全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

广义线性纵横波联合反演的gpu加速

Keywords: 纵横波联合反演,稀疏脉冲,广义线性反演,gpu,计算效率

Full-Text   Cite this paper   Add to My Lib

Abstract:

?文中进行了广义线性反演(gli)的gpu加速研究,通过理论分析和模型数据检验表明,通过应用l1模稀疏脉冲反褶积,数据质量得到了显著提高.该方法不仅对多波数据适用,对单一纵、横波数据都可以进行反演.反演结果表明:①应用l1模稀疏脉冲反褶积可以在反演前消除子波的效应,从而使gli在反射系数层面上发挥更好的作用.②将gpu并行计算技术整合到反演过程中,能够极大地提高计算速度和效率,对单个角度域共成像点道集(adcig)的加速达数百倍以上;在处理海量的叠前反演数据时,这种优势将更加明显.③在反演中采用并行计算,必须摒弃细粒度的任务设计思路,采用较粗粒度的设计,才能尽量减少延迟,提高计算效率.

References

[1]  cookeda,schneiderwa.generalizedlinearinversionofreflectionseismicdata[j].geophysics.1983,48(6):665-676[2]macdonaldc,davisp,jacksond.inversionofreflectiontraveltimesandamplitudes[j].geophysics.1987,52(5):606-617[3]russellbh.introductiontoseismicinversionmethods.segbooks,1988.[4]rijssene,hermang.resolutionanalysisofband-limitedandoffset-limitedseismicdataforplane-layeredsubsurfacemodels[j].geophysicalprospecting.1991,39(1):61-76[5]dehaasj,berkhouta.ontheinformationcontentofpp,p-sv,sv-sv,andsv-preflections.segtechnicalprogramexpandedabstracts,1988,7:1190-1194.[6]micikeviciusp.3dfinitedifferencecomputationongpususingcuda.proceedingsof2ndworkshopongeneralpurposeprocessingongraphicsprocessingunits.acm,2009,79-84.[7]张兵,赵改善,黄骏等.地震叠前深度偏移在cuda平台上的实现.勘探地球物理进展,2009,31(6):427-432]zhangbing,zhaogaishan,huangjunetal.seismicprestackdepthmigrationonthecudaplatform.progressinexplorationgeophysics,2009,31(6):427-432.[8]刘伟峰,王永胜,张天雷等.使用gpu模拟地震波传播的性能研究.系统仿真学报,2009,21(1):170-174liuweifeng,wangyongsheng,zhangtianleietal.performancestudyofseismicwavepropagationsimulationusinggpu.journalofsystemsimulation,2009,21(1):170-174.[9]李博,刘国峰,刘洪.地震叠前时间偏移的一种图形处理器提速实现方法.地球物理学报,2009,52(1):245-252]libo,liuguofeng,liuhong.amethodofusinggputoaccelerateseismicpre-stacktimemigration.chinesejournalofgeophysics,2009,52(1):245-252.[10]刘红伟,李博,刘洪等.地震叠前逆时偏移高阶有限差分算法及gpu实现.地球物理学报,2010,53(7):1725-1733liuhongwei,libo,liuhongetal.thealgorithmofhighorderfinitedifferencepre-stackreversetimemigrationandgpuimplementation.chinesejournalofgeophysics,2010,53(7):1725-1733.[11]menkew.geophysicaldataanalysis:discretein-versetheory.internationalgeophysicsseries,newyork:academicpress,1989.[12]martings,wileyr,marfurtkj.marmousi2:anelasticupgradeformarmousi[j].theleadingedge.2006,25(2):156-166
[2]  cookeda,schneiderwa.generalizedlinearinversionofreflectionseismicdata[j].geophysics.1983,48(6):665-676
[3]  macdonaldc,davisp,jacksond.inversionofreflectiontraveltimesandamplitudes[j].geophysics.1987,52(5):606-617
[4]  russellbh.introductiontoseismicinversionmethods.segbooks,1988.
[5]  rijssene,hermang.resolutionanalysisofband-limitedandoffset-limitedseismicdataforplane-layeredsubsurfacemodels[j].geophysicalprospecting.1991,39(1):61-76
[6]  dehaasj,berkhouta.ontheinformationcontentofpp,p-sv,sv-sv,andsv-preflections.segtechnicalprogramexpandedabstracts,1988,7:1190-1194.
[7]  micikeviciusp.3dfinitedifferencecomputationongpususingcuda.proceedingsof2ndworkshopongeneralpurposeprocessingongraphicsprocessingunits.acm,2009,79-84.
[8]  张兵,赵改善,黄骏等.地震叠前深度偏移在cuda平台上的实现.勘探地球物理进展,2009,31(6):427-432]zhangbing,zhaogaishan,huangjunetal.seismicprestackdepthmigrationonthecudaplatform.progressinexplorationgeophysics,2009,31(6):427-432.
[9]  刘伟峰,王永胜,张天雷等.使用gpu模拟地震波传播的性能研究.系统仿真学报,2009,21(1):170-174
[10]  liuweifeng,wangyongsheng,zhangtianleietal.performancestudyofseismicwavepropagationsimulationusinggpu.journalofsystemsimulation,2009,21(1):170-174.
[11]  李博,刘国峰,刘洪.地震叠前时间偏移的一种图形处理器提速实现方法.地球物理学报,2009,52(1):245-252]libo,liuguofeng,liuhong.amethodofusinggputoaccelerateseismicpre-stacktimemigration.chinesejournalofgeophysics,2009,52(1):245-252.
[12]  刘红伟,李博,刘洪等.地震叠前逆时偏移高阶有限差分算法及gpu实现.地球物理学报,2010,53(7):1725-1733
[13]  liuhongwei,libo,liuhongetal.thealgorithmofhighorderfinitedifferencepre-stackreversetimemigrationandgpuimplementation.chinesejournalofgeophysics,2010,53(7):1725-1733.
[14]  menkew.geophysicaldataanalysis:discretein-versetheory.internationalgeophysicsseries,newyork:academicpress,1989.
[15]  martings,wileyr,marfurtkj.marmousi2:anelasticupgradeformarmousi[j].theleadingedge.2006,25(2):156-166

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133