Identifying nonconformity root causes using applied knowledge discovery

Abstract

Quality control, failure analysis and improvement are central elements in manufacturing. Total Quality Management (TQM) provides several quality oriented tools and techniques which, in the event of things, are not always applicable. The increased use of Information Technology (IT) in manufacturing means increased data availability and improved potential for knowledge extraction. Exploiting this knowledge requires data storage and processing facilities with demanding, time consuming sessions for interpretation. Without suitable tools and techniques, knowledge remains hidden in databases. This paper presents a method to help identify root causes of nonconformities (NCs) using a pattern identification approach. Hereby, a general framework, Knowledge Discovery in Databases (KDD), is adapted. This adaptation involves incorporating an economic concentration measure, the Herfindahl-Hirschman Index (HHI), as the data mining algorithm. After presenting the theoretical background, a new methodology is proposed. The suggested approach can be regarded as a quality tool to help make root cause identification of failures simpler and more agile. A case study from the automotive industry is examined using this tool. Results are obtained and presented in the form of matrix based patterns. They suggest that concentration indices help indicate possible root causes of NCs, warranting further investigation in this area.