Topology Optimization for Acoustic-Structure Interaction Systems Using the Model Reduction Method

Purpose This paper develops a vibro-acoustic component modal synthesis method to solve the problem of high computation time in solving the eigenvalues and eigenvectors of coupled equations for noise reduction topology optimization. Methods The normalization condition of the left and right eigenvectors of the unsymmetrical vibro-acoustic equation is introduced, and Craig-Bampton method is extended to the structure-acoustic coupled problem, which reduces the time consumed by the response calculation part. The results show that the eigenvalues, the eigenvectors, and the response values calculated by the proposed method match well with the original model. To solve the problem of spurious eigenmodes caused by the traditional penalization model used in the classical bi-directional evolutionary optimization (BESO) method, the linear stiffness penalization model is introduced into the BESO method, and an improved penalization model is proposed. Conclusions Numerical examples of minimizing the response at a specified target point are studied. The results show that the iterative process is robust. Through these numerical design examples, the effectiveness and reliability of the model reduction method and the improved penalization model are verified.