A generalised collaborative optimisation method and its combination with kriging metamodels for engineering design

摘要

In this paper, a generalised collaborative optimisation (GCO) method is proposed in the field of multidisciplinary design optimisation (MDO). In this method, the concept of generalised multidisciplinary variables is put forward and summarised as four types to provide more comprehensive interdisciplinary compatibility constraints than traditional collaborative optimisation. The linear normalisation method is adopted in GCO to remove inconsistencies caused by the range discrepancies between design variables. The effectiveness of GCO is demonstrated by its successful application in the design of a gear reducer. Furthermore, to undertake the computational and organisational challenges in MDO, an effective approach combining kriging and GCO (KGCO) is presented. In this approach, kriging metamodels are employed to reduce the computational cost by replacing expensive simulation and analysis codes. System-level compatibility constraints are approximated by kriging metamodels to remove the non-smoothness of the original quadratic constraint forms. The feasibility and computing efficiency of the KGCO approach are illustrated by solving the gear reducer design problem. Through its innovative and successful application in the design of a small waterplane area twin hull ship, it is further demonstrated that the KGCO approach is a feasible and efficient design methodology in support of the design of engineering products.