| With the further development towards oil and gas fields,the requirement on the precision and efficiency of logging data improved to a new level.Experts with rich knowledge and experience can make up for part of the demand,but the time cost is too high.At the same time,the traditional manual logging interpretation method can not meet the needs of efficient oilfield development.Therefore,a cutting-edge method on advancing logging interpretation is urgently needed to improve both the precision of interpretation and efficiency of exploration.In order to handle the above problems,this passage discussed new thoughts on the way of logging interpretation which is based on theoretical knowledge of sedimentology and logging theory,relating geological,drilling and logging data in the west of the 7thblock of Gudong Oilfield as the data basis,fully harnessing advantages of Random Forest algorithms in data mining.Foremost,pretreat the logging data.It includes well log curves standardization and the standardization of logging interpretation model data.In the next step,study further on data mining technology and related algorithms,the establishment of the model using Random Forest algorithm is determined.In the final stage,we complete the evaluation of the importance of logging curve and establishment of logging interpretation model based on Random Forest algorithm for the west of the7thblock of Gudong Oilfield.Finally,the establishment of a three-dimensional model of oil and water distribution before development in the study area is completed.Through verification,the accuracy of model is strongly supported by fact that the consequence have the compliance rate of up to 91.7%with manual logging interpretation,and at the same time,highly response to the oil production and development documents.This model greatly improves the precision and the efficiency of logging interpretation while based exactly on fully utilize expert impact. |
PDF Full Text Request