基于机器学习BP算法和树模型算法的井筒流体流型预测

Wellbore Fluid Flow Pattern Prediction Based on Machine Learning BP Algorithm and Tree Model Algorithm

近年来随着石油行业的不断发展,为适应油藏在地下不同分布情况,除了传统采用的垂直井以外,大斜度井、水平井也不断被利用于对油藏的探测与开发。判断不同斜度井在不同流量、含水率下井中流体的流型就成了一个亟需解决的问题。本文通过分别使用BP神经网络和树模型,对不同流量、含水率的各斜度井井下流体流型进行预测,与实验所获取的数据进行对比,两种算法预测准确度分别能达到75%和91.7%,验证了BP神经网络算法和树模型算法在流型预测中的有效性,为后续井下流型研究提供了一种新的方案。

In recent years,with the continuous development of the petroleum industry,in order to meet the different distribution of oil reservoirs in the ground,in addition to the traditional vertical wells,highly deviated wells and horizontal wells have been continuously utilized to detect and develop oil reservoirs.Judging the flow pattern of the fluid in the well with different flow rates and water cuts has become an urgent problem to be solved.This paper uses BP neural network and tree model respectively.Compared with the data obtained in the experiment,the prediction accuracy of the two algorithms can reach 75% and 91.7% respectively,which verifies the effectiveness of the BP neural network algorithm and the tree model algorithm in the prediction of the flow pattern,which is a good follow-up Downhole flow pattern research provides a new solution.