航空发动机气路参数偏差值平滑处理技术  

Smoothing technology of aero-engine parameter's deviation

作　　者：钟诗胜[1] 崔智全[2] 付旭云[1] 王体春[3] 

机构地区：[1]哈尔滨工业大学(威海)船舶工程学院,山东威海264209 [2]哈尔滨工业大学(威海)汽车工程学院,山东威海264209 [3]南京航空航天大学机电工程学院,江苏南京210026

出　　处：《计算机集成制造系统》2013年第6期1199-1204,共6页Computer Integrated Manufacturing Systems

基　　金：中国民航总局科技资助项目(MHRD201052);国家863计划重点资助项目(2012AA040911);国家自然科学基金重点资助项目(60939003)~~

摘　　要：为保留反映航空发动机故障信息或操作状态信息的真实突变数据,提出一种离群点挖掘与类指数平滑算法相结合的数据平滑处理技术。采用基于统计学的滚动样本离群点挖掘算法进行离群点挖掘,对非离群点的突变数据采用基于权重分配的类二次指数平滑算法进行滚动平滑。以厂家系统输出的平滑值作为学习目标,以均方误差最小化为目标进行参数优化。采用所提出的平滑处理技术及十点平滑算法对偏差值进行平滑处理,并与厂家系统的原始平滑值对比可知,提出的数据平滑处理技术能够保留真实突变数据,同时实现发动机气路参数偏差值的合理平滑。To keep the true mutation data which contained aero-engine fault or operation status information, a data smoothing technology was proposed by combining the outlier mining and similar exponential smoothing algorithm. The rolling sample outlier mining algorithm based on statistics was used to mine the outliers, and the similar double exponential smoothing algorithm based on weight was used to smooth the un-outlier in rolling mode. The smoothed data of factory was taken as learning objectives to optimize the parameters with minimum mean square error. Deviations were smoothed by the proposed smoothing technology and ten point smoothing algorithm. By comparing with original smoothing data of factory system, the result showed that the proposed smoothing technology could preserve the real mutation data, and the other engine gas path parameter deviation could smoothed reasonably.

关 键 词：航空发动机 气路参数 偏差值 离群点 类指数平滑算法 

分 类 号：TP311.132[自动化与计算机技术—计算机软件与理论] V263.6[自动化与计算机技术—计算机科学与技术]