基于基追踪的测井信号分离
|
摘要
石油是关系到我国国计民生的重要战略资源,而我国对外依存度占石油消费的一半以上,这严重威胁着我国的能源安全。因而,大力发展石油开采技术,不仅关系到我国经济能否持续、快速发展,而且关系到我国能源供应的战略安全。
声波测井信号处理技术是目前勘探石油等能源资源的有力手段,但这种技术仍存在一定的缺陷。目前国内外仅利用首波信息,来获取地层的岩性、孔隙和油气含量等关键信息,通常与实际情况存在误差。近年来,基追踪(Basis Pursuit,BP)算法成为信号稀疏表示领域研究的热点,该算法的思想是从过完备原子库中找到信号最稀疏的表示,也就是用尽可能少的原子表示原信号,进而获取信号的内在特性。根据声波测井信号的特性将该方法应用到测井数据处理中,为获取更加全面的测井信息提供了新的方向。
本文首先介绍了信号表示及信号稀疏分解的基础知识,接着以短时傅里叶变换为例介绍了常用时频分析方法的缺陷,引出了基于BP的稀疏分解方法。通过对模型信号的处理,实验结果证明利用BP算法进行信号分离是可行的。在将声波测井信号模型成功分离的基础上,把BP算法应用到实际声波测井信号中。通过对不同条件下实际声波测井信号的处理,实验证明利用BP算法仍能实现分离。根据分离结果分析不同条件下测井信号的特点,并且实验证明不同条件下测井信号的分离效果不同。最后,对含噪的实际声波测井信号进行处理,实验结果证明BP算法仍能将信号进行分离,但是由于噪声的干扰,含噪信号的分离结果比预处理后稍差。
Oil is not only related to our national economy but also is an important strategic resource, and our country dependence on foreign oil consumption accounted for more than half of this severe, which threat to China's energy security. Thus, to develop oil exploration technology is not only for the sustainability and rapid development of our economy, but also for the strategic security of the energy supply.
Acoustic logging signal processing technology is an powerful means for oil and other energy resources prospecting currently, but this technology still has some flaws. At home and abroad using only the information of the first wave, to obtain the formation of lithology, porosity, oil content and other key information, Usually there are some errors with the actual situation. In recent years, the Basis Pursuit(BP) algorithm is the research focus of sparse representation of the signal field, the idea of the algorithm is to find the most sparse representation of signals based on the over-complete dictionary. In other words, using as little atoms to respect the original signal as possible, then obtaining the intrinsic properties of the signal, using this method to handle the logging data based on the characteristics of acoustic logging signal, providing a new direction to get more comprehensive logging information.
The signal representation and signal sparse decomposition were introduced firstly in this thesis, and then to the basic knowledge of short-time Fourier transformation for example, the defects of common time-frequency analysis method were introduced, and based on BP the sparse decomposition method was raised. Based on model signal processing, experimental results prove BP signal separation is feasible. In will acoustic logging signal model on the basis of successful separation, the BP algorithm applied to actual acoustic logging signal. According to the different conditions actual acoustic logging signal processing, the experiment proved that BP algorithm can still achieve separation. According to the analysis of separation results under different conditions, and the characteristic of well logging signal experimental proof under various conditions, the separation effect of different logging signal. Finally, the practical with noise acoustic logging signal processing, the experimental results show the BP algorithm can still will signal, but because the separation of the noise, with noise signals than separation results after pretreatment is a bit poor.
声波测井信号处理技术是目前勘探石油等能源资源的有力手段,但这种技术仍存在一定的缺陷。目前国内外仅利用首波信息,来获取地层的岩性、孔隙和油气含量等关键信息,通常与实际情况存在误差。近年来,基追踪(Basis Pursuit,BP)算法成为信号稀疏表示领域研究的热点,该算法的思想是从过完备原子库中找到信号最稀疏的表示,也就是用尽可能少的原子表示原信号,进而获取信号的内在特性。根据声波测井信号的特性将该方法应用到测井数据处理中,为获取更加全面的测井信息提供了新的方向。
本文首先介绍了信号表示及信号稀疏分解的基础知识,接着以短时傅里叶变换为例介绍了常用时频分析方法的缺陷,引出了基于BP的稀疏分解方法。通过对模型信号的处理,实验结果证明利用BP算法进行信号分离是可行的。在将声波测井信号模型成功分离的基础上,把BP算法应用到实际声波测井信号中。通过对不同条件下实际声波测井信号的处理,实验证明利用BP算法仍能实现分离。根据分离结果分析不同条件下测井信号的特点,并且实验证明不同条件下测井信号的分离效果不同。最后,对含噪的实际声波测井信号进行处理,实验结果证明BP算法仍能将信号进行分离,但是由于噪声的干扰,含噪信号的分离结果比预处理后稍差。
Oil is not only related to our national economy but also is an important strategic resource, and our country dependence on foreign oil consumption accounted for more than half of this severe, which threat to China's energy security. Thus, to develop oil exploration technology is not only for the sustainability and rapid development of our economy, but also for the strategic security of the energy supply.
Acoustic logging signal processing technology is an powerful means for oil and other energy resources prospecting currently, but this technology still has some flaws. At home and abroad using only the information of the first wave, to obtain the formation of lithology, porosity, oil content and other key information, Usually there are some errors with the actual situation. In recent years, the Basis Pursuit(BP) algorithm is the research focus of sparse representation of the signal field, the idea of the algorithm is to find the most sparse representation of signals based on the over-complete dictionary. In other words, using as little atoms to respect the original signal as possible, then obtaining the intrinsic properties of the signal, using this method to handle the logging data based on the characteristics of acoustic logging signal, providing a new direction to get more comprehensive logging information.
The signal representation and signal sparse decomposition were introduced firstly in this thesis, and then to the basic knowledge of short-time Fourier transformation for example, the defects of common time-frequency analysis method were introduced, and based on BP the sparse decomposition method was raised. Based on model signal processing, experimental results prove BP signal separation is feasible. In will acoustic logging signal model on the basis of successful separation, the BP algorithm applied to actual acoustic logging signal. According to the different conditions actual acoustic logging signal processing, the experiment proved that BP algorithm can still achieve separation. According to the analysis of separation results under different conditions, and the characteristic of well logging signal experimental proof under various conditions, the separation effect of different logging signal. Finally, the practical with noise acoustic logging signal processing, the experimental results show the BP algorithm can still will signal, but because the separation of the noise, with noise signals than separation results after pretreatment is a bit poor.
引文
[1]李仙枝,乔文孝,车小花.声波测井波形信号波至点提取方法研究[J].声学技术,2008,27(5)：274-275.
[2]张珂.测井信号的采集与处理[D].西南石油大学硕士论文,2006.5.
[3]房文静.测井多尺度分析方法及应用研究[D].中国石油大学(华东)博士论文,2007.1.
[4]王栋.声波信号在测井中的影响及应用[J].石油和化工设备,2009.12(10)：29-30.
[5]Li Xiaoyan, Yin Zhongke. A Method Approach to the Extraction of the signal arrival from sonic logging signal on energy operater[J].2010:4842-4845.
[6]Chen S, Donoho D, Saunders M. Atomic decomposition by basis pursuit[J]. SIAM Journal on Scientific Computing,1999,20:36-60.
[7]桂志先,贺振华.裂隙介质中纵波地面地震记录模拟及特征[J],计算物理,2004,21(4)：91-94.
[8]李荣华,赵敏.盲信号分离的理论与发展现状[J].移动通信.2008,4：67-72.
[9]朱茉,季策,于洋.盲信号分离问题的分类与现状[J].化工自动化及仪表.2009,36(3)：7-11.
[10]Cao X R, Liu R W. A general approach to blind source separation[J]. IEEE Trans Signal Processing 1996,44(3):562-571.
[11]王纪伟,高宝成.一种盲信号分离算法的改进研究[J].计算机与现代化.2010 2：52-54.
[12]Comon P. Independent component analysis a new concept[J]. Signal Processing 1994,36(3):287-314.
[13]游荣义,陈忠.一种基于ICA的盲信号分离快速算法[J].电子学报,2004,32(4)：201-211.
[14]吴小培,冯焕清,周培琴.基于独立分量分析的混合声音信号分离[J].中国科学技术大学学报,2001,31(1)：121～126.
[15]李木森,毛剑琴.盲信号分离的现状和展望[J].信息与电子工程,2003,1(1)：69-78.
[16]Jutten C. Herault J Blind separation of sources Part Ⅰ An adaptive algorithm based on neurom in etic arch itecture[J]. Signal Processing 1991,24(1): 1-10.
[17]S, Amari. Natural gradient works eHciently in learning[J]. Neural Computation,1998.10(2):251-276.
[18]杜军,基于稀疏编码和ICA的带噪混叠语音盲分离[J].山东师范大学学报.2008,23(4)：136-138.
[19]蔡权伟,魏平,肖先赐.单通道多信号分量分离[J]通信学报,2006,27(6)：49-56.
[20]蔡权伟,魏平,肖先赐.基于能量算子的单信道重叠信号盲分离方法[J].《中国科学》杂志社,2008,38(4)：607-609.
[21]于宁宇,石荣,石磊.单通道时频重叠多信号处理技术进展与展望[J].电子信息对抗技术.2010,25(9)：1-7.
[22]冯林,程璐,吴振宇.基于第二代小波变换的测井信号处理[J].石油仪器.2004,6(3)：6-8.
[23]林盛,刘业新,李衍达.基于时—频分解技术的全波列声波测井信号处理[J].清华大学学报(自然科学版).1997,37(3)：63-66.
[24]BLUMENSATHT, DAVIES M. Shift-Invariant Sparse Coding for Single Channel Blind Souece Separation[J]. IEEE Transactions on Audio, Speech, and Language Processing,2006,14(1):50-57.
[25]SANJEEV A M, SIMONM, NELLG. A Tutorial on Particle Filters for Online Nonlinear/Norr Gaussian Bayesian Tracking[J]. IEEE Transactions on Signal Processing,2002,50(2):3-13.
[26]何昭水,谢胜利,傅予力.信号的稀疏性分析[J].自然科学进展.2006,16(9)：1167-1173.
[27]Li Y Q, Wang J, Sequential blind extraction of instantaneously mixed sources[J]. IEEE Trans. Signal Processing,2002,50(5):997-1006.
[28]Zhang L, Cichocki A, Amari S. Self-adaptive blind source separation based on activation function adaptation[J]. IEEE Trans. Neural Networks, 2004,15(2):233-244.
[29]Mallat S, Zhang Z. Matching pursuits with time-frequency dictionaries[J].IEEE Trans. Signal Processing,1993,41:3397-3415.
[30]高瑞,徐华楠,胡钢.基于GA和过完备原子库划分的MP信号稀疏分解算法[J].科学技术与工程.2008,8(4)：914-920.
[31]杨愚,李恒建,利用粒子群优化算法实现基于MP的信号稀疏分解[J],IECT,2006.11.
[32]邵君,尹忠科,王建英.基于FFT的MP信号稀疏分解算法的改进[J］.西南交通大学学报,2006,41(4)：466～470.
[33]刘浩,潘伟.基于FHT的实信号稀疏分解快速算法[J].铁道学报,2009.44(1)：45-48.
[34]Y.C.Pati, R.Rezaifar, P.S.Krishnaprased. Orthogonal matching pursuit recursive function approximation with applications to wavelet decomposition[C]. In 27th Asilomar Conf. on signals, Systems and Comput. November,1993.
[35]R. Gbrinoal, E.Baeyr, S.Mallat. Analysis of sound signals with high resolution matching pursuit[C], Time-Frequency and Time-Scale Analysis, Proceeding of the IEEE-SP International Symposium,1996:125-128.
[36]Pennec E, Mallat S. Image compression with geometrical wavelets[A]. Proceeding of IEEE ICIP[C]. Vancouver, Canada,2000.661-664.
[37]Do M, Vetterli M. Contourlet[A]. J Stoeckler, GV Welland. Beyond Wavelts[C]. Academic Press.2002.
[38]Gribonval R, Nielsen M. Sparse decompositions in 'incoherent' dictionaries[A], Proceedings of IEEE ICIP[C],2003:33-36.
[39]Bertsekas D. Non-Linear Programming[J]. Athena Scientific, Belmont, MA, 2nd edition,1995.
[40]Georgiev, P, Cichoki, A. Sparse component analysis of overcomplete mixtures by improved basis pursuit method[J]. Circuits and Systems,2004. ISCAS '04. Proceedings of the 2004 International Symposium on Volume 5,23-26.
[41]Huggins, P. S, Zucker, S.W. Greedy Basis Pursuit Signal Processing[J], IEEE Transactions on Volume 55, Issue 7, Part 2,2007,7:3760-3772
[42]Seung-Jean Kim, K. Koh, M. LUstig, Stephen Boyd, and Dimitry Gorinevsky. An Interior-Piont Method for Large-Scale l1-Regularized Lease Squares IEEE Journal of Selected Topics In Signal Processing[J],2007,1(4):606-617.
[43]王潇.MP和BP稀疏分解在盲源分离中的应用[D].西南交通大学硕士论文,2009.6.
[44]桂丽华.BP在面波分离和压制中的应用[D].西南交通大学硕士论文,2010.5.
[45]徐兵.基于BP的地震信号分离及初步应用[D].西南交通大学硕士论文,2010.5.
[46]P. Bloomfield and W.Steiger, Least A bsolute Deviations Theory Applications and Algorithms[J]. Birkhauser, Boston,1983,26(2):126-131.
[47]Portnoff MR. Time-Frequency Representation of Digital Signals and Systems Based on Short-Time Fourier Analysis[J]. IEEE Transactions on Acoustics, Speech and Signal Processing,1980,ASSP-28(1).
[48]田鑫,章成广,江万哲.短时傅里叶变换在阵列声波信息提取中的应用[J].国外测井技术,2005,20(3)：53-55.
[49]张跃飞.基于稀疏分解的图像压缩[D].硕士学位论文,西南交通大学,2006.1.
[50]裴正林,董玉珊,彭苏萍.裂隙煤层弹性波场方位各向异性特征数值模拟研究[J].石油地球物理勘探,2007,42(6)：665-672.
[51]常子楠.基于匹配追踪的正弦建模语言编码研究[D].硕士学位论文,南京师范大学,2006.
[52]李庆忠.走向精确勘探的道路-多分辨地震勘探系统工程剖析[M].北京：石油工业出版社,1994.
[53]Sheriff R.E. Encyclopedic dictionary of exploration geophsics. Tulsa Society of Exploration Geophysicists[J],1994.
[54]邓小英Ricker子波最小二乘支持向量回归机及其在地震勘探资料去噪中的应用[D].吉林大学博士论文,2007.
[55]袁野,罗正玮,李月Ricker子波视主频的变化对混沌振子检测效果的影响[J],地球物理学进展,2006,21(1)：70-73.
[56]强琳,刘贵忠.阵列声波全波列测井信号波至提取的小波变换方法[J].石油物探,1997,36(2)：55-61.
[57]尹忠科,王建英,邵君.基于原子库结构特性的信号稀疏分解[J].西南交通大学学报,2005,40(2)：173-178.
[58]陈磊.信号稀疏分解在空间谱估计中的应用.西南交通大学硕士论文[D].2004.
[59]Gribonval R Fast Matching Pursuit with a Multiscale Dictionary of Gaussian Chirps[J]. IEEE Transactions on Signal Processing.2005,49(5):994-1001.
[60]汪雄良,王春玲.基于改进基追踪放大的信号去噪[J].计算机应用,2005.8(1),19-21.
[2]张珂.测井信号的采集与处理[D].西南石油大学硕士论文,2006.5.
[3]房文静.测井多尺度分析方法及应用研究[D].中国石油大学(华东)博士论文,2007.1.
[4]王栋.声波信号在测井中的影响及应用[J].石油和化工设备,2009.12(10)：29-30.
[5]Li Xiaoyan, Yin Zhongke. A Method Approach to the Extraction of the signal arrival from sonic logging signal on energy operater[J].2010:4842-4845.
[6]Chen S, Donoho D, Saunders M. Atomic decomposition by basis pursuit[J]. SIAM Journal on Scientific Computing,1999,20:36-60.
[7]桂志先,贺振华.裂隙介质中纵波地面地震记录模拟及特征[J],计算物理,2004,21(4)：91-94.
[8]李荣华,赵敏.盲信号分离的理论与发展现状[J].移动通信.2008,4：67-72.
[9]朱茉,季策,于洋.盲信号分离问题的分类与现状[J].化工自动化及仪表.2009,36(3)：7-11.
[10]Cao X R, Liu R W. A general approach to blind source separation[J]. IEEE Trans Signal Processing 1996,44(3):562-571.
[11]王纪伟,高宝成.一种盲信号分离算法的改进研究[J].计算机与现代化.2010 2：52-54.
[12]Comon P. Independent component analysis a new concept[J]. Signal Processing 1994,36(3):287-314.
[13]游荣义,陈忠.一种基于ICA的盲信号分离快速算法[J].电子学报,2004,32(4)：201-211.
[14]吴小培,冯焕清,周培琴.基于独立分量分析的混合声音信号分离[J].中国科学技术大学学报,2001,31(1)：121～126.
[15]李木森,毛剑琴.盲信号分离的现状和展望[J].信息与电子工程,2003,1(1)：69-78.
[16]Jutten C. Herault J Blind separation of sources Part Ⅰ An adaptive algorithm based on neurom in etic arch itecture[J]. Signal Processing 1991,24(1): 1-10.
[17]S, Amari. Natural gradient works eHciently in learning[J]. Neural Computation,1998.10(2):251-276.
[18]杜军,基于稀疏编码和ICA的带噪混叠语音盲分离[J].山东师范大学学报.2008,23(4)：136-138.
[19]蔡权伟,魏平,肖先赐.单通道多信号分量分离[J]通信学报,2006,27(6)：49-56.
[20]蔡权伟,魏平,肖先赐.基于能量算子的单信道重叠信号盲分离方法[J].《中国科学》杂志社,2008,38(4)：607-609.
[21]于宁宇,石荣,石磊.单通道时频重叠多信号处理技术进展与展望[J].电子信息对抗技术.2010,25(9)：1-7.
[22]冯林,程璐,吴振宇.基于第二代小波变换的测井信号处理[J].石油仪器.2004,6(3)：6-8.
[23]林盛,刘业新,李衍达.基于时—频分解技术的全波列声波测井信号处理[J].清华大学学报(自然科学版).1997,37(3)：63-66.
[24]BLUMENSATHT, DAVIES M. Shift-Invariant Sparse Coding for Single Channel Blind Souece Separation[J]. IEEE Transactions on Audio, Speech, and Language Processing,2006,14(1):50-57.
[25]SANJEEV A M, SIMONM, NELLG. A Tutorial on Particle Filters for Online Nonlinear/Norr Gaussian Bayesian Tracking[J]. IEEE Transactions on Signal Processing,2002,50(2):3-13.
[26]何昭水,谢胜利,傅予力.信号的稀疏性分析[J].自然科学进展.2006,16(9)：1167-1173.
[27]Li Y Q, Wang J, Sequential blind extraction of instantaneously mixed sources[J]. IEEE Trans. Signal Processing,2002,50(5):997-1006.
[28]Zhang L, Cichocki A, Amari S. Self-adaptive blind source separation based on activation function adaptation[J]. IEEE Trans. Neural Networks, 2004,15(2):233-244.
[29]Mallat S, Zhang Z. Matching pursuits with time-frequency dictionaries[J].IEEE Trans. Signal Processing,1993,41:3397-3415.
[30]高瑞,徐华楠,胡钢.基于GA和过完备原子库划分的MP信号稀疏分解算法[J].科学技术与工程.2008,8(4)：914-920.
[31]杨愚,李恒建,利用粒子群优化算法实现基于MP的信号稀疏分解[J],IECT,2006.11.
[32]邵君,尹忠科,王建英.基于FFT的MP信号稀疏分解算法的改进[J］.西南交通大学学报,2006,41(4)：466～470.
[33]刘浩,潘伟.基于FHT的实信号稀疏分解快速算法[J].铁道学报,2009.44(1)：45-48.
[34]Y.C.Pati, R.Rezaifar, P.S.Krishnaprased. Orthogonal matching pursuit recursive function approximation with applications to wavelet decomposition[C]. In 27th Asilomar Conf. on signals, Systems and Comput. November,1993.
[35]R. Gbrinoal, E.Baeyr, S.Mallat. Analysis of sound signals with high resolution matching pursuit[C], Time-Frequency and Time-Scale Analysis, Proceeding of the IEEE-SP International Symposium,1996:125-128.
[36]Pennec E, Mallat S. Image compression with geometrical wavelets[A]. Proceeding of IEEE ICIP[C]. Vancouver, Canada,2000.661-664.
[37]Do M, Vetterli M. Contourlet[A]. J Stoeckler, GV Welland. Beyond Wavelts[C]. Academic Press.2002.
[38]Gribonval R, Nielsen M. Sparse decompositions in 'incoherent' dictionaries[A], Proceedings of IEEE ICIP[C],2003:33-36.
[39]Bertsekas D. Non-Linear Programming[J]. Athena Scientific, Belmont, MA, 2nd edition,1995.
[40]Georgiev, P, Cichoki, A. Sparse component analysis of overcomplete mixtures by improved basis pursuit method[J]. Circuits and Systems,2004. ISCAS '04. Proceedings of the 2004 International Symposium on Volume 5,23-26.
[41]Huggins, P. S, Zucker, S.W. Greedy Basis Pursuit Signal Processing[J], IEEE Transactions on Volume 55, Issue 7, Part 2,2007,7:3760-3772
[42]Seung-Jean Kim, K. Koh, M. LUstig, Stephen Boyd, and Dimitry Gorinevsky. An Interior-Piont Method for Large-Scale l1-Regularized Lease Squares IEEE Journal of Selected Topics In Signal Processing[J],2007,1(4):606-617.
[43]王潇.MP和BP稀疏分解在盲源分离中的应用[D].西南交通大学硕士论文,2009.6.
[44]桂丽华.BP在面波分离和压制中的应用[D].西南交通大学硕士论文,2010.5.
[45]徐兵.基于BP的地震信号分离及初步应用[D].西南交通大学硕士论文,2010.5.
[46]P. Bloomfield and W.Steiger, Least A bsolute Deviations Theory Applications and Algorithms[J]. Birkhauser, Boston,1983,26(2):126-131.
[47]Portnoff MR. Time-Frequency Representation of Digital Signals and Systems Based on Short-Time Fourier Analysis[J]. IEEE Transactions on Acoustics, Speech and Signal Processing,1980,ASSP-28(1).
[48]田鑫,章成广,江万哲.短时傅里叶变换在阵列声波信息提取中的应用[J].国外测井技术,2005,20(3)：53-55.
[49]张跃飞.基于稀疏分解的图像压缩[D].硕士学位论文,西南交通大学,2006.1.
[50]裴正林,董玉珊,彭苏萍.裂隙煤层弹性波场方位各向异性特征数值模拟研究[J].石油地球物理勘探,2007,42(6)：665-672.
[51]常子楠.基于匹配追踪的正弦建模语言编码研究[D].硕士学位论文,南京师范大学,2006.
[52]李庆忠.走向精确勘探的道路-多分辨地震勘探系统工程剖析[M].北京：石油工业出版社,1994.
[53]Sheriff R.E. Encyclopedic dictionary of exploration geophsics. Tulsa Society of Exploration Geophysicists[J],1994.
[54]邓小英Ricker子波最小二乘支持向量回归机及其在地震勘探资料去噪中的应用[D].吉林大学博士论文,2007.
[55]袁野,罗正玮,李月Ricker子波视主频的变化对混沌振子检测效果的影响[J],地球物理学进展,2006,21(1)：70-73.
[56]强琳,刘贵忠.阵列声波全波列测井信号波至提取的小波变换方法[J].石油物探,1997,36(2)：55-61.
[57]尹忠科,王建英,邵君.基于原子库结构特性的信号稀疏分解[J].西南交通大学学报,2005,40(2)：173-178.
[58]陈磊.信号稀疏分解在空间谱估计中的应用.西南交通大学硕士论文[D].2004.
[59]Gribonval R Fast Matching Pursuit with a Multiscale Dictionary of Gaussian Chirps[J]. IEEE Transactions on Signal Processing.2005,49(5):994-1001.
[60]汪雄良,王春玲.基于改进基追踪放大的信号去噪[J].计算机应用,2005.8(1),19-21.