3-D Bayesian variational full waveform inversion
نویسندگان
چکیده
Seismic full-waveform inversion (FWI) provides high resolution images of the subsurface by exploiting information in recorded seismic waveforms. This is achieved solving a highly nonnlinear and nonunique inverse problem. Bayesian inference therefore used to quantify uncertainties solution. Variational method that probabilistic, solutions efficiently using optimization. The has been applied 2D FWI problems produce full posterior distributions. However, due higher dimensionality more expensive computational cost, performance 3D remains unknown. We apply three variational methods analyse their performance. Specifically we automatic differential (ADVI), Stein gradient descent (SVGD) stochastic SVGD (sSVGD), problem, compare results cost. show ADVI most computationally efficient but systematically underestimates uncertainty. can be provide relatively rapid approximate insights into together with lower bound estimate demands highest still produces biased results. In contrast, including randomized term dynamics, sSVGD becomes Markov chain Monte Carlo accurate at intermediate thus conclude practically applicable, least small problems, image Earth's interior reasonable uncertainty estimates on those images.
منابع مشابه
Image-guided full waveform inversion
Multiple problems, including high computational cost, spurious local minima, and solutions with no geologic sense, have prevented widespread application of full waveform inversion (FWI), especially FWI of seismic reflections. These problems are fundamentally related to a large number of model parameters and to the absence of low frequencies in recorded seismograms. Instead of inverting for all ...
متن کامل3D acoustic frequency-domain full-waveform inversion
We present one of the first attempt at implementing a massively parallel frequency-domain full-waveform inversion algorithm for imaging 3D acoustic media. The inverse method is based on a classic steepest-descent algorithm. The algorithm was designed so that one or several frequencies are inverted at a time. Wave propagation modeling, a key component of the inversion algorithm, is performed wit...
متن کاملA projected Hessian for full waveform inversion
A Hessian matrix in full waveform inversion (FWI) is difficult to compute directly because of high computational cost and an especially large memory requirement. Therefore, Newton-like methods are rarely feasible in realistic largesize FWI problems. We modify the BFGS method to use a projected Hessian matrix that reduces both the computational cost and memory required, thereby making a quasiNew...
متن کاملFull waveform inversion with image-guided gradient
The objective of seismic full waveform inversion (FWI) is to estimate a model of the subsurface that minimizes the difference between recorded seismic data and synthetic data simulated for that model. Although FWI can yield accurate and high-resolution models, multiple problems have prevented widespread application of this technique in practice. First, FWI is computationally intensive, in part ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Geophysical Journal International
سال: 2023
ISSN: ['1365-246X', '0956-540X']
DOI: https://doi.org/10.1093/gji/ggad057