基于稀疏表示的阵列声波测井仪数据无损压缩传输方法  

Data Lossless Compression Transporting Method of Array Acoustic Logging Tool Based on Signal Sparse Representation

作　　者：李明 尹时松 张宁 李波宏 庄献华 李江山 LI Ming;YIN Shi-song;ZHANG Ning;LI Bo-hong;ZHUANG Xian-hua;LI Jiang-shan(Huabei Branch of China Petroleum Logging Co.,Ltd.,Renqiu 062552,China)

机构地区：[1]中国石油集团测井有限公司华北分公司,河北任丘062552

出　　处：《测控技术》2022年第5期106-112,共7页Measurement & Control Technology

摘　　要：交叉多极子阵列声波测井仪在井下信号采集的同时,采用无损压缩提升单位时间内上传数据量,是目前主流的仪器上传带宽增加方式。针对传统压缩方法压缩率较低,导致仪器在单位深度地层工作时长过长的问题,从信号稀疏表示的角度出发,对采集的多路声波波列采用预先构建的稀疏变换矩阵进行稀疏变换,将求解的稀疏表示系数和其重构信号与原始信号的误差进行压缩编码上传;地面系统通过相同的稀疏变换矩阵进行信号重构,实现解码;其中,稀疏变换矩阵采用K-SVD算法进行预训练,提升稀疏变换系数的稀疏度与重构信号精度,进一步降低上传的压缩编码长度。在HB油田3口井实际测井资料的实验中,本方法与目前主流的测井数据压缩方法相比,压缩率平均提升约17.3%;在4口井的阵列声波实际测井作业的应用测试中,作业效率平均提升约20.2%。结果表明,数据压缩传输算法极大地提升了阵列声波测井时效,在保证数据采集质量的同时,实现了阵列声波仪器的高速测量。Improving up-load data volume in unit time by lossless compression while signals are acquired down-hole is the main stream pattern for crossed multipole array acoustic logging tool to increase up-load transmitting bandwidth.Pointing to the problem of longer working time on unit depth zone caused by lower compression ratio of traditional methods,sparse transformation is conducted by pre-built sparse transforming matrix to upload solved sparse representing coefficients and error between reconstructed signal and raw signal from aspect of signal spares representation.Decoding process is realized by signal reconstructing adopting same sparse transforming matrix on logging ground system which is pre-trained by K-SVD algorithm to improve sparsity and reconstructing precision to decrease the length of uploaded codes further.Data compression ratio is improved by about 17.3%averagely compared with current mainstream methods on actual data from 3 wells of HB oilfield.Logging efficiency is improved by about 20.2%averagely on actual operations of 4 wells from HB oilfield.Result shows that efficiency is greatly improved by the algorithm which realizes high-speed data acquisition with the insurance of data acquiring quality.

关 键 词：阵列声波测井 信号稀疏表示 K-SVD算法 

分 类 号：TP63[自动化与计算机技术—控制理论与控制工程]