Study On Prediction Method Of Rock Physical Properties Based On Logging Curve

The core of logging interpretation is to determine the relationship between logging information and geological information,quickly and accurately identify lithology and divide strata,which has important reference significance in the basic research of reservoir,and efficient and accurate acquisition of reservoir porosity is an important part of oil and gas exploration.Due to the low efficiency and low accuracy of traditional logging curve identification,it is not conducive to the practical engineering application.Therefore,it is of great significance to build a logging curve model with fast calculation speed and high accuracy.In this paper,the sandy and argillaceous strata of Zhuxianzhuang mine in Huaibei are studied.Firstly,according to the specific geological conditions of the study area,the collected logging data are preprocessed,and the stratigraphic division and correlation of the study area are completed by using the response characteristics of logging curves;Secondly,combined with the existing logging data in the study area,the logging curve response model and porosity and permeability calculation model of the target interval in the study area are established to realize the preliminary identification of lithology and the preliminary calculation of porosity and permeability in the study area;Thirdly,the BA-BP neural network prediction model is constructed to further optimize the lithology and porosity;Finally,comparative analysis of the above research results shows that:in the interpretation and evaluation of conventional logging curves of 17-2 well and 17-5 well in the study area,the prediction accuracy of porosity is quite different,the average error of porosity calculated by conventional logging method of 17-2 well is 13.92%,and the average error of porosity calculated by conventional logging method of 17-5 well is 42.31%;The BA-BP neural network model is used to identify the lithology and predict the porosity of the two wells,and the average error is less than 3%.In this study area,the ba-bp neural network algorithm has higher accuracy and wider applicability.It can be seen that the BA-BP neural network algorithm in the study area has higher accuracy and better prediction effect than the traditional logging method in dealing with the nonlinear relationship.The accurate prediction of reservoir porosity can also provide reference for the exploration and development of oil and gas resources.Fig.[35]table[16]reference[109]...