Prestack seismic hybrid inversion based on simulated annealing algorithm with adaptive optimization parameters
GUO Qiang1,2, LUO Cong2, LIU Hongda3, HUANG Jian4
1. School of Resources and Geosciences, China University of Mining and Technology, Xuzhou, Jiangshu 221116, China; 2. School of Earth Sciences and Engineering, Hohai University, Nanjing, Jiangshu 211100, China; 3. No.1 Geological Team of Shandong Provincial Bureau Institute of Geology and Mineral Resources, Jinan, Shandong 250100, China; 4. College of Information Engineering, China Jiliang University, Hangzhou, Zhejiang 310018, China
Abstract：Prestack seismic inversion is an important technique for the quantitative processing and interpretation of seismic data, and the global optimization algorithm is an effective approach in prestack seismic inversion to estimate the elastic parameters. As a main global optimization algorithm, the simulated annealing algorithm has been widely used in inverting seismic data. However, the algorithm involves multiple optimization parameters (e.g., initial temperature, disturbance range, etc.), which have great effects on the inversion results. The parameter optimization by simulated annealing is mainly based on model trial calculation, but such an empirical method is easy to introduce errors and lacks promotion. To this end, this paper proposes a prestack seismic hybrid inversion method based on the simulated annealing algorithm with adaptive optimization parameters. Specifically, the Bayesian linear inversion and simulated-annealing nonlinear inversion are combined. The linear inversion results are used to estimate subsequent optimization parameters and construct prior models; according to the data differences of different seismic tracks, the applicable initial temperature and disturbance range are calculated track by track, which effectively improves the applicability and stability of the algorithm. The synthetic data tests and the application of measured data show that compared with those of the conventional simulated annealing inversion method, the inversion results of elastic parameters by the proposed method have a higher correlation coefficient with the logging data.
ZONG Z, YIN X, WU G. Geofluid discrimination incorporating poroelasticity and seismic reflection inversion[J]. Surveys in Geophysics, 2015, 36(5): 659-681.
[2]
HAMMER H, KOLBJØRNSEN O, TJELMELAND H, et al. Lithology and fluid prediction from prestack seismic data using a Bayesian model with Markov process prior[J]. Geophysical Prospecting, 2012, 60(3): 500-515.
[3]
印兴耀, 崔维, 宗兆云, 等. 基于弹性阻抗的储层物性参数预测方法[J]. 地球物理学报, 2014, 57(12): 4132-4140.YIN Xingyao, CUI Wei, ZONG Zhaoyun, et al. Petrophysical property inversion of reservoirs based on elastic impedance[J]. Chinese Journal of Geophy-sics, 2014, 57(12): 4132-4140.
[4]
BULAND A, OMRE H. Bayesian linearized AVO inversion[J]. Geophysics, 2003, 68(1): 185-198.
[5]
FATTI J L, SMITH G C, VAIL P J, et al. Detection of gas in sandstone reservoirs using AVO analysis: a 3-D seismic case history using the Geostack technique[J]. Geophysics, 1994, 59(9): 1362-1376.
[6]
黄捍东, 王彦超, 郭飞, 等. 基于佐普里兹方程的高精度叠前反演方法[J]. 石油地球物理勘探, 2013, 48(5): 740-749.HUANG Handong, WANG Yanchao, GUO Fei, et al. High precision prestack inversion algorithm based on Zoeppritz equations[J]. Oil Geophysical Prospecting, 2013, 48(5): 740-749.
[7]
LUO C, LI X, HUANG G. Pre-stack AVA inversion by using propagator matrix forward modeling[J]. Pure and Applied Geophysics, 2019, 176(10): 4445-4476.
[8]
周林, 李景叶, 陈小宏. 基于精确Zoeppritz方程的非线性AVO三参数反演[J]. 地球物理学报, 2016, 59(7): 2663-2673.ZHOU Lin, LI Jingye, CHEN Xiaohong. Nonlinear three-term AVO inversion based on exact Zoeppritz equation[J]. Chinese Journal of Geophysics, 2016, 59(7): 2663-2673.
[9]
TARANTOLA A, VALETTE B. Generalized nonlinear inverse problems solved using the least squares criterion[J]. Reviews of Geophysics, 1982, 20(2): 219-232.
[10]
霍国栋, 杜启振, 王秀玲, 等. 纵向和横向同时约束AVO反演[J]. 地球物理学报, 2017, 60(1): 271-282.HUO Guodong, DU Qizhen, WANG Xiuling, et al. AVO inversion constrained simultaneously in vertical and lateral directions[J]. Chinese Journal of Geophysics, 2017, 60(1): 271-282.
[11]
王宁, 周辉, 王玲谦, 等. 数据驱动的块排列正则化多道叠前地震反演[J]. 地球物理学报, 2022, 65(7): 2681-2692.WANG Ning, ZHOU Hui, WANG Lingqian, et al. Data-driven multichannel pre-stack seismic inversion via patch-ordering regularization[J]. Chinese Journal of Geophysics, 2022, 65(7): 2681-2692.
[12]
张繁昌, 李栋, 代荣获. 基于边缘保持平滑正则化的地震反演方法[J]. 中国矿业大学学报, 2015, 44(2): 255-261.ZHANG Fanchang, LI Dong, DAI Ronghuo. Seismic inversion based on edge preserving smooth regularization[J]. Journal of China University of Mining & Technology, 2015, 44(2): 255-261.
[13]
WANG K, SUN Z, DONG N. Prestack inversion based on anisotropic Markov random field-maximum posterior probability inversion and its application to identify shale gas sweet spots[J]. Applied Geophy-sics, 2015, 12(4): 533-544.
[14]
YAN Z, GU H. Non-linear prestack seismic inversion with global optimization using an edge-preserving smoothing filter[J]. Geophysical Prospecting, 2013, 61(4): 747-760.
[15]
方中于, 王丽萍, 杜家元, 等. 基于混合智能优化算法的非线性AVO反演[J]. 石油地球物理勘探, 2017, 52(4): 797-804.FANG Zhongyu, WANG Liping, DU Jiayuan, et al. Nonlinear AVO inversion based on hybrid intelligent optimization algorithm[J]. Oil Geophysical Prospecting, 2017, 52(4): 797-804.
[16]
罗红梅, 王长江, 智龙霄, 等. 基于精确Zoeppritz方程和改进人工蜂群算法的AVO反演方法研究[J]. 石油物探, 2022, 61(4): 659-672.LUO Hongmei, WANG Changjiang, ZHI Longxiao, et al. AVO inversion method based on exact Zoeppritz equations and improved artificial bee colony algorithm[J]. Geophysical Prospecting for Petroleum, 2022, 61(4): 659-672.
[17]
SEN M K, STOFFA P L. Global Optimization Methods in Geophysical Inversion[M]. Cambridge University Press, Cambridge, 2013.
[18]
ROTHMAN D H. Nonlinear inversion, statistical mechanics, and residual statics estimation[J]. Geophysics, 1985, 50(12): 2784-2796.
[19]
李景叶, 陈小宏. 用改进的模拟退火算法进行叠后时移地震数据反演[J]. 石油地球物理勘探, 2003, 38(4): 392-395.LI Jingye, CHEN Xiaohong. Inversion of poststack time-lapse seismic data by modified simulated annealing algorithm[J]. Oil Geophysical Prospecting, 2003, 38(4): 392-395.
[20]
MISRA S, SACCHI M D. Global optimization with model-space preconditioning: application to AVO inversion[J]. Geophysics, 2008, 73(5): R71-R82.
[21]
周琦, 印兴耀, 李坤. 煤系地层地震岩石物理建模及横波预测方法[J]. 石油地球物理勘探, 2022, 57(2): 357-366.ZHOU Qi, YIN Xingyao, LI Kun. Seismic rock physics modeling and shear wave velocity prediction method of coal measure strata[J]. Oil Geophysical Prospecting, 2022, 57(2): 357-366.
[22]
GUO Q, ZHANG H, CAO H, et al. Hybrid seismic inversion based on multi-order anisotropic Markov random field[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(1): 407-420.
[23]
张凌远, 张宏兵, 尚作萍, 等. 基于Zoeppritz方程的叠前和叠后混合多参数非线性地震反演[J]. 石油地球物理勘探, 2021, 56(1): 164-171.ZHANG Lingyuan, ZHANG Hongbing, SHANG Zuoping, et al. Nonlinear multi-parameter hybrid inversion of pre-stack and post-stack seismic data based on Zoeppritz equation[J]. Oil Geophysical Prospecting, 2021, 56(1): 164-171.
[24]
BASU A, FRAZER L N. Rapid determination of the critical temperature in simulated annealing inversion[J]. Science, 1990, 249(4975): 1409-1412.
[25]
姚姚. 地球物理非线性反演模拟退火法的改进[J]. 地球物理学报, 1995, 38(5): 643-650.YAO Yao. Improvement on nonlinear geophysical inversion simulated annealing[J]. Chinese Journal of Geophysics, 1995, 38(5): 643-650.
[26]
AARTS E, KORST J. Simulated Annealing and Boltzmann Machines: A Stochastic Approach to Combinatorial Optimization and Neural Computing[M]. John Wiley & Sons, New York,1989.
[27]
AKI K,RICHARDS P G.Quantitative Seismology:Theory and Methods[M].W.H.Freeman,San Francisco,1980.
[28]
张赛民, 陈灵君, 刘海飞, 等. 基于模型分块交叉移动的学习型模拟退火算法地震反演[J]. 中南大学学报(自然科学版), 2012, 43(3): 1040-1046.ZHANG Saimin, CHEN Lingjun, LIU Haifei, et al. Seismic inversion based on simulated annealing of learning and cross-moving with divided block model[J]. Journal of Central South University (Science and Technology), 2012, 43(3): 1040-1046.
[29]
张霖斌, 姚振兴, 纪晨, 等. 快速模拟退火算法及应用[J]. 石油地球物理勘探, 1997, 32(5): 654-660.ZHANG Linbin, YAO Zhenxing, JI Chen, et al. Fast simulated annealing algorithm and its application[J]. Oil Geophysical Prospecting, 1997, 32(5): 654-660.
