摘要
根据不同浅海环境参数在不同的频段对匹配场功率具有不同贡献的特点 ,提出了宽带匹配场反演的多步优化策略 ,并与直接全参数反演进行了实测数据的性能对比。采用多步反演策略对亚洲海实验 ( ASIAEX2 0 0 1 )数据进行处理 ,按照不同海区水文条件的差异 ,分别进行了水平不变和水平变化的环境参数反演 ,得到了东中国海
Inversion by MFP (Matched Field Processing) can be considered as a complex non linear optimization problem. The accuracy with which a parameter can be inverted effectively depends mainly on the extent of its influence on the acoustic field. Sensitive parameters can be estimated accurately and quickly as compared with less sensitive parameters. In order to get satisfactory inversion results of environmental parameters from ASIAEX2001 (Asian Seas International Acoustics Experiment) in the ECS (East China Sea) in June 2001, we studied the sensitivity of the inverted parameters including geometric and geoacoustic parameters at different frequencies. Fig.3 (a) gives the sensitivity indices of five geometric parameters--1.range, 2.water depth, 3.source depth, 4.receiver depth, 5.array tilt--at four typical frequencies--40Hz, 210Hz, 470Hz, 670 Hz. Fig.3 (b) gives sensitivity indices of 4 sediment parameters --1.thickness, 2.attenuation, 3.density, 4.sound speed--at the same four typical frequencies. Fig.3(c) gives sensitivity indices of 3 bottom parameters --1.sound speed, 2.attenuation, 3.density--also at the same four typical frequencies. Taking into consideration the fact that different parameters have different sensitivitiy indices on MFI (Matched Field Inversion) power, we propose a new approach named MS MFI (Multi Step MFI). The MS MFI scheme presented in Fig.4 is, in our opinion, ideal for complex geoacoustic models involving large number of unknowns. We divide the total parameter space into a few subspaces and perform inversion in each subspace. Firstly, in inverting the more sensitive parameters, we use high frequency data. Secondly, in inverting the less sensitive parameters, we use one or two high and several low frequency data. Finally, in inverting the least sensitive parameters, we use low frequency data. If necessary, we execute these three steps iteratively until we get suitable values. Subsection 1.2 gives four advantages for MS MFI as compared with DAP MFI (Direct All Parameters MFI). Using multi
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2003年第5期611-615,共5页
Journal of Northwestern Polytechnical University
基金
国家自然科学基金 (6 0 0 72 0 2 5 )
西北工业大学"英才培养计划"基金资助
关键词
宽带匹配场反演
亚洲海实验
多步反演策略
MFP (Matched Field Processing), ASIAEX (Asian Seas International Acoustic Experiments), DAP MFI (Direct All Parameters Matched Field Inversion), MS MFI (Multi Step Matched Field Inversion)
