Dynamic optimization design of thin plate structure based on surrogate model

Abstract To enhance dynamic optimization efficiency in traditional thin plate structures, surrogate models are integrated for structural dynamic size optimization, reducing dependence on high-precision finite element simulations. This paper focuses on optimizing the aluminum alloy sheet as the subject of research. The parameters of the specimen are optimized through sensitivity analysis, based on the comparison between the free mode Finite Element Analysis (FEA) and Experimental Modal Analysis (EMA) results. The modified specimen undergoes validation under a four-edge clamped boundary condition. Subsequently, random vibration analysis is conducted to determine the root mean square (RMS) value of the acceleration response. Employing surrogate model technology, four models are constructed using parametric modeling and experimental design methods. These models are then compared for their fitting effectiveness. The results indicate that the RBF model exhibits the highest accuracy in fitting each response, effectively replacing dynamic simulation. Finally, the RBF combined with the NSGA-II method, optimizes the specimen’s structural size, resulting in a 21.6% reduction in mass and an 18.60% decrease in acceleration response. These outcomes demonstrate satisfactory optimization results, ensuring accuracy while enhancing the dynamic characteristics of the specimen structure. This research offers valuable insights for related engineering applications.