Surrogate Model-based Acoustic Optimization Design for Underwater Cylindrical Shells

Jia-peng CHENG,De-qing YANG,Gui-lian YI
DOI: https://doi.org/10.3969/j.issn.1006-1335.2016.03.001
2016-01-01
Abstract:To improve the efficiency of underwater structure acoustic calculation, the FE software is used to analyze the sound radiation of an underwater cylindrical shell structure. An acoustic approximate agent model is established. Then, the analytical expression of the relation between the geometry size and underwater acoustic radiation characteristics of the shell structure is given to simplify the acoustic computation. An efficient structural acoustic optimization design model is established. The Latin Hypercubic Sampling method is utilized to select the sampling points. Using three types of surrogate models, polynomial response surface fitting, Kriging function and radial basis neural network, the acoustic radiation model of the double cylindrical shell structure is established for optimal design. Through comparison of the accuracy of three types of surrogate models, the appropriate surrogate models are chosen to construct the optimization model. The optimization model is solved by using Matlab optimization toolbox. As a result, the weight of the shell structure is reduced. This research provides references for predicting the acoustic radiation and performing the acoustic design of underwater structures.
What problem does this paper attempt to address?