Abstract: In order to eliminate the systematic error of Gas Path Analysis(GPA) caused by the averaging effect stemmed from Multiple Operating Points Analysis(MOPA), Sequential Operating Points Analysis(SOPA) is proposed for the first time based on aeroengine transient operating process, based on which a systematic GPA parameter selection method is introduced. This method takes advantage of the enhanced delineation of time signal provided by wavelet transforms to extract health parameter signatures, according to which the least number of requisite gas path sensors is determined, under the constraint of limited measurements. Singular Value Decomposition(SVD) is applied to SOPA subsystem matrices to attain their identifiabilities of the to-be-estimated health parameters. Applications to a high bypass ratio dual-spool separated flow turbofan engine indicate that the irreducible sensor configuration determined by wavelet analysis of output sensitivities is sufficient to identify the deviation of every to-be-estimated health parameter. The condition number of transient snapshot matrix is an effective metric of SOPA subsystem identifiability to determine the optimal position for SOPA application within the transient process, which guarantees the estimation accuracy of engine gas path component health status.

Key words: Transient state based gas path analysis;Aeroengine;Sequential operating points analysis;Wavelet transform;Singular value decomposition

摘要： 为了降低由多工作点分析（MOPA）方法的平均效应所产生的气路分析（GPA）系统误差，提出了基于航空发动机过渡工作过程的序列工作点分析（SOPA）技术，并以此为基础提出了一种系统的气路分析参数选择方法。该方法利用连续小波变换对时间信号的增强解析能力，提取待求健康参数在备选测量传感器上的参数特征，实现了在传感器安装受限条件下必要测量参数的选择。通过对SOPA子系统矩阵进行奇异值分解（SVD），获得了在过渡工作过程中不同时间片段上的健康参数可辨识性。针对大涵道比双轴分排涡扇发动机的参数分析结果表明：通过对待求健康参数的敏感性输出信号进行小波分析所确定的最简传感器布局，具备对全部待求健康参数的可辨识性；而以时间片段矩阵的条件数作为判据评估SOPA子系统的参数辨识能力，能够有效地确定具有高可靠性的SOPA时间片段位置，保证了对发动机气路部件健康状态的估计精度。

关键词: 过渡态气路分析;航空发动机;序列工作点分析;小波变换;奇异值分解