A Step-by-Step Forward Modeling Approach for Bridging Reservoir Properties to Pre-Stack and Post-Stack Seismic Data in Hydrocarbon Identification
LI Shengge;ZHANG Ji;YEHEYA Zhayier;LIU Guiping;CHENG Yuanfeng;Forward modeling methods for seismic data are crucial for revealing the lithology and fluid-bearing characteristics of underground formations, but directly correlating microscopic reservoir physical properties with macroscopic seismic data is extremely challenging. By introducing elastic parameters as a bridge and combining rock physics theory, a successful connection between microscopic reservoir parameters and macroscopic seismic data has been established. Based on this, a step-by-step forward modeling method from reservoir physical properties to seismic Amplitude Versus Offset(AVO) data was developed using Python scripting. Through case studies of a post-stack wedge model and a pre-stack three-layer cake model, the thin-layer tuning effect and the differences between the convolution results of the AVO gather and the theoretical results of the exact Zoeppritz equation were analyzed. The results show that high-porosity reservoirs exhibit more significant responses on pre-stack AVO gathers than low-porosity reservoirs, with AVO intercepts being 4.1 times higher and slopes being 4.6 times steeper,which facilitates oil-water identification. Furthermore, amplitude responses are stronger for natural gas reservoirs compared to liquid oil reservoirs, with AVO intercepts being 32.9% higher and slopes being 8.3% steeper, making natural gas identification more accurate.
Study on Multi-Element Coupled Hydrocarbon Migration Model of Cretaceous Qingshuihe Formation in Block 3,Central Junggar Basin
ZHAO Wenshuo;YAO Zongquan;WANG Yong;ZHANG Xuecai;WANG Haoyi;HAN Haibo;HUANG Yu;MU Yuqian;The accumulation and differential distribution of hydrocarbon in the Cretaceous Qingshuihe formation in Block 3 in central Junggar Basin mainly depend on the three-dimensional spatial configuration of the composite transport system of fault-weathering leach zone, sand-body and unconformity, in which the fault system and weathering leach zone dominate the vertical migration and multi-zone adjustment, while the sand-body and unconformity plane control the dominant lateral migration path. Based on the characterization and spatial configuration combination of transport systems such as bottom conglomerate, unconformity, weathering leach zone and fault, the spatial coupling relationship between storage layer(bottom conglomerate, unconformity plane, weathering leach zone) and reservoir forming elements are defined. Three hydrocarbon migration modes of Qingshuihe formation are established, including the A-type fault-assisted upper storage(fault-assisted, lateral escape, upper reservoir formation), B-type carpet sand storage(horizontal closure, carpet sand accumulation), C-type fault-assisted blanket sand storage(fault-assisted, lateral containment, blanket sand accumulation). These systematically elucidate the controlling effect of the migration mode on oil and gas, and the spatial evolution law.
Optimization Strategy of Genetic Algorithm in Transient Electromagnetic Deep Learning Inversion
WU Wenyu;ZHANG Yingying;WU Xinyu;XIE Bin;Transient electromagnetic(TEM) 1D inversion has been widely applied in geological engineering, yet these conventional methods remain constrained by strong dependence on initial models, limited noise resistance,and inefficiency in achieving real-time inversion. To address these challenges, we propose a convolutional neural network-long short-term memory hybrid architecture tailored to the inherent characteristics of TEM inversion.Using loop-source TEM 1D forward modeling, we generate training data comprising sampling time-decay voltage pairs as network inputs. An optimization strategy combining the Adam optimizer with the ReduceLROnPlateau learning rate scheduler is implemented to adaptively adjust learning rates during parameter updates. In view of the problem that the hyper-parameter setting of the current network structure depends on the empirical value and lacks scientificity, which leads to the waste of computing power and time, the genetic algorithm is proposed to optimize the hyper-parameters of the neural network structure in the model training stage to reduce the training cost and improve the model performance. The output layer provides subsurface resistivity-depth profiles corresponding to input electromagnetic responses, enabling deep learning-based TEM inversion. The trained GA-CNN-LSTM network demonstrates robust performance in real-time predictions for randomly generated three-layer and fivelayer models, with validation metrics yielding R2>0.9. Further evaluation using noise-contaminated data reveals that the optimized network achieves an average inversion time of 0.13 s and a structural similarity index of 90.138%across four common models, outperforming both Occam and LSTM inversion methods. Generalization capability is validated through successful inversion of 3D forward modeling data. These results demonstrate the algorithm's reliability, computational efficiency, and practical utility in complex geological scenarios. Finally, Occam inversion and neural network inversion are carried out on the measured data respectively. The trained neural network only take 0.73 s to complete the inversion accuracy, which verifies the practicability of the algorithm in this paper.
Lithofacies Prediction Based on Wavelet Transform and Convolutional Neural Networks
HUANG Yongbo;HAN Changcheng;WEI Yatao;ZHOU Yanxu;JIANG Xin;LI Jiaxuan;ZHANG Yuqi;Lithofacies analysis serves as the foundation for identifying high-quality reservoirs. However, in areas devoid of well data or constrained by complex inter-well geological conditions, traditional techniques struggle to rapidly and accurately recognize lithofacies types and their spatial distribution. This paper proposes a convolutional neural network(CNN)-based lithofacies identification method integrated with continuous wavelet transform(CWT),achieving efficient lithofacies recognition through deep learning. Applied to the Karamay formation in the Zhengshacun area of the Junggar Basin, the methodology involves: classifying typical lithofacies based on core and logging characteristics, performing synthetic record-based well-to-seismic matching to align logging lithofacies with post-stack seismic data, converting the matched seismic waveforms into time-frequency spectrum maps using Morlet wavelet transform, generating a time-frequency spectrum dataset for different lithofacies, and constructing and training a CNN model for validation. Under horizon constraints, the planar distribution of various lithofacies is investigated. Results demonstrate that the Morlet-CNN model achieves high identification accuracy, with recognition rates exceeding 85% for 4 lithofacies types in blind well X2, significantly enhancing both the efficiency and accuracy of lithofacies identification.
Prediction of Coal Tar Yield and Oil-Richness in the Badaowan Formation of Tiaohu Sag, Santanghu Basin
WU Hongfei;LI Xin;WANG Xinggang;JIAO Lixin;CAO Zhixiong;LI Bin;WEI Bo;FENG Shuo;Tar-rich coal integrates the properties of coal, oil and gas. Promoting its exploration and development has important strategic value for ensuring the supply of oil and gas resources in China, and realizing the clean and efficient utilization of coal. Therefore, based on the data of tar yield, industrial components, elemental analysis and coal petrography analysis of coal samples in Tiaohu sag of Santanghu Basin, combined with logging response,a logging prediction model of tar-rich coal tar yield is established, and the tar-rich coal resources of Badaowan formation in the study area are predicted. The results show that: the yield of coal tar is positively correlated with volatile yield, hydrogen content and vitrinite content, and negatively correlated with ash yield and inertinite content. There is a good negative correlation between volatile yield, hydrogen content, vitrinite content and acoustic time difference, compensated density logging values, and a poor correlation with natural gamma logging values. In addition, the prediction model of coal tar yield based on machine learning is established. The correlation coefficient between the predicted value and the actual value is 0.92. The relative error of tar yield prediction results of 90% coal samples is less than 20%, and the relative error of 75% coal samples is less than 15%.
Tracking the information about your manuscript
Communicate with the editorial office
Query manuscript payment status Edit officeCollecting, editing, reviewing and other affairs offices
Managing manuscripts
Managing author information and external review Expert Information Expert officeOnline Review
Online Communication with the Editorial Department
