基于改进DenseNet的大地电磁智能反演

doi:10.11720/wtyht.2024.1275

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (4789 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract

Magnetotelluric (MT) sounding is a vital exploration method in tunnel engineering. Inversion methods can assist geologists in interpreting geological data by converting MT data into geoelectric parameters. However, conventional inversion methods exhibit inferior timeliness and reliance on initial model settings. In this study, deep learning was applied to the one-dimensional inversion of magnetotelluric data. First, an improved DenseNet model was constructed and trained to invert geological models of various resistivity-variable strata, yielding a fast computational speed and high accuracy. Then, the robustness of the improved DenseNet model was tested, suggesting that its network structure can achieve satisfactory inversion results for noisy data. Finally, this artificial intelligence technique was applied to the MT data inversion of the Hongjiaqian tunnel in the Huangshan area, obtaining geophysical exploration results that match the geological research results. Additionally, relevant construction recommendations were given based on the inversion results.

Key words： magnetotellurics intelligent inversion deep learning tunnel engineering

Received: 26 June 2023 Published: 27 June 2024

ZTFLH:

P631

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Yu YAO
	Zhi-Hou ZHANG

Cite this article:

Yu YAO,Zhi-Hou ZHANG. Intelligent inversion of magnetotelluric data based on improved DenseNet[J]. Geophysical and Geochemical Exploration, 2024, 48(3): 759-767.

URL:

https://www.wutanyuhuatan.com/EN/10.11720/wtyht.2024.1275 OR https://www.wutanyuhuatan.com/EN/Y2024/V48/I3/759

Technology roadmap of one-dimensional intelligent inversion based on improved DenseNet

Construction of sample data

Schematic diagram of improved DenseNet

Error loss curve

Improved DenseNet inversion results of two-layer model and its apparent resistivity and phase comparison
a—inversion results of two-layer model; b—comparison of apparent resistivity; c—comparison of phase

Improved DenseNet inversion results of three-layer model and its apparent resistivity and phase comparison
a—inversion results of three-layer model; b—comparison of apparent resistivity; c—comparison of phase

Improved DenseNet inversion results of four-layer model and its apparent resistivity and phase comparison
a—inversion results of four-layer model; b—comparison of apparent resistivity; c—comparison of phase

Improved DenseNet inversion results of six-layer model and its apparent resistivity and phase comparison
a—inversion results of six-layer model; b—comparison of apparent resistivity; c—comparison of phase

Inversion results under different noise levels
a—inversion results of noisy data from two-layer geoelectric model; b—inversion results of noisy data from three-layer geoelectric model; c—inversion results of noisy data from four-layer geoelectric model; d—inversion results of noisy data from five-layer geoelectric model

Comparison for inversion results of the improved DenseNet and Occam

The misfit of the normalized objective function with iteration number for four-layer model

Aether system field measurement point layout

Comparison of inversion results
a—inversion results of SCS2D; b—inversion results of improved DenseNet

[1]	唐海敏, 张吉振, 王银, 等. 大地电磁法探测中条山隧道及断层的结构特征[J]. 铁道工程学报, 2013, 30(4):9-13.
[1]	Tang H M, Zhang J Z, Wang Y, et al. Exploration of structure characteristics of fault for Zhongtiaoshan tunnel with magnetotelluric method[J]. Journal of Railway Engineering Society, 2013, 30(4):9-13.
[2]	范祥泰, 张志厚, 苏建坤, 等. 大地电磁测深法探测山区深埋隧道隐伏构造——以安石隧道探测为例[J]. 科学技术与工程, 2021, 21(15):6211-6220.
[2]	Fan X T, Zhang Z H, Su J K, et al. Magnetotelluric sounding method for detecting concealed structures of deep-buried tunnels in mountainous areas:Taking the detection of anshi tunnel as an example[J]. Science Technology and Engineering, 2021, 21(15):6211-6220.
[3]	Li B, Xu Q, Yu J H, et al. Conductivity characteristics of magmatic rock intrusions contained in metamorphic rock strata in Mupi Highway Tunnel[J]. Applied Geophysics, 2022, 19(2):1-15.
[4]	Constable S C. Occam’s inversion:A practical algorithm for generating smooth models from electromagnetic sounding data[J]. Geophysics, 1987, 52(3):289.
[5]	Smith J T, Booker J R. Rapid inversion of two- and three-dimensional magnetotelluric data[J]. Journal of Geophysical Research:Solid Earth, 1991, 96(B3):3905-3922.
[6]	Rodi W, Mackie R L. Nonlinear conjugate gradients algorithm for 2D magnetotelluric inversion[J]. Geophysics, 2001, 66(1):174-187.
[7]	王鹤, 蒋欢, 王亮, 等. 大地电磁人工神经网络反演[J]. 中南大学学报:自然科学版, 2015, 46(5):1707-1714.
[7]	Wang H, Jiang H, Wang L, et al. Magnetotelluric inversion using artificial neural network[J]. Journal of Central South University:Science and Technology, 2015, 46(5):1707-1714.
[8]	胡祖志, 何展翔, 杨文采, 等. 大地电磁的人工鱼群最优化约束反演[J]. 地球物理学报, 2015, 58(7):2578-2587.
[8]	Hu Z Z, He Z X, Yang W C, et al. Constrained inversion of magnetotelluric data with the artificial fish swarm optimization method[J]. Chinese Journal of Geophysics, 2015, 58(7):2578-2587.
[9]	陈杰, 杨磊. 基于GA和PSO智能算法的大地电磁一维反演分析[J]. 工程地球物理学报, 2021, 18(5):561-567.
[9]	Chen J, Yang L. One-dimensional magnetotelluric inversion analysis based on GA and PSO intelligent algorithms[J]. Chinese Journal of Engineering Geophysics, 2021, 18(5):561-567.
[10]	王天意, 侯征, 何元勋, 等. 基于改进差分进化算法的大地电磁反演[J]. 地球物理学进展, 2022, 37(4):1605-1612.
[10]	Wang T Y, Hou Z, He Y X, et al. Magnetotelluric inversion based on the improved differential evolution algorithm[J]. Progress in Geophysics, 2022, 37(4):1605-1612.
[11]	杨凯, 唐卫东, 刘诚, 等. 基于LSTM循环神经网络的大地电磁方波噪声抑制[J]. 物探与化探, 2022, 46(4):925-933.
[11]	Yang K, Tang W D, Liu C, et al. Suppression of magnetotelluric square wave noise based on a LSTM recurrent neural network[J]. Geophysical and Geochemical Exploration, 2022, 46(4):925-933.
[12]	Li J F, Liu Y H, Yin C C, et al. Fast imaging of time-domain airborne EM data using deep learning technology[J]. Geophysics, 2020, 85(5):E163-E170.
[13]	Liu W, Wang H, Xi Z Z, et al. Physics-driven deep learning inversion with application to magnetotelluric[J]. Remote Sensing, 2022, 14(13):3218.
[14]	Li J, Liu Y C, Tang J T, et al. Magnetotelluric data denoising method combining two deep-learning-based models[J]. Geophysics, 2023, 88(1):E13-E28.
[15]	Liao X L, Shi Z Y, Zhang Z H, et al. 2D inversion of magnetotelluric data using deep learning technology[J]. Acta Geophysica, 2022, 70(3):1047-1060.
[16]	陈乐寿, 王光锷. 大地电磁测深法[M]. 北京: 地质出版社,1990.
[16]	Chen L S, Wang G E. Magnetotellurics[M]. Beijing: Geological Publishing House,1990.
[17]	范振宇. 基于卷积神经网络的大地电磁深度学习反演研究[D]. 北京: 中国地质大学(北京), 2020.
[17]	Fan Z Y. Magnetotelluric deep learning inversion based on convolutional neural network[D]. Beijing: China University of Geosciences, 2020.
[18]	冯兵, 李建国, 赵斌, 等. 音频大地电磁法在南岭于都—赣县矿集区银坑示范区深部矿产资源探测中的应用[J]. 地质学报, 2014, 88(4):669-675.
[18]	Feng B, Li J G, Zhao B, et al. The application of audio magnetotelluric method(AMT)in Nanling yudu-Gan county ore-concentrated area Yinkeng demonstration plot to survey deep mineral resources[J]. Acta Geologica Sinica, 2014, 88(4):669-675.