Transducer Position Optimization for Acoustic Modal Identification of Thin Annular Duct

Transducer position optimization is a classic combinatorial optimization problem. In acoustics, the layout of transducers affects the accuracy of acoustic inverse problems. This paper investigates the optimal transducer layout for acoustic modal identification by using the particle swarm optimization (PSO) algorithm in a thin annular duct, with the condition number as the optimization criterion. As the number of transducers increased, PSO exhibited convergence difficulties due to the high particle dimension. To address this issue, the optimization algorithm was improved by introducing a prior assumption for the optimal layout and relaxing the fitness function, transforming the high-dimensional optimization problem into parallel low-dimensional ones. Additionally, relaxing the fitness function helped to enhance its sensitivity to minor positional changes, thereby improving the optimization effect of the algorithm. Through numerical experiments, the improved algorithm was verified to converge to a lower condition number under lower particle swarm sizes and fewer iterations, and the running time of the algorithm was significantly reduced. Furthermore, this study applied the improved algorithm to search for the optimal transducer positions under different physical parameters and analyzed the influence of each physical parameter on the distribution and corresponding condition number of the optimized layout.