【作者】 熊磊； 余鹏； 陈美霞；

【Author】 XIONG Lei;YU Peng;CHEN Meixia;Huazhong University of Science and Technology;

【机构】 华中科技大学船舶与海洋工程学院；

【摘要】 针对圆柱壳结构声辐射预报问题,提出单元辐射叠加法结合信息采集更加完整的QR分解改进的遗传算法布置测点。该预报方法首先通过无限长圆柱形活塞声源叠加得到声传递向量ATV,然后通过QR分解法筛选初始测点,并结合遗传算法选出最优测点组,最后由传感器数据计算获得振速,把ATV作为系数矩阵对振速加权求和得到辐射声压。通过与基本遗传算法和有限元的结果对比表明,声辐射预报精度较高,大大缩短了搜索算法所需的计算时间,是一种快速的预报方法。更多还原

【Abstract】 Point at the problem of acoustic radiation prediction of cylindrical shell structure,an improved genetic algorithm based on element radiation superposition and QR decomposition is proposed.Firstly,the sound transfer vector ATV is obtained by superposition of infinite cylindrical piston sound sources,then the initial measuring points are selected by QR decomposition method,and the optimal group of measuring points is selected by combining with genetic algorithm.Finally,the vibration velocity is calculated from sensor data,and the radiation sound pressure is obtained by weighted summation of vibration velocity with ATV as coefficient matrix.By comparing with the results of basic genetic algorithm and boundary element method,it is shown that the prediction accuracy of acoustic radiation is higher,and the calculation time of search algorithm is greatly reduced,so it is a fast prediction method.更多还原