Knowledge-driven intelligent quality problem-solving system in the automotive industry. (October 2018)

• Record Type: Journal Article
• Authors: Xu, Zhaoguang; Dang, Yanzhong; Munro, Peter
• Is Part Of: Advanced engineering informatics. Volume 38(2018)
• Page Start: 441
• Page End: 457
• Publication Date: 2018-10
• Subjects: Quality management -- Intelligent quality problem-solving -- Knowledge management -- Automotive industry -- Digital fishbone diagram -- Ontology; Computer-aided engineering -- Periodicals; Engineering -- Data processing -- Periodicals; 620.00285
• DOI: 10.1016/j.aei.2018.08.013
• Languages: English
• ISSNs: 1474-0346

Abstract

Highlights: To obtain knowledge from automotive quality problem-solving data through data mining. To extract the relationship matrix between the components and faults. Ontology library provides a common language between different departments. Intelligent Quality Problem Solving System improves the efficiency of problem-solving. Digital Fishbone Diagram reduces problem analysis time and costs. Abstract: In the current automotive industry, quality management, especially quality problem-solving (QPS), plays an important role in fulfilling the expectations of demanding customers who seek high-quality products at low-cost. During the problem-solving process, various real-time and historical quality data are often not fully used, yet these data are of high value. This paper provides a comprehensive quality data mining process and method, as well as an intelligent quality problem-solving system (IQPSS). First, based on original quality problem data, an ontology library is constructed using the ontology generating module (OGM). Second, based on the generated ontology and the textual data of the original quality problem, this study builds a quality problem-solving knowledge base (QPSKB) by employing relevant algorithms in the knowledge transformation module (KTM). The component and fault relational matrix mining (CFRMM) algorithm is designed to extract the relationship matrix between the components and faults. The semi-supervised classification algorithm based on the K-nearest neighbor algorithm (KNN) is used to classify the immediate measures, causes and long-term measures into the corresponding ontology and express the ontology as their knowledge. Furthermore, the binary tree-based support vector machine (SVM) approach is applied to classify the cause texts into the factors of Man, Machine, Material, Method, and Environment (4M1E), which are the five factors in a fishbone diagram. In particular, the digital fishbone diagram is a brand-new type of fishbone diagram that subverts the traditional method of fishbone diagram analysis through brainstorming. A pilot run of the IQPSS has been undertaken in an automotive manufacturing company to demonstrate how quality management employees obtain this knowledge by searching in the IQPSS. The results show that the IQPSS contributes appreciably to the quality problem-solving in the manufacturing industry.