AUTHOR=Liu Jinzi , Liu Xinyu 

TITLE=Recognition and Classification for Inter-well Nonlinear Permeability Configuration in Low Permeability Reservoirs Utilizing Machine Learning Methods

JOURNAL=Frontiers in Earth Science

VOLUME=Volume 10 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/earth-science/articles/10.3389/feart.2022.849407

DOI=10.3389/feart.2022.849407

ISSN=2296-6463

ABSTRACT=Machine learning method had become mainly popular research algorithm for reservoir engineering evaluation. This paper selection and optimization one machine learning method to classification and recognition for nonlinear permeability configuration between injection and production wells in low permeability reservoir. The paper contains four classes of inter-well nonlinear permeability configurations between injection and production wells. Respectively, homogeneous, Linear increment, convexity increasing (Logarithmic function) and convex downward increasing (Exponential function). According to four kinds of nonlinear permeability distribution in low permeability reservoir, increased effect of threshold pressure gradient, it is considered to established productivity formula. And then, Decision tree (Tree), Neural Networks (NN) and Support Vector Machines (SVM) are used to train dynamic data under the influence of nonlinear permeability configuration in low permeability reservoirs as the training model. The data set is trained with dynamic production data under different permeability configuration, well spacing, thickness, pressure, production. The permeability configuration is classification and recognition by using different machine learning model. The results show that compared with NN and Decision Tree, SVM had better performance in accuracy of verification, TPR, FNR and Roc. Moreover, SVM verification results are close to the train methods. This paper provides new insights and methods to classification and recognition inter-well nonlinear permeability configuration in low permeability reservoirs. Moreover, it can also be extended to other unconventional reservoirs to solve similar theoretical problems.