航空发动机性能参数预测方法  

Forecasting method for aeroengine performance parameters

作　　者：李晓白[1] 崔秀伶[1] 郎荣玲[1] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100083

出　　处：《北京航空航天大学学报》2008年第3期253-256,共4页Journal of Beijing University of Aeronautics and Astronautics

基　　金：武器装备预研基金资助项目(51417010201HK0155)

摘　　要：航空发动机性能参数预测对于发动机的视情维修具有重要的意义.为了提高预测精度,在分析发动机性能参数数据特点的基础上,提出了一种新的应用于此领域的组合预测模型.首先利用小波变换将原始数据分解为不同尺度上的几组子序列,根据各子序列的特点分别选用自回归滑动平均(ARMA,Autoregressive Moving Average)模型或求和自回归滑动平均(ARIMA,Autoregressive Integrated Moving Average)模型进行预测,然后将所有预测结果合成,得到最终预测结果.通过仿真实验,验证了该组合模型提高短期和中长期预测精度的有效性,并分析了小波分解层数对于预测精度的影响.The forecasting of aeroengine performance parameters is very important for aeroengine maintenance based on condition. To improve the forecasting accuracy, a new combination method was proposed for forecasting parameters based on analyzing the data. Firstly, the original sequence was decomposed by wavelet transform and some sub-sequences in different frequency band were obtained. Then these sub-sequences were forecasted by ARMA/ARIMA respectively. Finally, the forecasting results of all sub-sequences were reconstructed and taken as the final forecasting result. Through the test, the proposed combination model was proved to be highly effective on improving the accuracy of the short-term and long-term forecasting, and the effect of wavelet decomposition levels on forecasting accuracy was analyzed.

关 键 词：组合预测 自回归滑动平均模型 求和自回归滑动平均模型 排气温度裕度 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]