Participative Method to Identify Data-Driven Design Use Cases

Abstract

Paradigms such as smart factory and industry 4.0 enable the collection of data in enterprises. To enhance decision making in design, computational support that is driven by data seems to be beneficial. With this respect, an identification of data-driven use cases is needed. Still, the state of practice does not reflect the potential of data-driven design in engineering product development. With this respect, a method is proposed addressing the business and data understanding in industrial contexts and corresponding Product Lifecycle Management (PLM) environments. This allows to identify use cases for data-driven design taking into account business processes as well as the related data. In the proposed method, first the main process tasks are analyzed using a SIPOC analysis that is followed by a process decomposition to further detail and highlight corresponding applications using Enterprise Architecture principles. Following this, value stream mapping and design process failure mode effect analysis are used to identify sources of waste and the related causes. With this, a feature analysis of given data is proposed to identify use cases and enable to further use standard data science methods like CRISP-DM. The method is validated using the infrastructure of the Pilotfabrik at TU Vienna. The use case shows the applicability of the method to identify features that influence the cost of a product during the manufacturing without changing the functional specifications. The results highlight that different methods need to be combined to attain a comprehensive business and data understanding. Further, a comprehensive view of the processes is yielded that enables to further identify use cases for data-driven design. This work lays a foundation for future research with respect to data-driven design use cases identification in engineering product development. Keywords: Data-driven-design • PLM • Enterprise architecture • ArchiMate • SysML This work has been partially supported and funded by the Austrian Research Promotion Agency (FFG) via the "Austrian Competence Center for Digital Production" (CDP) under the contract number 854187.

A Survey on the Challenges Hindering the Application of Data Science, Digital Twins and Design Automation in Engineering Practice

“…Although these methods are generally applicable, domain-specific difficulties concerning data collection or processing are not explicitly depicted. Therefore, methods have been proposed to extend CRISP-DM for the applicability in engineering (Huber et al, 2019;Rädler and Rigger, 2020;Stanula et al, 2018) by better align with the requirements in the specific domain, e.g. extension of the data understanding to gather knowing about tribological experiments (Bitrus et al, 2020).…”

“…The focus is on the exchange of data between the physical product and the digital twin. (Grieves and Vickers, 2017) To foster the integration of digital engineering in design practice, methods have been proposed in the context of design automation (Curran et al, 2010;Zheng et al, 2012) and data science (Bitrus et al, 2020;Rädler and Rigger, 2020;Wiemer et al, 2019). Design automation studies show that the potential and the opportunities in industry are still not entirely reflected (Rigger and Vosgien, 2018).…”