This paper is from Session 5D: Development of statistical thinking in the workplace

which comes under Topic 5: Statistics education in the disciplines and the workplace

Development of training methods to accelerate the competencies in Weibull analysis: case study in the automotive industry

Presenter

• Halimatou Ndiaye (Renault SA, France)

Co-authors

• Michael Huchette (UMR Sciences Techniques Education Formation, France)
• Paul Schimmerling (Renault SA, France)
• Sébastien Castric (Compiègne University of Technology, France)
• Zohra Cherfi (Compiègne University of Technology, France)

Abstract

In the automotive industry, statistical methods are well used in areas such as design, production or analysis of customer returns. However, real problems are often poorly formalized which creates special difficulties of implementation: choosing the appropriate method, selection of data for the model, physical and technical interpretation of results… Weibull analysis is a statistical method used to characterize the lifetime of a product during testing or in-service. It gives indication on failure mechanisms and the causes of product failures. It also predicts the expected number of failures based on real observed failure data. This paper focuses on the characterization of professional knowledge for Weibull analysis. It is based on analysis of technical notes written by engineers from RENAULT SA. The aim is to improve internal training. We present the context of the study, describe the implementation of the method and then explain how to use our results to improve training.