Fresnel层析成像并行算法研究
|
摘要
与射线层析成像相比,Fresnel层析成像考虑波频率的影响,具有较高的分辨率,但所需的存储空间和计算量更大,因此提出了Fresnel层析成像的并行算法.把大型层析反演方程组的求解,转化成对其中的各个方程进行相互独立的计算,避免了大型系数矩阵的存储问题;把一个Fresnel带的正演和反演计算放在一个进程,不同Fresnel带的计算相互独立进行,不需要信息传递,达到了极高的并行度;从进程之间没有通信,仅当从进程计算结束后,在主进程与各从进程之间有少量的数据传递,使通信开销达到了极小的程度.应用MPI在LinuxPC集群环境下实现了该算法,实际测试表明,该算法具有较高的并行度和加速比.
In contrast with ray-based traveltime tomography,Fresnel tomography accounts for the band-limited nature of seismic waves and gives the higher resolution tomograms.Because Fresnel tomography demands much computer memory and much running time,a parallel algorithm for it is proposed.The tomographic inversion is transformed to resolving respectively a series of single equation in light of backprojection principle,each equation corresponding to a Fresnel zone.The forward and inverse computation concerning a Fresnel zone is allocated to one process and is independent of other processes.Then the storage and calculation of the large-scale matrix in the tomography are avoided.No message delivers between the slave processes,and only a little of data delivers between a master process and the slave ones.By using the portable message passing interface standard(MPI) for the communication,the computing code of the algorithm is implemented on Linux system,which allows to distribute the work on several PCs connected via standard Ethernet in an in-house network,and greatly expands the applicability of Fresnel tomography.The tests on the synthetic and observed seismic travel time data show that this parallel algorithm has a good performance on Linux PCs.
In contrast with ray-based traveltime tomography,Fresnel tomography accounts for the band-limited nature of seismic waves and gives the higher resolution tomograms.Because Fresnel tomography demands much computer memory and much running time,a parallel algorithm for it is proposed.The tomographic inversion is transformed to resolving respectively a series of single equation in light of backprojection principle,each equation corresponding to a Fresnel zone.The forward and inverse computation concerning a Fresnel zone is allocated to one process and is independent of other processes.Then the storage and calculation of the large-scale matrix in the tomography are avoided.No message delivers between the slave processes,and only a little of data delivers between a master process and the slave ones.By using the portable message passing interface standard(MPI) for the communication,the computing code of the algorithm is implemented on Linux system,which allows to distribute the work on several PCs connected via standard Ethernet in an in-house network,and greatly expands the applicability of Fresnel tomography.The tests on the synthetic and observed seismic travel time data show that this parallel algorithm has a good performance on Linux PCs.
引文
[1]De Nicola S,Fedele R,Man’ko M A,et al·Fresnel tomography:A novel approach to wave-function reconstruction based on the Fresnel representation of tomograms[J]·Theoretical and Mathematical Physics,2005,144(2):1206-1213
[2]J Sheng,G T Schuster·Finite-frequency resolution li mits of wave path travelti me tomography for smoothly varying velocity models[J]·Geophysical Journal International,2003,152(3):669-676
[3]T CJohnson,P S Routh,MD Knoll·Fresnel volume georadar attenuation-difference tomography[J]·Geophysical Journal International,2005,162(1):9-24
[4]J Spetzler,R Snieder·The Fresnel volume and transmitted waves[J]·Geophysics,2004,69(3):653-663
[5]T CJohnson,P S Routh,WBarrash,et al·Afield comparison of Fresnel zone and ray-based GPR attenuation-difference tomography for ti me-lapse i maging of electrically anomalous tracer or contaminant plumes[J]·Geophysics,2007,72(2):G21-G29
[6]Luo Shengxian,Li Luming·Parallel computation collective communication based on MPI andits application[J]·Computer Applications,2003,23(6):51-53(in Chinese)(罗省贤,李录明·基于MPI的并行计算集群通信及应用[J]·计算机应用,2003,23(6):51-53)
[7]Liu Jinsong,Liu Futian,Liu Jun,et al·The opti mization and parallelization of sequential orthogonal triangularization algorithms used in seismic tomography[J]·Progress in Geophysics,2005,20(4):911-915(in Chinese)(刘劲松,刘福田,刘俊,等·天然地震走时反演矩阵顺序三角化算法的优化和并行化[J]·地球物理学进展,2005,20(4):911-915)
[8]Liu Jinsong,Liu Futian,Liu Jun,et al·The parallelization of LSQR algorithms used in seismic tomography[J]·Super Computation Letter,2005,3(2):5-11(in Chinese)(刘劲松,刘福田,刘俊,等·地震层析成像LSQR算法的并行化[J]·超级计算通信,2005,3(2):5-11)
[9]D W Vasco,J E Peterson,E L Jr Majer·Beyond ray tomography:Wavepaths and Fresnel volumes[J]·Geophysics,1995,60(6):1790-1804
[10]Wang Hongwei·Travelti me seismic tomography with Fresnel volume ray tracing[J]·Oil Geophysical Prospecting,1996,31(3):309-326(in Chinese)(王宏伟·菲涅耳体层析成像[J]·石油地球物理勘探,1996,31(3):309-326)
[11]Zhang Jianzhong,Chen Shijun,Xu Chuwei·A method of shortest path raytracing with dynamic networks[J]·Chinese Journal of Geophysics,2004,47(5):899-904(in Chinese)(张建中,陈世军,徐初伟·动态网络最短路径射线追踪[J]·地球物理学报,2004,47(5):899-904)
[12]Zhang Jianzhong,Chen Shijun·Numerical modeling of seismic first break in complex media[J]·Chinese Journal of Computational Physics,2003,20(5):429-433(in Chinese)(张建中,陈世军·复杂介质地震初至波数值模拟[J]·计算物理,2003,20(5):429-433)
[13]J A Hole·Nonlinear high-resolution three-di mensional seismic travel ti me tomography[J]·Journal of Geophysical Research,1992,97(B5):6553-6562
[14]Zhang Linbo,Chi Xuebin,Mo Zeyao,et al·Introduction to Parallel Computing[M]·Beijing:Tsinghua University Press,2006·210-215(in Chinese)(张林波,迟学斌,莫则尧,等·并行计算导论[M]·清华大学出版社,2006·210-215)
[2]J Sheng,G T Schuster·Finite-frequency resolution li mits of wave path travelti me tomography for smoothly varying velocity models[J]·Geophysical Journal International,2003,152(3):669-676
[3]T CJohnson,P S Routh,MD Knoll·Fresnel volume georadar attenuation-difference tomography[J]·Geophysical Journal International,2005,162(1):9-24
[4]J Spetzler,R Snieder·The Fresnel volume and transmitted waves[J]·Geophysics,2004,69(3):653-663
[5]T CJohnson,P S Routh,WBarrash,et al·Afield comparison of Fresnel zone and ray-based GPR attenuation-difference tomography for ti me-lapse i maging of electrically anomalous tracer or contaminant plumes[J]·Geophysics,2007,72(2):G21-G29
[6]Luo Shengxian,Li Luming·Parallel computation collective communication based on MPI andits application[J]·Computer Applications,2003,23(6):51-53(in Chinese)(罗省贤,李录明·基于MPI的并行计算集群通信及应用[J]·计算机应用,2003,23(6):51-53)
[7]Liu Jinsong,Liu Futian,Liu Jun,et al·The opti mization and parallelization of sequential orthogonal triangularization algorithms used in seismic tomography[J]·Progress in Geophysics,2005,20(4):911-915(in Chinese)(刘劲松,刘福田,刘俊,等·天然地震走时反演矩阵顺序三角化算法的优化和并行化[J]·地球物理学进展,2005,20(4):911-915)
[8]Liu Jinsong,Liu Futian,Liu Jun,et al·The parallelization of LSQR algorithms used in seismic tomography[J]·Super Computation Letter,2005,3(2):5-11(in Chinese)(刘劲松,刘福田,刘俊,等·地震层析成像LSQR算法的并行化[J]·超级计算通信,2005,3(2):5-11)
[9]D W Vasco,J E Peterson,E L Jr Majer·Beyond ray tomography:Wavepaths and Fresnel volumes[J]·Geophysics,1995,60(6):1790-1804
[10]Wang Hongwei·Travelti me seismic tomography with Fresnel volume ray tracing[J]·Oil Geophysical Prospecting,1996,31(3):309-326(in Chinese)(王宏伟·菲涅耳体层析成像[J]·石油地球物理勘探,1996,31(3):309-326)
[11]Zhang Jianzhong,Chen Shijun,Xu Chuwei·A method of shortest path raytracing with dynamic networks[J]·Chinese Journal of Geophysics,2004,47(5):899-904(in Chinese)(张建中,陈世军,徐初伟·动态网络最短路径射线追踪[J]·地球物理学报,2004,47(5):899-904)
[12]Zhang Jianzhong,Chen Shijun·Numerical modeling of seismic first break in complex media[J]·Chinese Journal of Computational Physics,2003,20(5):429-433(in Chinese)(张建中,陈世军·复杂介质地震初至波数值模拟[J]·计算物理,2003,20(5):429-433)
[13]J A Hole·Nonlinear high-resolution three-di mensional seismic travel ti me tomography[J]·Journal of Geophysical Research,1992,97(B5):6553-6562
[14]Zhang Linbo,Chi Xuebin,Mo Zeyao,et al·Introduction to Parallel Computing[M]·Beijing:Tsinghua University Press,2006·210-215(in Chinese)(张林波,迟学斌,莫则尧,等·并行计算导论[M]·清华大学出版社,2006·210-215)
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心