A Knowledge Based Engineering approach to automation of conceptual design option selection
E.J. Schut, Michel J.L. van Tooren · 45th AIAA Aerospace Sciences Meeting and Exhibit · 2007
*† The Multidisciplinary Design and Optimization (MDO) process can be supported by partial automation of analysis and optimization steps. Design and Engineering Engines (DEEs) are a useful concept to structure this automation. Within a DEE a product is parametrically defined using Knowledge Based Engineering (KBE). This parametric product model requires an initial design solution before multidisciplinary optimization can be performed. The Initiator component of the DEE is responsible for finding this initial design solution. However, at the start of the design process the designer is faced with too many requirements, too many design properties, a too complex model, and too little resources. By using a simplified design problem the designer is able to find an initial design solution. Via iteratively adjusting the level of simplification the designer tries each time to find a feasible solution, in order to eventually find the initial design solution. This process highly iterative, requiring the same design knowledge to be accessed multiple times. To support accessibility of design knowledge surrogate models can be used that capture the optimized feasible solutions, representing the design space. At the same model aggregation level design options have identical objectives, based on which the optimal design option can be selected automatically. An implementation of the Initiator, featuring this methodology, is used in a sample DEE for aircraft vertical tail design. The selection process is implemented for panel structural design, based on four design options.