Abstract: It is discussed that when a mechanical product is under innovation, the changes in its working background and the varying market requirements have critical impacts on the conceptual design process. Characterizing this kind of innovation process, a dynamic model of E-R-F-B-S (environment, requirement, function, behaviour and structure) is put forward, which is featured by the directed mappings between the five kinds of variable sets. Named innovation chains, four kinds of innovation patterns of the mechanical products depicting the designers’ usual trains of creative thoughts in the design activities are outlined, represented by the directed combination of the seven mapping units in the E-R-F-B-S model. The mathematical model of innovation chain is established by determining the correspondence between innovation chain and directed graph, then directed graph and sparse matrix, thus making it possible to suggest the setting up of a computer-aided reasoning system for conceptual design, sparse matrices acting as its input variables. A simple case study is presented to illustrate the proposed approach.

Key words: Conceptual design, Directed graph, Innovation chain, Mapping between variable sets, Product innovation, Sparse matrix, ultrasonic assisted cutting, shape characteristic, Nomex honeycomb composite material