Multi-Model Dashboard : making interdisciplinary dependencies explicit in complex systems engineering environments

Abstract

Nowadays, systems engineering is an important discipline to manage the complexity of multidisciplinary engineering projects. A typical example of such a project is the development of industrial production plants, such as power plants or steelworks. Engineers from different disciplines, e.g. mechanical engineering, electrical engineering, building engineering, and software engineering, have to collaborate to succeed in their creation. But, when engineers from different fields work together, one can often observe the lack of a holistic view on the system. One reason for this is that engineers from one domain have none or only limited understanding of the models stemming from other domains. The engineers tend to develop domain-specific models in isolation, unaware of the dependencies to other domain¿s models. However, for the success of the overall project, it is necessary to define design constraints based on important parameters stemming from the models of several disciplines. If important design parameters change in a way that conflicts design constraints, mutually agreed upon with engineers of a partner discipline, then the timely notification of the partners is essential to mitigate the risk of costly design rework later in the project. Currently, a common approach is to manually check the design constraints either during a few predefined milestones, which often includes a delay before problems are discovered, or, which is even worse, at times when incidents already happened. In this thesis ways are explored to improve upon this situation. The key points are a theoretical concept, called the Multi-Model Dashboard (MMD) approach, and a research prototype implementation of the tool support for this approach. The goal of the MMD approach is to define a methodology on how to identify, specify, design and monitor multi-model parameters and design constraints. The approach is designed as a process model according to the requirements gathered from industry partners and the results of a systematic literature survey. It provides a framework and guidelines for engineers that seek a structured way to deal with interdisciplinary parameters and design constraints. The second core contribution of this thesis, is a research prototype that is designed complementary to the theoretical MMD approach. By iterative prototyping, the tool support was implemented as a proof-of-concept for the approach and as a reference implementation for interested parties. The thesis contains an in depth description of how the prototype¿s components are implemented and how interaction between them is realized. The MMD components are then evaluated by means of real-life use cases based in the systems engineering domain. Key points of the evaluation are a feasibility study and a cost-benefit assessment. The thesis is concluded by a discussion of the results, which identifies strengths and limitations of the MMD.