How to build up an engineering change dependency model based on past change data?

Coping with engineering changes (EC) is a challenge in engineering design of complex products, since changes prone to propagate and produce further changes on components and processes. This is due to the high connectivity of components in complex systems. There has been a lot of research regarding methods based on product structure models (e.g. in form of DSMs) to predict and assess the propagation of engineering changes. These models are normally generated in interviews with experts, which estimate the propagation of ECs on component-level. Thus, the procedure is very time-consuming and methods are often not profitable applicable. This paper aims to present an approach of how an EC dependency model can be generated with less effort by applying the MDM methodology combined with data mining techniques by using commonly available EC data in industry. Data mining techniques enable the extraction and quantification of the dependencies in the model.