A SELF-TRAINING CLASSIFICATION JUDGEMENT OPTIMIZATION METHOD FOR THE IMPACT-RESISTANT STRUCTURAL SIZE OPTIMIZATION

Impact-resistant structures have a broad application and play a crucial role in military and civilian safety. However, such structure optimization suffers from the time-consuming issue of the structure dynamic response analysis and finds sensitivity analysis difficult due to the strong non-linearity associated with the extreme loading. In this study, a self-training classification judgment optimization method is proposed. This method utilizes a self-training classification judgment surrogate model based on support vector machines and a genetic algorithm to solve the size optimization problem of impact-resistant structures. Different from the regression-based conventional surrogate models, the self-training classification judgment surrogate model reduces the computational cost of the sample dataset. Moreover, a constraint-handling strategy and a fitness calculation method are introduced to integrate the classification judgment surrogate model into the genetic algorithm. Two examples, including the optimization of the blast-resistant corrugated sandwich structures and the penetration-resistant polyurea/ceramic composite plate, are presented to showcase the effectiveness and efficiency of the proposed method. It is expected that the proposed novel method with high efficiency for impact-resistant structures optimization is capable of ensuring the service performance of equipment structures and personnel safety under extreme impact loadings.