Study On The Recognition Method Of Formation Hydrate Based On Integrated Learning Algorithm

Natural gas hydrate is a kind of clean energy,which is widely distributed in the world and rich in reserves.Researchers have successfully drilled hydrate in Shenhu area of South China Sea and Qilian permafrost.At present,the main methods of gas hydrate identification include core sampling and identification based on formation logging data and seismic data.The traditional hydrate identification method is mainly analyzed by experts based on sampling,logging or seismic data,which depends on expert experience and level,and takes time and effort.With the development of logging while drilling technology,it is very important to identify formation hydrate in real time and quickly during drilling.It is of great significance to discover hydrate intervals in time for intelligent recognition in the future.In this paper,based on the logging data of 8 wells,through the optimization of typical machine learning methods,the hydrate identification is carried out by using the integrated learning method based on the logging curve.The digital tool of petroleum cloud logging curve is used to collect logging data,the integrated learning Adaboost algorithm is used to identify hydrate intervals,and the prediction results are evaluated according to a series of parameters such as F₁score,accuracy and recall rate.In addition,by optimizing the number of logging parameters,we can find the influence degree of different logging parameters on the test results.The test results show that the integrated learning algorithm Adaboost has high accuracy and timeliness,which provides a new idea for the formation hydrate recognition and is of great significance for the realization of intelligent recognition in the future.