The L-Chart for Non-Normal Processes

Abstract

This paper presents a proposal to construct control charts for non-normal processes and an application. This control chart is based on the least-square L-estimator, which replaces the average and the standard deviation usually calculated for Shewhart charts. This estimator has a minimum variance for estimation of the process position and scattering whatever the data distribution. We focus our attention on “multi-generators” processes, like screw-machines or multi-die holder for injection molding, these processes are non-normally distributed.

Key Words:

• Control chart
• injection press
• L-statistics
• non-normal process
• statistical process control

Additional information

Notes on contributors

Emmanuel Duclos

Emmanuel Duelos He is an expert in the field of Statistical Process Control. He is consulting for different companies.

Pillet Maurice

Maurice Pillet He is Professor in University of Savoy in the Department of Industrial Engineering, laboratory LISTIC. He is a member of ASQ and RUFEREQ. His research works concern Quality and industrial performance.

Laetitia Avrillon

Laetitia Avrillon He is a PhD student at the University of Savoy (France). Her main area of research is quality engineering in the high tech industry and more particularly problemsolving methods. Her research work is carried out in collaboration with Trixell, a joint-venture company (Thales Electron Devices, Philips Medical Systems and Siemens Medical Solutions).