基于随机森林算法的煤层预测方法研究

Research on Coal Seam Prediction Method Based on Random Forest Algorithm

煤层的识别对于煤层气的开发至关重要,使用密度测井是煤层识别的主要工具,但它增加了运行的成本。采用随机森林机器学习方法从钻井数据中识别煤,能够大大降低勘探成本。通过与实测密度测井验证,所有井的煤层识别率均超过90%,这表明机器学习方法是识别煤层并减少测井成本的有效方法。

Coal seam identification is very important for the development of coalbed methane.Density logging is the main tool for coal identification,but it increases the cost of operation.Using random forest machine learning method to identify coal from drilling data can greatly reduce the exploration cost.Through the verification with the measured density logging,the coal recognition rate of all wells is more than 90%,which shows that the machine learning method is an effective method to identify coal seams and reduce the logging cost.