Geoenergy Science and Engineering, Год журнала: 2024, Номер unknown, С. 213550 - 213550
Опубликована: Ноя. 1, 2024
Язык: Английский
Geoenergy Science and Engineering, Год журнала: 2024, Номер unknown, С. 213550 - 213550
Опубликована: Ноя. 1, 2024
Язык: Английский
Nature Reviews Earth & Environment, Год журнала: 2023, Номер 4(5), С. 319 - 332
Опубликована: Май 2, 2023
Язык: Английский
Процитировано
120Nature Reviews Earth & Environment, Год журнала: 2023, Номер 4(8), С. 568 - 581
Опубликована: Июль 11, 2023
Язык: Английский
Процитировано
56Scientific Reports, Год журнала: 2024, Номер 14(1)
Опубликована: Март 12, 2024
Abstract Rate of penetration (ROP) is a key factor in drilling optimization, cost reduction and cycle shortening. Due to the systematicity, complexity uncertainty operations, however, it has always been problem establish highly accurate interpretable ROP prediction model guide optimize operations. To solve this Tarim Basin, study proposes four categories hybrid physics-machine learning (ML) methods for modeling. One which residual modeling, an ML learns predict errors or residuals, via physical model; second integrated coupling, output used as input third simple average, predictions from both are combined; last bootstrap aggregating (bagging), follows idea ensemble combine different models’ advantages. A total 5655 real data points Halahatang oil field were test performance various models. The results showed that modeling model, with R 2 0.9936, had best performance, followed by average bagging values 0.9394 0.5998, respectively. From view accuracy, interpretability, physics-ML optimal method prediction.
Язык: Английский
Процитировано
22Remote Sensing of Environment, Год журнала: 2024, Номер 303, С. 114001 - 114001
Опубликована: Фев. 2, 2024
Язык: Английский
Процитировано
7Remote Sensing of Environment, Год журнала: 2025, Номер 318, С. 114609 - 114609
Опубликована: Янв. 22, 2025
Язык: Английский
Процитировано
0Remote Sensing, Год журнала: 2025, Номер 17(4), С. 686 - 686
Опубликована: Фев. 18, 2025
Interferometric synthetic aperture radar (InSAR) technology plays a crucial role in monitoring surface deformation and has become widely used volcanic earthquake research. With the rapid advancement of satellite technology, InSAR now generates vast volumes data. Deep learning revolutionized data analysis, offering exceptional capabilities for processing large datasets. Leveraging these advancements, automatic detection from extensive datasets emerged as major research focus. In this paper, we first introduce several representative deep architectures commonly including convolutional neural networks (CNNs), recurrent (RNNs), generative adversarial (GANs), Transformer networks. Each architecture offers unique advantages addressing challenges We then systematically review recent progress identification signals images using techniques. This highlights two key aspects: design network methodologies constructing Finally, discuss propose potential solutions. study aims to provide comprehensive overview current applications extracting features, with particular focus on monitoring.
Язык: Английский
Процитировано
0Опубликована: Март 1, 2025
Язык: Английский
Процитировано
0Water Resources Management, Год журнала: 2025, Номер unknown
Опубликована: Март 24, 2025
Язык: Английский
Процитировано
0ISPRS Journal of Photogrammetry and Remote Sensing, Год журнала: 2025, Номер 225, С. 1 - 18
Опубликована: Апрель 24, 2025
Язык: Английский
Процитировано
0Science Bulletin, Год журнала: 2023, Номер 68(8), С. 773 - 774
Опубликована: Март 30, 2023
Язык: Английский
Процитировано
10