Abstract:The updating of reservoir geological models has become a research hotspot. Nevertheless, two difficulties continue to hinder the development of reservoir geological model updating techniques. First, logging while drilling (LWD) is used mainly to guide geosteering operations and effectively identify the lithology; few scholars have researched the interpretation of reservoir physical characteristics while drilling, which is the basis of updating geological models. Second, interpretation results are difficult to transmit to geological models in real time. Based on the LWD technique, this paper uses logging interpretation, machine learning, computer science, and reservoir geological modeling theories and methods to conduct the research of real-time geological model updating around the well. First, based on effective logging data, two machine learning algorithms which are random forest (RF) and extreme gradient boosting tree (XGBoost) are used to establish interpretation models of the reservoir lithology, porosity, and permeability. The parameters of each model are optimized through cross-validation method, and LWD data are interpreted in real time by interpretation models. Second, based on the convenience of the Ocean secondary development platform and the functionality of Petrel software, a real-time transmission plug-in for the current well trajectory and reservoir interpretation results is compiled, and an automatic updating module for the geological model is established. A case study is performed with data from the Sulige gas field in the Ordos Basin, China. For the real-time interpretation of reservoir characteristics while drilling, after 228 trials, the XGBoost algorithm is chosen to establish reservoir lithology, porosity, and permeability interpretation models. For the real-time updating of the geological model around the well, given the consistent probability distributions and the agreement between adjacent wells, we obtain relative errors between the simulated and real values of the lithofacies, porosity, and permeability of 3.90%, 4.50%, and 7.60%, respectively. Therefore, this paper provides a new method for the real-time modification and updating of reservoir geological models, which preliminarily resolved the contradiction between accuracy and real time of geological model real-time updating.

Well Logging Reconstruction Based on Bidirectional GRU

Well Logging Reconstruction Based on a Temporal Convolutional Network and Bidirectional Gated Recurrent Unit Network with Attention Mechanism Optimized by Improved Sand Cat Swarm Optimization

The Bidirectional Gated Recurrent Unit Network Based on the Inception Module (Inception-BiGRU) Predicts the Missing Data by Well Logging Data

Well Logs Reconstruction Based on Deep Learning Technology

Machine-learning-based automatic well-log completion and generation: Examples from the Ordos Basin, China

USING INFORMATION GRANULATION TO CONSTRUCT THE WELL-LOGGING LITHOLOGY INTERPRETATION MODEL

Seismic and Well Logs Integration for Reservoir Lateral Porosity Prediction Based on Semi-Supervised Learning

A method for well log data generation based on a spatio-temporal neural network

Missing Well Log Reconstruction Using A Sequence Self-attention Deep Learning Framework

Deep-learning missing well-log prediction via long short-term memory network with attention-period mechanism

Missing well logs prediction using deep learning integrated neural network with the self-attention mechanism

Well logging curve reconstruction based on kernel ridge regression

Well-Logging Prediction Based on Hybrid Neural Network Model

Research on Quantitative Inversion Characterization of High-Definition Electrical Imaging Logging in Oil-Based Mud Based on BPNN and MPGA-LM Algorithm

LogRegX: An Explainable Regression Network for Cross-Well Geophysical Logs Generation

Log interpretation method of resistivity low-contrast oil pays in Chang 8 tight sandstone of Huanxian area, Ordos Basin by support vector machine

Real-time updating method of local geological model based on logging while drilling process

Geophysical Log Interpretation Using Neural Network

Simultaneous prediction of multiple physical parameters using gated recurrent neural network: Porosity, water saturation, shale content

Handling missing data in well-log curves with a gated graph neural network

Regression Committee Machine and Petrophysical Model Jointly Driven Parameters Prediction From Wireline Logs in Tight Sandstone Reservoirs