This work brings a contribution to the improvement of the product development process in the engineering domain. The overall goal of the design process is to realize in a short time and with minimal costs a high quality design solution that satisfies all customer requirements. In order to achieve this goal and to master the product complexity, a divide-and-conquer approach is applied: The design task is split in several development branches, each of them being concerned with only one aspect of the product. At different stages of the design process, the design models resulting from the different development branches are synchronized with each other in order to ensure the consistency of the overall design. However, checking the consistency of a design model and between different design models is a very challenging task. One of the reasons is that the engineering tools employed by the development branches operate on models that have different conceptualization of a product according to their own viewpoint. For instance, a computer-aided design (CAD) tool will model the geometrical characteristics of a product, while a simulation tool will model the behavior of the product. Although the two design models represent different viewpoints on the product, they must be consistent with each other, if a common implementation (i.e., the product) of the two design models should be realized.
Besides the consistency checking of design models, the propagation of design changes from one model to another is also important. These two tasks are very frequently done in the design process, and ideally they should be automated. However, there are several hindrances to an automation of these tasks. First, the engineering tools do not typically interoperate. Exchanging model information between engineering tools is often done manually, is error-prone and difficult, considering the size and complexity of the design models. Second, the design models are represented using different modeling languages with their own syntax and modeling methodology. Third, the design models operate with different conceptualization of the product according to their own view on the product. Very often many aspects of the design model remain implicit and are known only to the engineers who have actually built the model. Forth, there are no explicit or formal correspondences between the design models that could be used to asses the consistency between the models and to propagate the design changes.
The contribution that this work brings to the improvement of the design process is two fold. First, it provides a framework for building higher quality design models. Second, it provides a framework that supports the automation of the consistency checking between the design models and for change propagation. The two goals are achieved by using a formal approach to modeling engineering systems based on ontologies. Ontologies are formal descriptions of the objects, of their properties and relationships in a certain domain. Ontologies also contain axioms that make domain assumption explicit both for humans and computers. Tools that commit to using the same ontology have a better chance to achieve interoperability and consistency between their models. As part of this work, an engineering ontology has been developed for representing engineering systems. The engineering ontology has been used to enrich the representation of the design models with domain knowledge. Another ontology has been developed for representing the requirements of a product in a formal way. Logical reasoning on the enriched design models has been used to check the fulfillment of the requirements. The second goal was achieved by defining in a formal way the consistency between different design models. This is supported by a mechanism for representing the correspondences (or mappings) between concepts in the design models. The mapping mechanism can be used to check the consistency of the design models or to propagate changes in one design model to another.
The concepts in this work have been implemented in two prototypes that have been used in three scenarios in the design process of the automotive domain. Employing ontologies to make domain assumptions explicit and for enriching design models with domain knowledge has proved to enhance the quality of the models, to enable the reuse of design models and to support the interoperation of the engineering tools.
