3D traveltime calculation of first arrival waveusing physics-informed neural network
DU Guoning1, TAN Jun1,2,3, SONG Peng1,2,3, XIE Chuang1, WANG Shaowen1
1. College of Marine Geosciences, Ocean University of China, Qingdao, Shandong 266100, China; 2. Qingdao National Laboratory for Marine Science and Technology, Qingdao, Shandong 266100, China; 3. Key Lab of Submarine Geosciences and Prospecting Techniques, Ministry of Education, Qingdao, Shandong 266100, China
Abstract:In seismic exploration, accurate calculation of the traveltime of the first arrival wave is an important basis for processing techniques such as migration imaging and traveltime inversion. The finite-difference algorithms based on an eikonal equation have shown excellent effect in solving seismic wave traveltime. However, It requires huge computational consumption, especially for calculating the traveltime of multiple-source and high-density grids. This paper develops an algorithm with a 3D eikonal equation for calculating traveltime based on a physics-informed neural network (PINN). Specifically, the algorithm trains the neural network by minimizing a loss function composed of 3D eikonal equation and other physical conditions information and finally outputs a traveltime satisfying the eikonal equation. Numerical simulation experiments based on different velocity models show that the proposed method has higher computational accuracy and efficiency than traditional finite-difference algorithms.
LAWTON D C.Computation of refraction static corrections using first-break traveltime differences[J].Geophysics,1989,54(10):1289-1296.
[2]
SCHIJNS H,HEINONEN S,SCHMITT D R,et al.Seismic refraction traveltime inversion for static corrections in a glaciated shield rock environment:a case study[J].Geophysical Prospecting,2009,57(6):997-1008.
[3]
ALKHALIFAH T.Efficient traveltime compression for 3D prestack Kirchhoff migration[J].Geophysical Prospecting,2011,59(1):1-9.
[4]
MAUFROY E,GAFFET S,OPERTO S,et al.Travel time inversion from ground level to gallery:protocol for the characterization of P-wave seismic signature in a fractured-porous Urgonian platform at hectometric scale[J].Near Surface Geophysics,2014,12(6):697-708.
[5]
ALVARO R,AKHIL D G,BALLIN P.Assisted history matching in the presence of frequent well intervention using generalize travel time inversion[J].Journal of Petroleum Science and Engineering,2011,78(2):415-430.
[6]
ZHAO C J,ZHANG L L,YU P,et al.Combined inver-sion of first-arrival travel times and reflection travel times[J].Geophysical Prospecting,2019,67(7):1764-1777.
[7]
NASR M,GIROUX B,DUPUIS J C.A hybrid approach to compute seismic travel times in three-dimensional tetrahedral meshes[J].Geophysical Prospecting,2020,68(4):1291-1313.
[8]
JIANG W B,ZHANG J.First-arrival traveltime tomography with modified total-variation regularization[J].Geophysical Prospecting,2017,65(5):1138-1154.
[9]
GIROUX B,GLOAGUEN E.Geostatistical traveltime tomography in elliptically anisotropic media[J].Geophysical Prospecting,2012,60(6):1133-1149.
[10]
GRECHKA V,DE LA PENA A,SCHISSELÉ-REBEL E,et al.Relative location of microseismicity[J].Geophysics,2015,80(6):WC1-WC9.
[11]
LOEWENTHAL D,HU L Z.Two methods for computing the imaging condition for common-shot pre-stack migration[J].Geophysics,1991,56(3):378-381.
[12]
SAMBRIDGE M S,KENNETT B L N.Boundary va-lue ray tracing in a heterogeneous medium:a simple and versatile algorithm[J].Geophysical Journal International,1990,101(1):157-168.
[13]
ERVENY V,PENIK I.Gaussian beams and paraxial ray approximation in three-dimensional elastic inhomogeneous media[J].Journal of Geophysics,1983,53(1):1-15.
[14]
COMAN R,GAJEWSKI D.Traveltime computation by wavefront-orientated ray tracing[J].Geophysical Prospecting,2005,53(1):23-36.
[15]
JULIAN B R,GUBBINS D.Three-dimensional seismic ray tracing[J].Journal of Geophysics,1977,43(1):95-113.
[16]
UM J,THURBER C.A fast algorithm for two-point seismic ray tracing[J].Bulletin of the Seismological Society of America,1987,77(3):972-986.
[17]
VIDALE J.Finite-difference calculation of travel times[J].Bulletin of the Seismological Society of America,1988,78(6):2062-2076.
[18]
QIN F H,LUO Y,OLSEN K B,et al.Finite-difference solution of the eikonal equation along expanding wavefronts[J].Geophysics,1992,57(3):478-487.
[19]
VAN TRIER J,SYMES W W.Upwind finite-differen-ce calculation of traveltimes[J].Geophysics,1991,56(6):812-821.
[20]
SETHIAN J A.Level Set Methods and Fast Mar-ching Methods:Evolving Interfaces in Computational Geometry,Fluid Mechanics,Computer Vision,and Materials Science[M].2nd ed.Cambridge University Press,Cambridge,1999.
[21]
QIAN J L,ZHANG Y T,ZHAO H K.Fast sweeping methods for eikonal equations on triangular meshes[J].SIAM Journal on Numerical Analysis,2007,45(1):83-107.
[22]
FOMEL S,LUO S T,ZHAO H K.Fast sweeping method for the factored eikonal equation[J].Journal of Computational Physics,2009,228(17):6440-6455.
[23]
LUO S T,QIAN J L.Factored singularities and high-order Lax-Friedrichs sweeping schemes for point-source traveltimes and amplitudes[J].Journal of Computational Physics,2011,230(12):4742-4755.
[24]
VIDALE J E.Finite-difference calculation of travel times in three dimensions[J].Geophysics,1990,55(5):521-526.
[25]
SOUKINA S M,GAJEWSKI D,KASHTAN B M.Traveltime computation for 3D anisotropic media by a finite-difference perturbation method[J].Geophysical Prospecting,2003,51(5):431-441.
孙章庆,孙建国,岳玉波,等.基于快速推进迎风双线性插值法的三维地震波走时计算[J].地球物理学报,2015,58(6):2011-2023.SUN Zhangqing,SUN Jianguo,YUE Yubo,et al.3D traveltime computation using fast marching upwind bilinear interpolation method[J].Chinese Journal of Geophysics,2015,58(6):2011-2023.
[28]
HU J T,CAO J X,WANG H Z,et al.3D traveltime computation for quasi-P-wave in orthorhombic media using dynamic programming[J].Geophysics,2018,83(1):C27-C35.
[29]
LAGARIS I E,LIKAS A,FOTIADIS D I.Artificial neural networks for solving ordinary and partial differential equations[J].IEEE Transactions on Neural Networks,1998,9(5):987-1000.
[30]
LEE H,KANG I S.Neural algorithm for solving differential equations[J].Journal of Computational Physics,1990,91(1):110-131.
[31]
SIRIGNANO J,SPILIOPOULOS K.DGM:a deep learning algorithm for solving partial differential equation[J].Journal of Computational Physics,2018,375:1339-1364.
[32]
TOMPSON J,SCHLACHTER K,SPRECHMANN P,et al.Accelerating eulerian fluid simulation with convolutional networks[C].Proceedings of the 34th International Conference on Machine Learning,2017,3424-3433.
[33]
RAISSI M,PERDIKARIS P,KARNIADAKIS G E.Physics-informed neural networks:a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J].Journal of Computational Physics,2019,378:686-707.
[34]
XU Y R,LI J Y,CHEN X H.Physics informed neural networks for velocity inversion[C].SEG Technical Program Expanded Abstracts,2019,38:2584-2588.
SONG C,ALKHALIFAH T,WAHEED U B.Solving the frequency-domain acoustic VTI wave equation using physics-informed neural networks[J].Geophysical Journal International,2021,225(2):846-859.
[37]
WAHEED U B,HAGHIGHAT E,ALKHALIFAH T,et al.PINNeik:eikonal solution using physics-informed neural networks[J].Computers & Geo-sciences,2021,155:104833.
[38]
HORNIK K,STINCHCOMBE M,WHITE H.Multilayer feedforward networks are universal approximators[J].Neural Networks,1989,2(5):359-366.
[39]
BISHOP C M.Pattern Recognition and Machine Learning[M].Springer,New York,2006.
[40]
JAGTAP A D,KAWAGUCHI K,EM KARNIADAKIS G.Locally adaptive activation functions with slope recovery for deep and physics-informed neural networks[J].Proceedings of the Royal Society A:Mathematical,Physical and Engineering Sciences,2020,476(2239):20200334.
[41]
WANG S F,YU X L,PERDIKARIS P.When and why PINNs fail to train:a neural tangent kernel perspective[J].Journal of Computational Physics,2022,449:110768.
[42]
SLOTNICK M M.Lessons in Seismic Computing[M].Society of Exploration Geophysicists,Tulsa,1959,178-196.