A Light-Weight Design Method for Vehicle with Welded Structure Based on Surrogate Model

Abstract

In this paper, a multi-disciplinary optimization frame considering weld fatigue constraints is constructed based on kriging surrogate model in order to improve the fatigue reliability of vehicle structures. Under this frame, kriging surrogate model is used to construct weight, strength, stiffness, frequency, fatigue life responses for optimization, which improves effectively the computing efficiency. For weld fatigue analysis, master S-N curve method in AMSE criterion is used for reliable fatigue life. The curb girders of a high-speed train are taken as an example, a light-weight optimization model considering static strengthen and weld fatigue constrains is constructed and solved by use of submodelling technology under the proposed optimization frame. After optimization, the weight of the curb girders is reduced about 33% from 1.73 ton to 1.16 ton. These results show the proposed optimization method is effective, and provide the preference for the light-weight design of high-speed trains.