A System Approach for Creating Employee-Oriented Quality Control Loops in Production for Smart Failure Management System in SMEs

Abstract

The basis of effective processes for failure elimination and prevention is formed by quality control loops, which aim to detect and analyse deviations/failures from quality specifications and to take appropriate measures to correct them in time. Quality methods can either be used as controllers (e.g. statistical process control, failure mode and effects analysis) in quality control loops or as a special form of these (e.g. Plan-Do-Check-Act cycle, 8D Report, Six Sigma methodology). For the successful introduction of quality control loops, relevant data should be systematically collected, evaluated and interpreted in order to derive targeted measures as a consequence. Small and Medium Enterprises (SME) rarely have a systematic quality management system for the comprehensive collection and analysis of quality data and their embedding in quality control loops. On the other hand, the increasing complexity of production systems requires the digitalisation and expansion of quality control loops already in use, although they have delivered good results so far. At this point, Artificial Intelligence (AI) is a future key technology that holds significant potential for future value creation. AI-driven data science methods (e.g. machine learning) enable the explanation of complex, correlationally directed relationships in large amounts of data and accordingly contribute to process improvement as well as failure management. In this context, the expansion of quality control loops through digitalised elements and AI methods can help to achieve a smart failure management system.In terms of content, failure management also includes the term "failure prevention" and is not a one-time process, but a continuous process that requires the motivation and understanding of all employees. Furthermore, the concept of "Total Productive Management" also aims at defect-free products and effective production processes and involves all employees in improvement activities to maximise plant efficiency and minimise losses. At this point, SMEs need intelligent, digital and employee-oriented error management systems. The core objective of the paper is to present the conceptual development of a smart failure management system that is in a continuous learning process through interaction with the employee and in this way learns human cognitive problem-solving skills. This approach is intended to detect failures on the shop floor at an early stage in order to identify possible causes of problems and derive measures. If the defect type, cause or measure are not known, the system suggests suitable methods/tools of quality and data science to support employees in problem solving process. In order for the assistance system to have human cognitive problem-solving capabilities, the system must be trained in advance by qualified employees who have extensive technical and methodological knowledge and can apply it confidently. With this in mind, the failure management system is expanded to include two additional subsystems. Firstly, the less competent employees must be taught missing methodological knowledge on the basis of digital learning methods. Then, the learning phase is followed by the employee training, in which the employee is supported digitally and dialogue-based in the selection and implementation of methods as well as the interpretation of results.