Fixture Failure Diagnosis for Autobody Assembly Using Pattern Recognition

Abstract

In this paper, a fault diagnostic method is proposed for autobody assembly fixtures. This method uses measurement data to detect and isolate dimensional faults of part caused by fixture. The proposed method includes a predetermined variation pattern model and a fault mapping procedure. The variation pattern model is based on CAD information about the fixture geometry and location of the measurement points. This fault mapping procedure combines Principal Component Analysis with pattern recognition approach. Simulations and one case study illustrate the proposed method.