基于组合模型的航空发动机推力衰减短期预测研究

Short-Term Prediction Research of Aero-Engine Thrust Deterioration Base on Combined Model

摘要: 为了提高航空发动机推力衰减的短期预测精度，提出了基于ARIMA和BP神经网络的组合模型。阐述了ARIMA-BP组合模型的预测步骤，建立了发动机推力衰减的预测模型。利用ARIMA模型预测推力整体变化趋势，利用BP神经网络预测非线性剩余变化量，进而完成对发动机推力的预测分析。并与ARIMA模型预测结果进行对比。结果表明：组合模型短期预测精度较高，为确保发动机试车安全提供可靠的方法，具有一定的工程应用意义。

Abstract: In order to improve the short-term prediction accuracy of aeroengine thrust deterioration, a combined model based on ARIMA and BP neural network was presented. The prediction procedure of ARIMA-BP combined model was expounded, and the prediction model of thrust deterioration was established. The ARIMA model was used to predict the overall thrust variation trend, and the BP neural network was used to predict the non-linear residual variation, and then the prediction and analysis of engine thrust were completed. The results were compared with ARIMA model predictions. The result showed that the short-term prediction accuracy of the combined model was higher, which provided a reliable method for ensuring the safety of engine test run and had certain engineering application significance.