水声宽带信号波形预报的并行算法

摘要

针对利用频域方法实现水声宽带信号波形预报过程中计算量较大的问题,引入频域宽带近似展开模型,研究了适合该宽带模型特点的并行算法,将串行模型并行化,并对宽带信号波形并行算法的可行性和有效性进行了分析,通过仿真分析利用模型的并行算法,在典型浅海负跃层和深海声道环境下,波形预报速度分别提高了26倍和24.6倍,计算效率都在75％以上,有效的实现了水下宽带声场的快速准确预报,为实现大规模信号级仿真奠定基础.%In order to solve the problem of computation amount in underwater broad-band signal prediction, the paper used the model of frequency approximate expansion and researched parallel computation algorithm of serial broadband model according to model characteristic. Then feasibility and effectiveness of parallel arithmetic were analyzed and testified. Under typical ocean environment including shallow water with a thermocline and deep sound channel, making use of parallel computation arithmetic, the computation speed is increased 26 times and 24.6 times respectively , and the efficiency is above 75% , which predicts broad-band signal quickly and exactly.