Optimization of structural variables of a flextensional transducer by the statistical multiple regression analysis method

The performance of an acoustic transducer is determined by the effects of many structural variables, and in most cases the influences of these variables are not linearly independent of each other. To achieve optimal performance of an acoustic transducer, we must consider the cross-coupled effects of its structural variables. In this study, with the finite-element method, the variation of the operation frequency and sound pressure of a flextensional transducer in relation to its structural variables is analyzed. Through statistical multiple regression analysis of the results, functional forms of the operation frequency and sound pressure of the transducer in terms of the structural variables were derived, with which the optimal structure of the transducer was determined by means of a constrained optimization technique, the sequential quadratic programming method of Phenichny and Danilin. The proposed method can reflect all the cross-coupled effects of multiple structural variables, and can be extended to the design of general acoustic transducers.