处理后的细节项数据进行预测;最后,将两部分预测结果进行叠加,得到最终的性能趋势预测结果。研究结果表明,采用ARIMA⁃LSTM的联合预测方法对飞机液压泵性能变化趋势的预测效果明显优于单一的ARIMA与LSTM预测方法,为飞机液压泵性能变化趋势预测的工程应用提供了一种新方法。

The traditional performance prediction method has low prediction accuracy for military aircraft hydraulic pump performance in high working intensity of military aircraft hydraulic pump and complicated working environment.Aiming at this problem,an military aircraft hydraulic pump performance trend prediction method is proposed based on the autoregressive integrated moving average(ARIMA)model and long short term memory(LSTM)network.Firstly,the obtained raw data of aircraft hydraulic pump performance characterization parameters are linearly decomposed to obtain trend item data and detail item data.Then,the ARIMA method is used to predict the trend item data,and the LSTM method is used to predict the normalized detail item data,and finally the two parts of the prediction result are superimposed to obtain the final performance trend prediction result.The research results show that the combined prediction method of ARIMA-LSTM is better than the single ARIMA or LSTM prediction methods for predicting the trend of aircraft hydraulic pump performance,which provides a new method for engineering application of aircraft hydraulic pump performance change trend prediction.