Lithology Identification of Uranium-Bearing Sand Bodies Using Logging Data Based on a BP Neural Network
Abstract
:1. Introduction
- ①
- The mathematical formulas of weight iteration are deduced in detail with examples in the geology domain, which provide theoretical support for geological staff to deeply understand the mathematical principle of a BP neural network model.
- ②
- An optimized gradient algorithm based on the AdaGrad method is proposed to improve the computational efficiency of the training model.
- ③
- The codes to realize the lithology prediction are self-written, and it is convenient to analyze the factors affecting the lithology classification by tracking the trend of the estimated parameters.
2. Methodology
2.1. Forward Propagation
2.2. Backward Propagation
2.3. Optimized Gradient Descent Algorithm
Optimized AdaGrad algorithm (O_AdaGrad) |
Global learning rate: η Initial parameter: w Small constant: ε Gradient cumulative variables: r While does not meet the stop criteria do n samples{x(1),…x(n)},the true value y(i) calculate gradient: g← Cumulative square gradient: r← Calculate update: Apply update: |
3. Example
3.1. Geologic Setting
3.1.1. Regional Geology
3.1.2. Deposit Geology
3.2. Data Characteristics
3.3. BP Neural Network Model for Lithology Identification
3.3.1. Model Building
3.3.2. Model Training
3.3.3. Prediction Validation
4. Conclusions
- 1.
- Through comparative analysis of logging data, the parameters of natural gamma (GP), density (DEN), natural potential (SP), resistivity (ρ), and well-diameter (Φ) are selected to establish the BP neural network used for the lithology identification of the Tarangaole uranium deposit. A total of 4578 sample data were selected from 8 boreholes to train the BP neural network model, and 599 samples for data verification.
- 2.
- The prediction accuracy rate using a BP neural network model reaches 88.31% in this contribution, supporting the effectiveness of the BP neural network model in predicting lithology in the studied area. The average sensitivity and the average specificity of the testing data are 74.83% and 92.23%, respectively.
- 3.
- The loss function value, the weight value, and the prediction accuracy under different iteration times in the prediction process are reported in detail, which is favorable to the objective evaluation of the influencing factors of data changes in the process of training and prediction.
Author Contributions
Funding
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Zhang, J.D.; Xu, G.Z.; Lin, J.R.; Peng, Y.B.; Wang, G. The implication of six kinds of new sandstone-type uranium deposits to uranium resources potential in North China. Geol. China 2010, 37, 1434–1449. [Google Scholar]
- Yao, W.S.; Wang, X.Q.; Zhang, B.M.; Xu, S.; Shen, W.; Du, X. Pilot study of deep-penetrating geochemical exploration for sandstone-type uranium deposit, Ordos Basin. Earth Sci. Front. 2012, 19, 167–176. [Google Scholar]
- Sun, D.; Jiang, M.; Luo, Y.; Jiang, X.; Gu, D. Comprehensive information prospecting model for sandstone uranium deposits in Teguida area, Niger. Geophys. Geochem. Explor. 2015, 39, 273–282. [Google Scholar]
- Jiang, C.; Zhang, D.; Chen, S. Lithology identification from well-log curves via neural networks with additional geologic constraint. Geophysics 2021, 86, M85–M100. [Google Scholar] [CrossRef]
- Hill, E.J.; Pearce, M.A.; Stromberg, J.M. Improving Automated Geological Logging of Drill Holes by Incorporating Multiscale Spatial Methods. Math. Geosci. 2021, 53, 21–53. [Google Scholar] [CrossRef] [Green Version]
- Wang, X.Y.; Xie, R.H.; Mao, Z.G.; Wu, B.H.; Xu, C.Y.; Wei, H.Y. Application research of principal component analysis method in lithology identification of tight sandstone. Nat. Gas Oil 2021, 39, 88–93. [Google Scholar]
- Zhao, Y. A new method for logging identification and evaluation of low-resistivity gas layers: A case study of the Dongsheng Gasfield in the Ordos Basin. Nat. Gas Ind. 2020, 40, 47–54. [Google Scholar]
- Tian, Y.K.; Zhou, H.; Yuan, S.Y. Lithologic discrimination method based on Markov random-field. Chin. J. Geophys. 2013, 56, 1360–1368. [Google Scholar]
- Huang, X.; Yi, X.; Jin, T.; Zhang, J.; Xu, B.; Li, F. Performance Evaluation of Digital Back-Propagation in the Presence of Power Excursion. IEEE Photonics Technol. Lett. 2021, 33, 1010–1013. [Google Scholar] [CrossRef]
- Badr, A. Awesome back-propagation machine learning paradigm. Neural Comput. Appl. 2021, 33, 13225–13249. [Google Scholar] [CrossRef]
- Nguyen, K.C.; Nguyen, C.T.; Nakagawa, M. Nom document digitalization by deep convolution neural networks. Pattern Recognit. Lett. 2020, 133, 8–16. [Google Scholar] [CrossRef]
- Liu, Y.F.; Liu, Y.; Zhu, H. Masked convolutional neural network for supervised learning problems. HHS Public Access 2020, 9, e290. [Google Scholar] [CrossRef]
- Xin, R.; Zhang, J.; Shao, Y. Complex Network Classification with Convolutional Neural Network. Tsinghua Sci. Technol. 2020, 25, 447–457. [Google Scholar] [CrossRef]
- Fang, T.; Chen, Z.; Fu, Y. Face recognition method based on neural network multi-layer feature information fusion. CAAI Trans. Intell. Syst. 2021, 16, 279–285. [Google Scholar]
- Chai, B.; Yang, L.; Wang, J.; Li, R. Text sentiment classification approach with integrated local and global prominent features. J. Hebei Univ. Nat. Sci. Ed. 2021, 41, 201–211. [Google Scholar]
- Qu, F.; Yang, H.T.; Chen, M. Identification of Asmari Formation Lithology in Fauqi Oilfield by BP Neural Network Based on Principal Component Analysis. Sino-Glob. Energy 2021, 26, 26–32. [Google Scholar]
- Tang, Y. Parameter selection and applicability of gas content logging interpretation methodology in coal seam. Coalgeolooy Explor. 2015, 43, 94–98. [Google Scholar]
- Xue, W.Z.; Chen, B.; Zhe, Z.H. Recognition of Stratigraphic Lithology by BP-Neural Network—A Case Study of Yiner Basin. Petrochem. Ind. Technol. 2019, 26, 103–107. [Google Scholar]
- Ma, J.; Yarats, D. Quasi-hyperbolic momentum and Adam for deep learning. arXiv 2018, arXiv:1810.06801. [Google Scholar]
- Keskar, N.S.; Socher, R. Improving Generalization Performance by Switching from Adam to SGD. arXiv 2017, arXiv:1712.07628. [Google Scholar]
- Ruder, S. An overview of gradient descent optimization algorithms. arXiv 2016, arXiv:1609.04747. [Google Scholar]
- Dauphin, Y.; Pascanu, R.; Gulcehre, C.; Cho, K.; Ganguli, S.; Bengio, Y. Identifying and attacking the saddle point problem in high-dimensional non-convex optimization. arXiv 2014, arXiv:1406.2572. [Google Scholar]
- Zhang, S.; Choromanska, A.; LeCun, Y. Deep learning with Elastic Averaging SGD. arXiv 2014, arXiv:1412.6651. [Google Scholar]
- Zeiler, M.D. Adadelta: An Adaptive Learning Rate Method. arXiv 2012, arXiv:1212.5701. [Google Scholar]
- Mahboubi, S.; Indrapriyadarsini, S.; Ninomiya, H.; Asai, H. Momentum Acceleration of Quasi-Newton Training for Neural Networks; Springer International Publishing: Cham, Switzerland, 2019; pp. 268–281. [Google Scholar]
- Robitaille, B.; Marcos, B.; Veillette, M.; Payre, G. Modified quasi-Newton methods for training neural networks. Comput. Chem. Eng. 1996, 20, 1133–1140. [Google Scholar] [CrossRef]
- Weinberger, N.; Shayevitz, O. On the Optimal Boolean Function for Prediction Under Quadratic Loss. IEEE Trans. Inf. Theory 2017, 63, 4202–4217. [Google Scholar] [CrossRef]
- Hajra, S. Making Convex Loss Functions Robust to Outliers using $e$-Exponentiated Transformation. arXiv 2019, arXiv:1902.06127. [Google Scholar]
- Xing, H.; Ji, M. Robust one-class support vector machine with rescaled hinge loss function. Pattern Recognit. 2018, 84, 152–164. [Google Scholar] [CrossRef]
- Zhou, Y.; Wang, X.; Zhang, M.; Zhu, J.; Zheng, R.; Wu, Q. MPCE: A Maximum Probability Based Cross-entropy Loss Function for Neural Network Classification. IEEE Access 2019, 7, 146331–146341. [Google Scholar] [CrossRef]
- Chen, L.; Chen, Y.; Feng, X.; Li, J.; Guo, H.; Miao, P.; Jin, R.; Tang, C.; Zhao, H.; Wang, G.; et al. Uranium occurrence state in the Tarangaole area of the Ordos Basin, China: Implications for enrichment and mineralization. Ore Geol. Rev. 2019, 115, 103034. [Google Scholar] [CrossRef]
NO. | Borehole | Depth (m) | Density (g/cm3) | Resistivity (Ω·m) | Natural Gamma (API) | Gamma Ray ((nc/kg·h) | Well-Diameter (cm) | Spontaneous Potential (mv) |
---|---|---|---|---|---|---|---|---|
1 | ZKC232-87 | 470.05 | 2 | 12.56 | 128.57 | 16.78 | 120.9 | −515.07 |
2 | 470.1 | 4.04 | 12.26 | 123.82 | 19.28 | 120.9 | −515.4 | |
3 | 470.15 | 1.57 | 12.34 | 125.4 | 22.15 | 120.9 | −516.4 | |
4 | 470.2 | 3.21 | 12.23 | 117.86 | 21.43 | 120.9 | −516.92 | |
5 | 470.25 | 3.87 | 12.24 | 117.67 | 25.72 | 120.9 | −516.76 | |
… | … | … | … | … | … | … | … | |
599 | 499.9 | 1.01 | 14.95 | 185.23 | 28.92 | 111.46 | −514.84 | |
600 | 499.95 | 1.51 | 17.62 | 182.62 | 30.71 | 111.46 | −514.51 | |
601 | ZKC232-91 | 470.05 | 2.22 | 12.46 | 104.17 | 20 | 109.55 | −193.17 |
602 | 470.1 | 2.22 | 12.54 | 103.52 | 18.81 | 109.55 | −193.46 | |
603 | 470.15 | 2.22 | 12.57 | 101.37 | 18.81 | 109.54 | −193.76 | |
604 | 470.2 | 2.22 | 12.56 | 100.19 | 17.86 | 109.54 | −193.99 | |
605 | 470.25 | 2.22 | 12.55 | 98.02 | 17.86 | 109.57 | −194.1 | |
… | … | … | … | … | … | … | … | … |
1198 | ZKC232-91 | 499.9 | 2.09 | 13.51 | 411.67 | 79.07 | 104.04 | −198.74 |
1199 | 499.95 | 2.1 | 13.63 | 401.76 | 78.95 | 104.04 | −198.88 | |
4040 | ZKC264-87 | 470.05 | 2.18 | 16.22 | 317.06 | 48.94 | 112.46 | −472.2 |
4041 | 470.1 | 2.18 | 16.05 | 355.55 | 47.15 | 112.46 | −472.25 | |
4042 | 470.15 | 2.18 | 16.05 | 348.81 | 59.3 | 112.46 | −472.15 | |
4043 | 470.2 | 2.18 | 16.11 | 361.11 | 58.58 | 112.5 | −472.16 | |
4044 | 470.25 | 2.17 | 16.44 | 380.86 | 57.15 | 112.48 | −472.23 | |
4045 | 470.3 | 2.16 | 16.36 | 377.98 | 64.31 | 112.46 | −472.03 | |
… | … | … | … | … | … | … | … | … |
4577 | ZKC264-87 | 496.95 | 1.87 | 52.1 | 127.05 | 15.43 | 133.35 | −182.92 |
4578 | 497 | 1.72 | 73 | 62.23 | 20.36 | 133.35 | −305.31 |
Stratum | Conglomerate (%) | Coarse-Grained Sandstone (%) | Medium-Grained Sandstone (%) | Fine-Grained Sandstone (%) | Mudstone (%) | Calcareous Sandstone (%) | Number of Samples |
---|---|---|---|---|---|---|---|
J2z1 | 13.70 | 36.98 | 38.14 | 8.24 | 1.83 | 1.11 | 4578 |
Lithology | Code | Label |
---|---|---|
Conglomerate | CG | 1 |
Coarse-grained sandstone | CS | 2 |
Medium-grained sandstone | MS | 3 |
Fine-grained sandstone | FS | 4 |
Mudstone | MD | 5 |
Calcareous sandstone | KS | 6 |
True | CG | CS | MS | FS | MD | Total |
---|---|---|---|---|---|---|
Prediction | ||||||
CG | 26 | 0 | 0 | 0 | 0 | 26 |
CS | 10 | 40 | 2 | 0 | 0 | 52 |
MS | 0 | 4 | 391 | 25 | 4 | 424 |
FS | 0 | 0 | 20 | 55 | 5 | 80 |
MD | 0 | 0 | 1 | 6 | 10 | 17 |
Sensitivity | 72.22% | 90.91% | 94.44% | 63.95% | 52.63% | 74.83% |
Specificity | 100.00% | 99.63% | 81.72% | 79.81% | 100.00% | 92.23% |
Precision | 100.00% | 76.92% | 92.22% | 68.75% | 58.82% | 79.34% |
Accuracy | 88.31% |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Sun, Y.; Chen, J.; Yan, P.; Zhong, J.; Sun, Y.; Jin, X. Lithology Identification of Uranium-Bearing Sand Bodies Using Logging Data Based on a BP Neural Network. Minerals 2022, 12, 546. https://doi.org/10.3390/min12050546
Sun Y, Chen J, Yan P, Zhong J, Sun Y, Jin X. Lithology Identification of Uranium-Bearing Sand Bodies Using Logging Data Based on a BP Neural Network. Minerals. 2022; 12(5):546. https://doi.org/10.3390/min12050546
Chicago/Turabian StyleSun, Yuanqiang, Jianping Chen, Pengbing Yan, Jun Zhong, Yuxin Sun, and Xinyu Jin. 2022. "Lithology Identification of Uranium-Bearing Sand Bodies Using Logging Data Based on a BP Neural Network" Minerals 12, no. 5: 546. https://doi.org/10.3390/min12050546