Introduction of decision support systems for failure management in manufacturing

Data-driven approaches are essential for optimizing failure management in manufacturing. However, there is no systematic approach that serves as a blueprint for data-driven failure management and thus data is predominantly used for custom-tailored and standalone solutions. Identifying and integrating cross-value chain data along the product lifecycle remains challenging. This complicates constructing comprehensive failure knowledge and prioritizing quality-centered measures.
This paper proposes a systematic approach to implementing data-enabled failure management within the value chain, requiring adapted organization and strategic alignment across all business levels. A guideline is presented for organizing data-enabled failure management. Firstly, an understanding of data-enabled failure management is established based on a reference structure. Secondly, structures and utilized data are analyzed. The paper examines how added value can be generated and utilized from existing information, as well as from AI-supported models, to derive optimization and failure prevention processes. Subsequently, implementation potentials and areas for improvement are identified. The focus in developing cross-value chain failure management is on achieving direct added value for the company throughout the build-up phase of establishing AI use cases. This methodology, initially developed for the commercial vehicle industry, is evaluated and refined through use cases involving field data, production process data, and configuration data. The presented results serve as a basis to further improve failure culture and understanding, as well as to introduce employee incentive programs in future research activities.
This research establishes a framework for achieving a cohesive decision support system for failure management. By outlining the organizational setup, including the definition of an idealistic representation on the strategic orientation, scope of the project and a status quo analysis, followed by the decision on project implementation and the implementation of the project itself in form of pre-defined use cases, this paper contributes to advancing integration of data-driven approaches in failure management for manufacturing.