Design for uncertainty

The development of engineering systems, whether those systems be products, plant or services, is a complex activity. The systems themselves are becoming more rather than less complex. Partly this arises from the need to optimise multiple system attributes including function, cost, resource usage, and style, among others. Also, most significant engineering projects involve diverse activities (including design, risk assessment, decision making, construction, and ongoing operation), and these invariably require teams of people with different skills. In turn that diversity requires effective management and leadership. Furthermore, the addition of electronic functionality and active control systems adds functionality but also introduces complex failure modes. Currently the development of complex engineering systems is typically managed by decomposing the project according to the bodies of knowledge involved, i.e. according to engineering design, manufacturing/construction, project management, management theory, etc. Consequently, there exist separate professional bodies of knowledge that interface with each other in limited ways. In particular, design theory, risk management, decision theory, and organisational behaviour are all important factors in determining the success of an engineering system, but relate together only weakly. This paper presents the development of a descriptive meta-model that seeks to integrate multiple separate bodies of knowledge, namely design theory, risk management, decision theory, and organisational behaviour, into one consistent epistemology. This paper presents results for a subset of the larger model. It focusses on key activities in the design process, namely how the design objectives (needs) are determined, how early design concepts are formed, and how uncertainties are treated during design. Results are presented for strengths and weaknesses of the constituent methodologies, and issues are identified for further study. Also, the features of a novel methodology, called DSI, for coping with design uncertainties are described and illustrated with data from a domestic dishwasher design