以小波分析和主元分析PCA理论为基础,针对PCA单尺度建模的局限性,将PCA的去线性变量相关性的能力以及小波变换提取变量局部特征和近似分解变量自相关性的能力结合起来形成多尺度主元分析法(MSPCA)。将其应用于浮式油轮生产储油卸油系统(FPSO)油气水分离系统。研究结果表明:基于多尺度主元分析的故障诊断方法可以准确地对FPSO生产过程的早期故障进行检测和诊断;且对于系统的细小扰动,多尺度主元分析法的故障诊断能力优于主元分析法。 Based on wavelet analysis and principal component analysis (PCA) theory, aiming at the limitations of single-scale modeling in PCA, the ability of correlation of de-linear variables in PCA and the ability of wavelet transform to extract the local features of variables and the auto-correlation of approximate decomposition variables Form multiscale principal component analysis (MSPCA). It is applied to FPSO oil-gas-water separation system. The results show that the fault diagnosis method based on multi-scale principal component analysis can accurately detect and diagnose the early failure of the FPSO production process, and the multi-scale PCA method has better fault diagnosis ability than the main element Analysis.