基于LGWO-LightGBM的混合模型在火山岩测井岩性识别中的应用OA

With the Bohai Sea's complex and varied volcanic rock types and absence of specialised logging data like imaging,the machine learning method in the context of big data offers a fresh solution to the logging lithology identification challenge.In order to improve the accuracy of lithology identification,a hybrid model of lithology identification of volcanic reservoirs embedded in the LightGBM algorithm of Lévy flight strategy(LFS)grew wolf optimizer(GWO)was proposed.With LightGBM serving as the foundation algorithm,the LGWO-LightGBM hybrid model employs LGWO to optimise the hyper-parameter optimisation procedure inside the LightGBM framework,finally producing the ideal hyper-parameter combinations that yield a better accuracy rate.The experimental results demonstrate that the model performs best when the following parameters are set:180,0.2,501,460 and 30 for the number of leaf nodes,maximum depth,number of leaf nodes,and minimum number of samples of leaf nodes,respectively.This results in a lithology identification accuracy of 93.9%,which is 7.2%higher than that of the LightGBM model(86.7%),and a shorter relative calculation run time.LightGBM solves the problem of covering a variety of hyperparameters that are difficult to determine,with the advantage of better recognition accuracy and operational efficiency when dealing with a large amount of data.The model not only improves the productivity of oil and gas exploration and development of volcanic reservoirs in the Bohai Sea waters,but it also serves as a reference for identifying the volcanic lithology of reservoirs under similar types of geological conditions.