石油地球物理勘探2025，Vol.60Issue(3)：588-598,11.DOI:10.13810/j.cnki.issn.1000-7210.20240333
基于NOA优化随机森林算法的砂岩孔隙度预测
Porosity prediction of sandstone based on NOA optimized random forest algorithm
摘要
Abstract
In oil and gas exploration and development,porosity is an important parameter to evaluate reservoir physical properties,especially in flooded well evaluation.Accurate porosity prediction is of key significance to the evaluation of remaining oil and subsequent production and development.The conventional linear porosity model has limitations in prediction accuracy,and the random forest regression model often faces the problems of low optimization efficiency,complex parameter adjustment,and large consumption of computing resources when traditional parameter optimization methods are employed.To improve the accuracy and efficiency of po-rosity prediction,this paper proposes a new method to optimize the random forest regression model based on the nutcracker optimization algorithm(NOA).This method is inspired by the foraging,storage,and food re-trieval behavior of the North American bird nutcracker.In this study,the random forest regression model is sub-jected to hyperparameter optimization through NOA with the acoustic time difference,compensated density,and compensated neutron curve as the input features of the model and the core porosity as the target value,which avoids locally optimal solutionand thus determines globally optimal parameter combination.Compared with the traditional grid search method,NOA shows higher efficiency in hyperparameter optimization.The re-sults of data analysis and model prediction show that this method not only speeds up the training speed of the ran-dom forest model but also effectively improves the fitting effect and prediction accuracy of the porosity model.关键词
NOA/随机森林/孔隙度/水淹井Key words
NOA/random forest/porosity/flooded wells分类
地质学引用本文复制引用
芦杨笛,邓瑞,汪益,张程恩,段宏臻..基于NOA优化随机森林算法的砂岩孔隙度预测[J].石油地球物理勘探,2025,60(3):588-598,11.基金项目
本项研究受中国石油天然气集团有限公司项目"地球物理井地立体探测方法与装备研发"(2023ZZ05-03)和中国石油集团测井有限公司项目"快速测井装备研制与配套"(CNLC2022-4D)联合资助. （2023ZZ05-03）
