A data-driven method for intelligent lithology identification in coal mines based on drilling parameters

A novel integrated strategy of easy pruning, parameter searching, and re-parameterization for lightweight intelligent lithology identification

Integrated lithology identification based on images and elemental data from rocks

It is important to accurately master the downhole drilling parameter information and drilling formation lithology for the sake of efficient and reasonable gas extraction, hole arrangement and construction safety. According to different detection information, an identification & interpretation model for drilling engineering parameters and a lithology identification & interpretation model in while-drilling azimuthal Gamma logging were established. Next, the drilling conditions were judged and the lithology was discriminated. However, this method relied heavily upon the professional quality and working experience of workers. By numerically simulating the influence laws of azimuthal Gamma logging on the amplitude values of detection data under different borehole sizes, a more suitable method was chosen and applied to the automatic lithology explanation and identification during the logging. An object identification method combining deep convolutional neural network (CNN) and multi-weight multi-task learning mechanism was established, followed by the residual network design, in an effort to elevate the network training rate.

Lithology identification using kernel Fisher discriminant analysis with well logs

Subsurface petrophysical properties usually differ between different reservoirs, which affects lithology identification, especially for unconventional reservoirs. Thus, the lithology identification of subsurface reservoirs is a challenging task. Machine learning can be regarded as an effective method for using existing data for lithology prediction. By combining the hidden Markov model and random forests, we have adopted a novel method for lithology identification. The hidden Markov model provides a new hidden feature from elastic parameters, which is associated with unsupervised learning. Because elastic parameters are determined by petrophysical properties, the hidden feature may reveal an inner relationship of the petrophysical properties, which can expand the sample space. Then, with the new feature and the elastic parameters, the random forest method is adopted for lithology identification. In the prediction framework, the parameters of the hidden Markov model are updated until a satisfactory hidden feature is obtained. By analysis of synthetic and well-logging data, the superiority of the proposed method is demonstrated. Field seismic data application further proves the validity of the method. Numerical results show that the predicted lithology and shale content match well with real logging data.

ABSTRACT: Real-time identification of formation lithology is of great significance to optimization of drilling rate and directional drilling. However, the experimental observation method and the statistical analysis method require researchers with rich experience and professional knowledge, and the identification efficiency is low and the cost is high. The goal of this study is to employ machine learning for identifying lithology from only the real-time drilling parameters without any downhole measurements. First, the real-time drilling data were cleaned, and then the correlation between each drilling parameter and formation lithology was analyzed by correlation analysis algorithm, and 13 drilling parameters were selected as model inputs. And then random forest (RF) and XGBoost algorithms are used to develop lithology identification models respectively. The results show that the XGBoost model has the best result in identifying formation lithology, with an accuracy of 79.21%. Finally, feature importance analysis shown that MWI, MWO and MFI have important effects on model performance. This study has important implications for improving the probability of drilling into reservoirs and reducing drilling costs. 1. INTRODUCTION Lithology identification is an important basic problem in the fields of geology, oil and gas exploitation, resource exploration, geotechnical engineering, etc (Xu et al., 2021). In drilling engineering, accurate and timely determination of the formation lithology at the positive bit is one of the most important factors to ensure safe and efficient drilling (Al-AbdulJabbar et al., 2019; Zhang et al., 2017; Shan et al., 2015). The classical lithology identification methods mainly include two kinds, one is the experimental observation method, through the direct observation of rock specimens or rock cast thin slices, the rocks can be classified according to their mineral composition, color, etc (Guojun et al., 2010; Hu et al., 2010). The other is the statistical analysis method, the logging cross-section chart is drawn by measuring the logging response characteristics of pure rocks, and then the rock type is roughly determined according to the actual logging values (Teama et al., 2016; Zhou et al., 2016). The above two traditional lithology identification methods require researchers with rich experience and professional knowledge, and the identification efficiency is low and the cost is high. Therefore, a fast and accurate method for lithology identification is urgently needed.

The lithology identification of deep carbonate rocks has always been a difficult field of research. The deep carbonate rock formations in the eastern section of the Tarim Basin have complex tectonism, deposition, and diagenesis. Therefore, it is difficult to use conventional logging information to identify the lithology of deep carbonate rocks. In this study, a natural gamma ray spectrometry log in unconventional logging was used to discover a new method (comprehensive superposition method) for the lithology identification of deep carbonate rocks. The research results showed that the TH, U, K, and KTH information in a natural gamma ray spectrometry log can reflect the lithological characteristics of carbonate rocks extremely well. In addition, this new identification method was able to optimize the typical logging parameters which could best reflect the lithology of the carbonate rocks after analyzing the original record curve and a variety of combined curves of the natural gamma ray spectrometry. This new method could comprehensively superpose the typical logging parameters and was capable of finally quantitatively identifying the lithology of dolomite, micrite, calcarenite, and finely crystalline limestone in deep carbonate rocks, with an identification accuracy of more than 79 %. On the basis of the above research results, a special statistical method was made to analyze the typical logging parameters of natural gamma ray spectrometry and to assist and test the identification of a comprehensive superposition method, in order to improve the lithology identification accuracy. This new method for the lithology identification of deep carbonate rocks, based on a natural gamma ray spectrometry log, achieved good application effects in the eastern section of the Tarim Basin.

A novel hybrid method of lithology identification based on k-means++ algorithm and fuzzy decision tree

Lithology identification is crucial in mineral and energy resource exploration as it determines geological composition and guides exploration activities, improving resource location and evaluation efficiency. The advancement of artificial intelligence technology has promoted the application of machine learning-based multi-source geophysical data fusion methods in lithology identification. However, due to the differences in geophysical exploration techniques and data types across mining areas, single machine learning methods often struggle to adapt to diverse geological environments, lacking necessary universality and robustness, which severely restricts the practical application of intelligent identification technology in actual exploration. To address these limitations, this study introduces a Multi-mode Adaptive Prediction System (MAPS) for lithology identification. MAPS innovatively integrates three learning models (supervised, semi-supervised, and unsupervised learning), and can automatically select the most suitable learning mode based on prior information such as the quantity and quality of existing labeled samples and the completeness of geological background information, achieving rapid and accurate lithology identification. We verified MAPS's performance advantages through extensive comparative experiments: in supervised learning mode, compared to Support Vector Machine (SVM) and Naive Bayes classifier, accuracy improved by 0.7% and 3.5% respectively, with F1 scores increasing by 3.4% and 4.5%; in semi-supervised learning mode, compared to semi-supervised fuzzy C-means algorithm and self-learning algorithm, accuracy and F1 scores improved by a minimum of 33.67% and 0.15 respectively; in unsupervised mode, compared to traditional fuzzy C-means and Gaussian mixture models, MAPS demonstrated superior ability to mine and construct internal data structures, showing stronger feature learning capabilities. Furthermore, MAPS has shown excellent performance in the practical application of coal seam location prediction. The coal seam locations predicted by the system are highly consistent with actual drilling results, further validating MAPS's significant application potential in practical engineering. In conclusion, MAPS significantly improves the efficiency and accuracy of lithology identification, providing reliable technical support for mineral and energy resource exploration with broad application prospects.  

The rapid and accurate identification of the formation lithology encountered during the drilling of oil and gas fields is an important step to control the trajectory of the drilling tool borehole and improve the optimal reservoir encounter rate. At present, the main way to distinguish the lithology of the formation encountered by drilling is to use artificial detection elements, which does not form a set of intelligent formation identification system. In view of the above problems, this paper proposes a method to identify the lithology of drilled formation by using element and gamma spectrum measurement and establishes a reasoning model of intelligent identification of formation lithology by using improved fuzzy clustering algorithm-SVM (IFCM-SVM) method. Field application shows that the accuracy of IFCM-SVM intelligent formation identification method proposed in this paper can reach 90.9%, and verifies the feasibility of using element and gamma spectrum measurement to realize the intelligent identification of formation lithology in drilling.

Modal identification with output-only measurements plays a key role in vibration-based damage detection, model updating, and structural health monitoring in civil engineering. This paper addresses the application of modal identification method to a triangle steel tube truss natatorium using the field measurement data. To obtain dynamic characteristics of the spatial structure, four different output-only system identification methods are employed. They are natural excitation technique–eigensystem realization algorithm, data-driven stochastic subspace identification method, frequency-domain decomposition/frequency-spatial domain decomposition method, and half spectra/rational fractional orthogonal polynomial method. First an analytical modal analysis was performed on the three-dimensional finite element model according to the factual layout design to obtain the calculated frequencies and mode shapes. Then the whole procedure of the field vibration tests on the natatorium was presented. Finally, practical issues and efficiency of the four output-only modal identification techniques are investigated, and compared with the results from a finite element model. The system identification results demonstrate that both methods can provide reliable information on dynamic characteristics of the spatial structure. The frequency-domain methods, however, can quickly identify the modal parameters, but the leakage error introduced by power-spectral density estimation is existent due to the limited length of data. And the time-domain methods can avoid the leakage error, but the computational modes and the computational cost are the main two drawbacks in application. The conclusion is that several system identification methods should be consulted to ensure the accuracy of the estimated modal parameters.

Lithology identification by adaptive feature aggregation under scarce labels

Lithology identification using electrical imaging logging image: A case study in Jiyang Depression, China

Accurate lithology identification is a crucial prerequisite for sedimentary environment analysis, oil and gas exploration, and development. Traditional identification methods have problems such as low efficiency, high cost, or limited applicability. This paper proposes an ADASYN-IRMO-CatBoost combined model. Firstly, the ADASYN algorithm is used to perform adaptive sampling on the logging datasets of three wells in the southern Ordos Basin to solve the problem of data imbalance. Then, the Improved Radial Movement Optimization (IRMO) algorithm is utilized to optimize the hyperparameters of the CatBoost model. Finally, CatBoost is used as the core classifier for lithology identification. Experimental results show that the overall accuracy of the combined model on the test set reaches 92%, which is 13% higher than that of the single CatBoost model. The F1 scores of various lithologies and the AUC values of the ROC curves are significantly better than those of the single model, demonstrating stronger classification performance and robustness. It provides an efficient and accurate new method for lithology identification of sandstone and mudstone reservoirs and has good application prospects in the field of geological exploration.

An automatic identification method of imbalanced lithology based on Deep Forest and K-means SMOTE