Mapping Criteria for Designing Machine Learning-Driven Integrated Quality-Maintenance Engineering Frameworks

Other conference contribution, 2026

Increasing global competition and the growing complexity of modern manufacturing systems demand operations that are quality-driven, resilient, and sustainable. Quality Engineering (QE) directly determines product reliability, customer satisfaction, and organisational financial viability, while Maintenance Engineering (ME) ensures equipment availability, prevents unplanned downtime, and sustains the conditions under which quality targets are met. However, both functions have traditionally operated in silos, managing separate data streams, KPIs, and decisions, despite defects and failures often sharing common root causes. Machine Learning (ML) has been identified as the bridging technology to achieve meaningful integration, enabling real-time analysis of heterogeneous data sources across both domains and creating conditions for a QE-ME unified approach to defect prevention, failure prognosis, and continuous improvement. Yet, no consolidated reference exists in the scientific literature that maps which components are necessary for such QE-ME integration. This paper addresses this gap through a literature review and structured comparative analysis of data-driven QE and ME frameworks, using comparison matrices. The objective is to identify and compare their key components to determine commonalities and differences towards their convergence in an “ML-driven Integrated Quality-Maintenance Engineering Reference Framework”. For industry practitioners, the findings offer actionable guidelines on which components to prioritise when designing such kind of frameworks. For researchers, they provide a structured evidence-based map of the field that surfaces the existing research gap and points to the most critical open challenges at the ML-driven QE-ME intersection.