时移地震数据互均化处理并行算法研究

发布时间：2018-09-14 20:16

【摘要】：随着油气勘探目标越来越复杂，人们正逐步加深对地震资料处理重要性的认识，并愈加迫切地需要高精度勘探技术。并行处理技术日益得到石油地球物理界的广泛关注，如何快速高效地并行处理大规模地震数据这一个问题已经成为急需解决的重大课题之一。本文中利用CUDA平台来实现计算复杂度较大的大数据量的处理；利用云计算平台Hadoop对大规模的地震数据进行处理；并将HadoopDB技术应用到时移地震数据处理中，来处理一些相对关系比较复杂的地震数据。近年来，GPU绘制流水线的高速度和并行性以及可编程功能，使其在通用计算领域的应用有着巨大的潜力。本文利用GPU并行计算技术实现对时移地震数据的相位校正和振幅校正中任务繁重的计算进行并行化处理。实验结果表明基于GPU的并行计算提高该算法的运算效率，并且该方法可以应用于多种时移地震数据互均化校正算法。云计算是一种高效处理数据密集型应用的计算模式，Hadoop平台是基于MapRecuce编程模型的云计算软件平台。本文基于Hadoop来实现时延校正和匹配校正的并行算法，，将大数据集分成小数据集，分配到异构的计算机集群上，有效的实现了大规模海量数据的并行处理。实验结果表明基于Hadoop的并行算法在处理大规模数据集时具有明显的优势。 HadoopDB是集高吞吐、大并发、海量数据存储和分析计算与一体的技术，它在Hadoop的基础上进行了改进并使用了关系型数据库保存数据，是一种混合型的海量数据存储计算技术。本文将HadoopDB技术应用到时移地震数据互均化处理上，将HadoopDB的基本构造和特性与时移地震数据互均化处理相结合，提出一种解决方案。最后我们对论文工作进行了总结，并讨论了对进一步工作的展望。
[Abstract]:With the increasing complexity of oil and gas exploration targets, people are gradually deepening the understanding of the importance of seismic data processing, and increasingly urgent need for high-precision exploration technology. The technology of parallel processing has been paid more and more attention by the petroleum geophysics. How to process large scale seismic data in parallel quickly and efficiently has become one of the important problems that need to be solved. In this paper, the CUDA platform is used to process large amount of data with large computational complexity; the cloud computing platform Hadoop is used to process large-scale seismic data; and the HadoopDB technology is applied to the processing of time-lapse seismic data. To deal with relatively complex seismic data. In recent years, the high speed, parallelism and programmable function of GPU rendering pipeline make it have great potential in the field of general computing. In this paper, the GPU parallel computing technique is used to parallelize the phase correction and amplitude correction of time-lapse seismic data. The experimental results show that the parallel computation based on GPU can improve the efficiency of the algorithm, and this method can be applied to a variety of time-lapse seismic data mutual equalization correction algorithms. Cloud computing is an efficient computing mode for data-intensive applications. Hadoop platform is a cloud computing software platform based on MapRecuce programming model. In this paper, the parallel algorithm of delay correction and matching correction is implemented based on Hadoop. Big data set is divided into small data sets and assigned to heterogeneous computer clusters, which effectively implements the parallel processing of massive data. The experimental results show that the parallel algorithm based on Hadoop has obvious advantages in dealing with large data sets. HadoopDB is a set of high throughput, large concurrency, mass data storage, analysis and calculation. It improves on Hadoop and uses relational database to save data. It is a hybrid computing technique for mass data storage. In this paper, the HadoopDB technique is applied to the reciprocal equalization of time-lapse seismic data, and a solution is proposed by combining the basic structure and characteristics of HadoopDB with the time-lapse seismic data mutual equalization. Finally, we summarize the work of the paper and discuss the prospects for further work.
【学位授予单位】：烟台大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P631.4;TP338.6

【相似文献】