Logging lithofacies identification of tight sandstone reservoir based on ensemble learning algorithm

Lithofacies of tight sandstone reservoir is an important basis for reservoir identification and reservoir parameter calculation. Intelligent algorithm can effectively improve the identification accuracy of reservoir lithofacies, but a single intelligent algorithm is easy to overfit and has poor stability. In this study, four Boosting ensemble learning algorithms, GBDT, LightGBM, XGBoost and CatBoost, are used. Based on the logging data of Z gas field in Xihu Sag, seven logging parameters with high sensitivity to lithofacies, including P16H and CNCF, are screened by decision tree feature selection and artificial deletion. Based on these parameters, a lithofacies intelligent recognition model is constructed, and the genetic algorithm is used to optimize the hyperparameters of the model, so as to realize the classification and recognition of 12 lithofacies types in Z gas field. The results show that the accuracy of LightGBM algorithm optimized by genetic algorithm reaches 87.3% in lithofacies recognition task, which is better than traditional machine learning methods such as Decision Tree and Support Vector Machine. The four Boosting algorithms proposed in this study, which combine feature selection and genetic algorithm optimization, provide a new research idea and method for lithofacies recognition based on logging data.