Article
Geosciences, Multidisciplinary
Melckzedeck M. Mgimba, Shu Jiang, Edwin E. Nyakilla, Grant Charles Mwakipunda
Summary: In this study, the GMDH technique was employed for the first time to predict formation pore pressures from well logs data in the Nanye 1 well. The results showed that the GMDH technique outperformed other machine learning techniques, with the lowest RMSE of 0.0308 MPa and a high coefficient of determination of 0.998. Furthermore, the GMDH technique was able to overcome challenges faced by other techniques by identifying data structure and automatically selecting relevant input data.
NATURAL RESOURCES RESEARCH
(2023)
Article
Chemistry, Multidisciplinary
Shuiqing Hu, Haowei Zhang, Rongji Zhang, Lingxuan Jin, Yuming Liu
Summary: Total organic carbon (TOC) is the basis of source rock evaluation, and the Majiagou formation in the Ordos Basin has complex lithology and low organic matter content. The introduction of a neural network model in TOC logging interpretation shows good predictive effects in complicated lithologic regions.
APPLIED SCIENCES-BASEL
(2021)
Article
Energy & Fuels
Richa, S. P. Maurya, Kumar H. Singh, Raghav Singh, Rohtash Kumar, Prabodh Kumar Kushwaha
Summary: Seismic inversion is a geophysical technique used to estimate subsurface rock properties by combining well log information and seismic data to extract high-resolution subsurface acoustic impedance. By analyzing well log and seismic data, potential oil and gas fields can be identified, and subsurface reservoir characteristics can be evaluated.
JOURNAL OF PETROLEUM EXPLORATION AND PRODUCTION TECHNOLOGY
(2022)
Article
Chemistry, Multidisciplinary
Gulbahar Yazmyradova, Nik Nur Anis Amalina Nik Mohd Hassan, Nur Farhana Salleh, Maman Hermana, Hassan Soleimani
Summary: The integrated approach of rock physics analysis, pre-stack seismic inversion, and artificial neural network successfully delineates the hydrocarbon potential in HPHT zone of the Malay Basin, recognizing and predicting reservoir rock and fluid properties.
APPLIED SCIENCES-BASEL
(2021)
Article
Energy & Fuels
Mohammad Ali Riahi, Mohammad Ghasem Fakhari
Summary: This study discusses the use of direct acoustic impedance for pore pressure estimation in a famous Iranian carbonate reservoir, which yielded acceptable results compared to traditional methods. The results revealed overpressure zones and pressure reduction areas, providing significant implications for determining reservoir boundaries and planning drilling trajectories for new wells.
JOURNAL OF PETROLEUM EXPLORATION AND PRODUCTION TECHNOLOGY
(2022)
Article
Engineering, Environmental
Abir Banerjee, Rima Chatterjee
Summary: This study successfully mapped the pore pressure distribution in the Raniganj basin in India using sonic log and seismic data. A model-based seismic inversion technique was used, combined with estimated pore pressure from sonic logs, to accurately determine the sub-surface pore pressure. The analysis also provided insights into the stress regime and stability of existing faults, helping to mitigate drilling risks and provide important inputs for future geomechanical studies.
BULLETIN OF ENGINEERING GEOLOGY AND THE ENVIRONMENT
(2022)
Article
Energy & Fuels
N. P. Szabo, F. Remeczki, A. Jobbik, K. Kiss, M. Dobroka
Summary: Innovative interpretation technologies are needed for the evaluation of Hungarian tight gas formations due to diverse geological environment and heterogeneous dataset. A new inversion methodology, combined with laboratory measurements, shows promise in estimating petrophysical parameters accurately. The joint analysis of well logs and core data using this approach may reveal further potential tight gas resources in the studied area.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2022)
Article
Environmental Sciences
Liurong Tao, Haoran Ren, Zhiwei Gu
Summary: Seismic impedance inversion is an important technique for geological interpretation and reservoir investigation. This research utilized a self-attention U-Net trained with seismic profiles and background impedance to obtain acoustic impedance profiles, showing robustness to noise and superior spatial continuity compared to conventional methods. The quality of predicted profiles was evaluated using various indexes. The proposed method was also applied to field data without labels, demonstrating its generalization capability from synthetic to field data.
Article
Geochemistry & Geophysics
Hongzhou Wang, Jun Lin, Xintong Dong, Shaoping Lu, Yue Li, Baojun Yang
Summary: Velocity model inversion is a challenging task in seismic exploration, and accurate velocity models are crucial for high-resolution seismic imaging. CNN-based velocity inversion methods have shown remarkable performance but are limited by their inability to capture long-range dependence and time-varying properties. To address these limitations, we propose SVIT, a DL framework based on Transformer, which utilizes self-attention mechanism to capture long-range dependencies in seismic data. SVIT outperforms conventional full-waveform inversion and existing CNN-based methods in terms of velocity estimation, subsurface structures, and geologic interfaces.
Article
Geochemistry & Geophysics
Xinpeng Pan, Zhishun Liu, Pu Wang, Ying Zheng, Lei Li, Xun Wang, Zhenwei Guo, Jianxin Liu
Summary: Horizontally transverse isotropy (HTI) induced by vertical or subvertical aligned fractures is common in unconventional fractured shale oil or gas reservoirs. Understanding fracture properties and in situ stresses is essential for optimizing well planning, hydraulic fracturing, and seismic inversion in these types of reservoirs.
Article
Geochemistry & Geophysics
Li Bonan, Si Wenpeng, Shen Hui, Zhang Xuebing
Summary: The article proposed a data-based description scheme for characterizing carbonate pore types by using virtual samples and supervised machine learning algorithm to improve sample richness and enhance identification accuracy. The radial basis function support vector machine trained with blended dataset in high dimensions accurately separated different carbonate pore systems.
Article
Geosciences, Multidisciplinary
Simona Gabrielli, Aybige Akinci, Luca De Siena, Edoardo Del Pezzo, Mauro Buttinelli, Francesco Emanuele Maesano, Roberta Maffucci
Summary: Deep fluid circulation is likely responsible for the large extensional events of the 2016-2017 Central Italy seismic sequence. This study demonstrates the use of peak delays, a proxy for scattering attenuation, in mapping thrusts and sedimentary structures and their influence on fluid overpressure and release. The results highlight the control of thrusts and paleogeography on the sequence and suggest the monitoring potential of this technique for seismic hazard assessment.
GEOPHYSICAL RESEARCH LETTERS
(2023)
Article
Geochemistry & Geophysics
Jian Sun, Kristopher Innanen, Tianze Zhang, Daniel Trad
Summary: Full waveform inversion (FWI) is a state-of-the-art method for imaging subsurface structures and physical parameters with seismic data, but it faces challenges in implementation and use. The implicit full waveform inversion (IFWI) algorithm, designed with deep neural representations, shows improved convergence and the ability to capture high-resolution subsurface structures. Although uncertainty analysis is not fully solved, IFWI addresses it meaningfully by approximating Bayesian inference. Numerical experimentation suggests that IFWI has a strong capacity for generalization and is suitable for multi-scale joint geophysical inversion.
JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH
(2023)
Article
Geochemistry & Geophysics
Xinyu Li, Yaojun Wang, Yu Liu, Hanpeng Cai
Summary: The paper presents a multichannel seismic impedance inversion method that combines logging and seismic data. By using dictionary learning method and sparse representation technology to add vertical and transverse distribution prior information simultaneously, this method solves the issue of inaccurately describing complex reservoir information in traditional methods and improves the efficiency of seismic inversion.
Article
Geosciences, Multidisciplinary
Chaoyang Hu, Fengjiao Wang, Chi Ai
Summary: This study utilizes surface displacement to calculate changes in reservoir pore pressure, implementing an improved CNN and machine learning for pressure inversion. The proposed method shows promising results in oilfield experiments.
FRONTIERS IN EARTH SCIENCE
(2021)