Bayesian deep networks for absolute permeability and porosity uncertainty prediction from image borehole logs from brazilian carbonate reservoirs

Valentin, Manuel Blanco; Fraga, Bernardo Machado de Oliveira; Campos, Jorge; Coutinho, Bernardo; Dias, Luciana Olivia; Faria, Elisângela L.; Albuquerque, Márcio P. de; Albuquerque, Marcelo P. de; Correia, Maury Duarte

doi:10.1016/j.petrol.2021.108361

Cited by 13 publications

(1 citation statement)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, machine learning (ML), especially deep learning (DL) techniques, started to be intensively explored in geophysical problems for a wide variety of scales, instruments and data (see, for instance, Yu & Ma, 2021;Bom et al, 2021;Dias et al, 2020), such as automatic detection of faults in seismic images (Araya-Polo et al, 2017), automatic detection of salt bodies in seismic images (Shi et al, 2019;Sen et al, 2020), seismic image denoising and resolution enhancement (Wang & Nealon, 2019;Bugge et al, 2021) and, seismic facies segmentation (Civitarese et al, 2019). One particular set of problems that attracted attention is the one related to VMB (AlAli & Deep-tomography 3 Anifowose, 2021), from automated NMO velocity picking (Smith, 2017;Park & Sacchi, 2020) to FWI assisted by ML algorithms (Zhang & Alkhalifah, 2019;He & Wang, 2021), until the complete model estimation (Biswas et al, 2019;Araya-Polo et al, 2018Yang & Ma, 2019;Li et al, 2020;Fabien-Ouellet & Sarkar, 2020;Kazei et al, 2020;Geng et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

Deep-tomography: iterative velocity model building with deep learning

Muller,

Bom,

Costa

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

The accurate and fast estimation of velocity models is crucial in seismic imaging. Conventional methods, like Tomography and Full-Waveform Inversion (FWI), obtain appropriate velocity models; however, they require intense and specialized human supervision and consume much time and computational resources. In recent years, some works investigated deep learning(DL) algorithms to obtain the velocity model directly from shots or migrated angle panels, obtaining encouraging predictions of synthetic models. This paper proposes a new flow to increase the complexity of velocity models recovered with DL. Inspired by the conventional geophysical velocity model building methods, instead of predicting the entire model in one step, we predict the velocity model iteratively. We implement the iterative nature of the process when, for each iteration, we train the DL algorithm to determine the velocity model with a certain level of

show abstract