Application of the decision tree method to lithology identification of volcanic rocks-taking the Mesozoic in the Laizhouwan Sag as an example

Abstract

Abstract The decision tree method can be used to identify complex volcanic rock lithology by dividing lithology sample data layer by layer and establishing a tree structure classification model. Mesozoic volcanic strata are widely developed in the Bohai Bay Basin, the rock types are complex and diverse, and the logging response is irregular. Taking the D oilfield of the Laizhouwan Sag in the Bohai Bay Basin as an example, this study selects volcanic rocks with good development scales and single-layer thicknesses of more than 0.2 m as samples. Based on a comparison of various lithology identification methods and both coring and logging data, using the decision tree analysis method and the probability density characteristics of logging parameters, six logging parameters with good sensitivity to the response of the volcanic rocks of the above formation are selected (resistivity (RD), spontaneous potential (SP), density (ZDEN), natural gamma ray (GR), acoustic (DT), and compensated neutron correction (CNCF) curves), which are combined to form a lithology classifier with a tree structure similar to a flow chart. This method can clearly express the process and result of identifying volcanic rock lithology with each logging curve. Additionally, crossplots and imaging logging are used to identify the volcanic rock structure, and the core data are used to correct the identified lithology. A combination of conventional logging, imaging logging and the decision tree method is proposed to identify volcanic rock lithology, which substantially improves the accuracy of rock identification.