基于非线性主成分分析的车身制造偏差源诊断

Fault diagnosis in auto-body assembly process based on non-linear principal component analysis

对主成分分析在轿车车身制造偏差源诊断方面的不足进行了分析,进而在其基础上提出采用非线性主成分分析,从而更有效、更准确地来对轿车车身制造偏差源进行诊断,并给出了计算实例。实例证明该方法优于主成分分析,并在工程上有较大的可用性。

This paper makes analysis of the shortages of Principal Component Analysis(PCA) in fault diagnosis of auto-body assembly process. Basing on this, another analysis method named as Non-Linear Principal Component Analysis(NLPCA) was proposed. The given actual case proves that Non-Linear PCA is better than PCA in searching root cause which brought in auto-body assembly process. And it has high usability in auto industry.