High resolution velocity model estimation from refraction and reflection data

Forgues, Éric; Scala, Emma; Pratt, R. G.

doi:10.1190/1.1820111

Cited by 15 publications

(21 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although the reflector inversion is not explicitly restricted to recovering only high wavenumbers, high vertical wavenumbers dominate the model updates as a consequence of the survey geometry. The same gradient method algorithm recovers a wide range of wavenumber information when applied to crosshole and offset vertical seismic profiling (VSP) data (Pratt, 1990(Pratt, , 1999Song et al, 1995;Pratt and Shipp, 1999) or to wide-angle (refraction) data (Pratt et al, 1996;Brittan et al, 1997;Forgues et al, 1998).…”

Section: Gradient Method High-wavenumber Inversionmentioning

confidence: 98%

Reflection waveform inversion using local descent methods: Estimating attenuation and velocity over a gas‐sand deposit

2001

Self Cite

View full text Add to dashboard Cite

Prestack seismic reflection data contain amplitudes, traveltimes, and moveout information; waveform inversion of such data has the potential to estimate attenuation levels, reflector depths and geometry, and background velocities. However, when inverting reflection data, strong nonlinearities can cause reflectors to be incorrectly imaged and can prevent background velocities from being updated. To successfully recover background velocities, previous authors have resorted to nonlinear, global search inversion techniques.We propose a two-step inversion procedure using local descent methods in which we perform alternate inversions for the reflectors and the background velocities. For our reflector inversion we exploit the efficiency of the back-propagation method when inverting for a large parameter set. For our background velocity inversion we use Newton inverse methods. During the background velocity inversions it is crucial to adaptively depth-stretch the model to preserve the vertical traveltimes. This reduces nonlinearity by largely decoupling the effects of the background velocities and reflectors on the data. Nonlinearity is further reduced by choosing to invert for slownesses and by inverting for a sparse parameter set which is partially defined using geological reflector picks.Applying our approach to shallow seismic data from the North Sea collected over a gas-sand deposit, we demonstrate that the proposed method is able to estimate both the geometry and internal velocity of a significant velocity structure not present in the initial model. Over successive iterations, the use of adaptive depth stretching corrects the pull-down of the base of the gas sand. Vertical background velocity gradients are also resolved. For an insignificant extra cost the acoustic attenuation parameter Q is included in the inversion scheme. The final attenuation tomogram contains realistic values of Q for the expected lithologies and for the effect of partial fluid saturation associated with a shallow bright spot. The attenuation image may also indicate the presence of fracturing.

show abstract

Section: Gradient Method High-wavenumber Inversionmentioning

confidence: 98%

Reflection waveform inversion using local descent methods: Estimating attenuation and velocity over a gas‐sand deposit

2001

Self Cite

View full text Add to dashboard Cite

show abstract

“…• at the near-surface scale (500 m depth) in application to localization of water resources [16]; • at the exploration scale (10 km depth) in application to localization of oil and gas, storage or reservoir monitoring [24,29,51,55]; • at the lithospheric scale (100 km depth), in application to fault slip reconstruction and earthquake interpretation [8,7,52,21] Imaging at very shallow near-surface scale (1 m depth) are also currently investigated [62]. The reconstruction of deep Earth structure at the global scale (10000 km scale) could also be considered.…”

Section: The Fwi Methodmentioning

confidence: 99%

Full Waveform Inversion and the Truncated Newton Method

Métivier¹,

Brossier²,

Virieux³

et al. 2013

SIAM J. Sci. Comput.

238

147

View full text Add to dashboard Cite

Abstract. Full Waveform Inversion (FWI) is a powerful method for reconstructing subsurface parameters from local measurements of the seismic wavefield. This method consists in minimizing a distance between predicted and recorded data. The predicted data is computed as the solution of a wave propagation problem. Conventional numerical methods for the resolution of FWI problems are gradient-based methods, such as the preconditioned steepest-descent, or more recently the l-BFGS quasi-Newton algorithm. In this study, we investigate the interest of applying a truncated Newton method to FWI. The inverse Hessian operator plays a crucial role in the parameter reconstruction. The truncated Newton method allows one to better account for this operator. This method is based on the computation of the Newton descent direction by solving the corresponding linear system through an iterative procedure such as the conjugate gradient method. The large-scale nature of FWI problems requires however to carefully implement this method to avoid prohibitive computational costs. First, this requires to work in a matrix-free formalism, and the capability of computing efficiently Hessian-vector products. To this purpose, we propose general second-order adjoint state formulas. Second, special attention must be payed to define the stopping criterion for the inner linear iterations associated with the computation of the Newton descent direction. We propose several possibilities and establish a theoretical link between the Steihaug-Toint method, based on trust-regions, and the Eisenstat stopping criterion, designed for method globalized by linesearch. We investigate the application of the truncated Newton method to two test cases: the first is a standard test case in seismic imaging based on the Marmousi II model. The second one is inspired by a nearsurface imaging problem for the reconstruction of high velocity structures. In the latter case, we demonstrate that the presence of large amplitude multi-scattered waves prevents standard methods from converging while the truncated Newton method provides more reliable results.

show abstract

“…Both approaches ͑using all of the frequency data components͒ are computationally expensive ͑Gauthier et al, 1986͒. There is data redundancy when we invert all frequency data components in the frequency-domain approach ͑or, equivalently, in the time-domain approach͒. Pratt and Worthington ͑1988͒, Liao andMcMechan ͑1996͒, andForgues et al ͑1998͒ point out that by inverting data at a very limited but select number of frequencies, one can produce an unaliased image at a reasonable computational expense. Furthermore, frequency-domain methods are especially preferable when a significant number of sources are involved ͑Marfurt, 1984͒ because, at least in the 2D configuration, the multisource forward solutions can be calculated at nearly no extra cost.…”

Section: Introductionmentioning

confidence: 97%

Simultaneous multifrequency inversion of full-waveform seismic data

2009

View full text Add to dashboard Cite

We present a simultaneous multifrequency inversion approach for seismic data interpretation. This algorithm inverts all frequency data components simultaneously. A dataweighting scheme balances the contributions from different frequency data components so the inversion process does not become dominated by high-frequency data components, which produce a velocity image with many artifacts. A Gauss-Newton minimization approach achieves a high convergence rate and an accurate reconstructed velocity image. By introducing a modified adjoint formulation, we can calculate the Jacobian matrix efficiently, allowing the material properties in the perfectly matched layers ͑PMLs͒ to be updated automatically during the inversion process. This feature ensures the correct behavior of the inversion and implies that the algorithm is appropriate for realistic applications where a priori information of the background medium is unavailable. Two different regularization schemes, an L 2 -norm and a weighted L 2 -norm function, are used in this algorithm for smooth profiles and profiles with sharp boundaries, respectively. The regularization parameter is determined automatically and adaptively by the so-called multiplicative regularization technique. To test the algorithm, we implement the inversion to reconstruct the Marmousi velocity model using synthetic data generated by the finite-difference time-domain code. These numerical simulation results indicate that this inversion algorithm is robust in terms of starting model and noise suppression. Under some circumstances, it is more robust than a traditional sequential inversion approach.

show abstract

High resolution velocity model estimation from refraction and reflection data

Cited by 15 publications

References 0 publications

Reflection waveform inversion using local descent methods: Estimating attenuation and velocity over a gas‐sand deposit

Reflection waveform inversion using local descent methods: Estimating attenuation and velocity over a gas‐sand deposit

Full Waveform Inversion and the Truncated Newton Method

Simultaneous multifrequency inversion of full-waveform seismic data

Contact Info

Product

Resources

About