摘要
声波测井是获得井眼附近地层参数的重要测井方法之一。然而许多钻井中存在着极端扩径情况,这对声波测井数据有一定的影响,因此有必要对声波测井数据作环境校正。已经证明声波测井极端扩径校正算法可以有效地校正因井径不规则产生的对声波测井数据的影响。然而该方法在计算机运行的过程中暴露出数据占用空间大、运行时间较长等弊端,无法满足测井解释工作现场快速处理的要求。针对声波反演校正算法的这些弊端,通过对声波测井极端扩径反演校正算法的研究,根据CUDA并行计算适合大规模重复计算的特点,设计了声波反演校正算法的CUDA并行算法。在搭建的CUDA编程平台上,实现了声波测井极端扩径校正算法的并行计算。通过对实际井资料的处理实验发现与串行计算相比,CUDA并行计算在保证精度的基础上处理200米井数据所用时间可降低30%。因此从计算结果和计算时间上说明声波测井极端扩径校正算法适合在CUDA上并行计算。
Acoustic logging is one of the important methods to gain formation parameters near the borehole.However there is extreme hole enlargement condition in drilling, which influences the acoustic logging data. Environmental correction is necessary.It has been proved that the extreme hole enlargement correction algorithm can correct the influence caused by irregular hole diameter on acoustic logging data. But some disvantages in the method expose such as taking up large space and long running time, etc., which can not meet the need of immediate processing of logging interpretation in the working field.Through research, it is found that CUDA parallel computation is suitable for large scale and repeat calculation. Therefore CUDA parallel algorithm is designed for acoustic wave reversion correction algorithmic.It is realized in CUDA programming platform.Comparing with serial computing in the experiments, CUDA parallel computation reduces 30% time in processing 200m interval data with accurate results.It shows from the results and time that acoustic logging extreme hole enlargement correction algorithm fits to CUDA parallel computation.
出处
《国外测井技术》
2013年第1期7-10,3,共4页
World Well Logging Technology
基金
国家科技重大专项(2011ZX05023-005-006)
关键词
声波测井
扩径反演校正
并行计算
CUDA
acoustic logging
hole enlargement reversion correction
parallel computation
CUDA
