A5027915805- Hydrocarbon exploration and reservoir analysis
- Hydraulic Fracturing and Reservoir Analysis
- Reservoir Engineering and Simulation Methods
- NMR spectroscopy and applications
- Seismic Imaging and Inversion Techniques
- Image Processing and 3D Reconstruction
- Geophysical Methods and Applications
- Geological Modeling and Analysis
- Drilling and Well Engineering
- Generative Adversarial Networks and Image Synthesis
- Seismology and Earthquake Studies
- Advanced Data Processing Techniques
- Earthquake Detection and Analysis
- High-Velocity Impact and Material Behavior
- Enhanced Oil Recovery Techniques
- Anomaly Detection Techniques and Applications
- Technology and Security Systems
Changzhou University
2025
China University of Petroleum, Beijing
2021-2024
We propose a novel machine learning approach to improve the formation evaluation from logs by integrating petrophysical information with neural networks using loss function. The can either be specific logging response equations or abstract relationships between data and reservoir parameters. compare our method's performances two datasets evaluate influences of multi-task learning, model structure, transfer petrophysics informed (PIML). Our experiments demonstrate that PIML significantly...
As an important part of industrial production, the optimization circulating water systems is great significance for improving energy efficiency and reducing operating costs. However, traditional methods lack real-time dynamic adjustment capabilities often cannot fully cope with complex changeable environment demands. Advances in computer technology can enable people to use machine learning models process information data ultimately help simplify simulation optimization. In this paper, system...
Machine learning algorithms have become powerful tools for modeling in the engineering field. They are suitable solving problems that can't be effectively solved by traditional physical models or empirical due to complex relationship of variables. Since interpretation method log data is based on petrophysical mechanisms and models, many assumptions needed, which may lead deviations practical application. Therefore, it great significance achieve reservoir fluid identification when using...
We propose a novel machine learning approach to improve the formation evaluation from logs by integrating petrophysical information with neural networks using loss function. The can either be specific logging response equations or abstract relationships between data and reservoir parameters. compare our method’s performances two datasets evaluate influences of multi-task learning, model structure, transfer petrophysics informed (PIML). Our experiments demonstrate that PIML significantly...
Nuclear magnetic resonance (NMR) is a powerful tool in biomedical, chemical analysis, the oil industry, and other scientific fields. It provides information on molecular structure for analysis of dynamics interactions. In recent years, deep learning (DL) has attracted great interest various research fields because availability high-performance computing. The employment DL methods to effectively address shortcomings NMR data processing new field, such as signal reconstruction, MRI peak...
Forward and inverse models are the central common problems in well-logging evaluation. Logging problems, however, ill-posed, pathological, uncertain, which caused by problem itself a difficult to be faced interpretation of logging data. Manual log relies solely on geophysical mechanics data understand downhole conditions, complex variable. The subjectivity expert can have significant impact results, re-analysis set using new analysis processing tools, such as machine learning, could lead...
Summary Intelligent geophysical logging inversion based on data-driven machine learning can efficiently realize traditional interpretation and formation evaluation, has broad application prospects. Compared with commercial artificial intelligence, intelligent faces the problem of small samples low-quality labels. It is necessary to study representative labeled data sets for scenario logging. However, logging, observed downhole strata cannot be seen or touched, coupled multi-solution...
Summary The existing study single predict neural network were used, that is, for a one petrophysical parameter can be predicted, such as porosity (POR) or water saturation (SW) with set of logging data. When the tasks are related to each other, underlying information extracted from input features model has certain commonality. In this case, multi-task extract higher quality help multi-dimensional output information, so obtain better performance and produce an effect similar "information...
Summary Using neural network to predict reservoir parameters, we can map the relationship between logs and parameters as long building a suitable model have large number of training data. With help network, unknown physical without much geological expertise. The existing research parameter prediction with only focuses on one kind modeling, ignoring parameters. In this paper, relevance transfer learning is introduced, which using knowledge petrophysics improve performance prediction....