Knowledge graph for manufacturing cost estimation of gear shafts - a case study on the availability of product and manufacturing information in practice

Abstract

Growing cost pressure forces companies to actively manage their product costs to secure profitability. Here, manufacturing cost estimation within product development estimates manufacturing and material costs. As most products are developed in generations, needed product and manufacturing information can origin from reference system elements (RSE), for example similar components of prior product generations. Problematically, this product and manufacturing information as well as the knowledge of its interrelation is often stored in an unstructured way, document based or at least not machine-readable. This makes manufacturing cost estimation an effortful, time consuming and mainly manual activity with low traceability, where a wide manufacturing knowledge is required. Trends in production, like new manufacturing processes and production systems further increase the need for manufacturing information and knowledge. Knowledge graphs as semantic technologies can improve the findability and reusability of reference system elements and enable automatic information processing.Within this research, cost estimation of research and development of a large automotive supplier was used as research environment. Guided by the model of PGE an ontology for the manufacturing cost estimation domain was developed. Then, a knowledge graph was instantiated based on product and manufacturing information from gear shafts of electric axles. A case study was carried out to evaluate process-specific cycle time calculation as exemplary use case of the knowledge graph. Process-specific cycle times are generally effortful estimated based on detailed manufacturing information and then used together with machine hourly rates to estimate manufacturing costs. Here, the structured and machine-readable manufacturing information of identified reference system elements is extracted from the knowledge graph to reduce the effort, increase the traceability and enable future automation. The case study shows exemplary, how a knowledge graph can support manufacturing cost estimation of gear shafts where product and manufacturing information is automatically identified using reference system elements.