基于机器学习聚类算法的水平井流型预测

Horizontal Well Flow Pattern Prediction Based on Machine Learning Clustering Algorithm

随着水平井钻井技术的日趋完善其规模不断扩大,井下流体流型判断也成为测井作业中的首要工作,通过利用聚类算法对现有测井数据进行预测分析并将结果与实验流体流型进行对比,算法整体预测准确度能够达到80%以上,验证了此方法在流型预测方面的有效性并为井下流体流型预测提供了一种新思路。

With the continuous improvement of horizontal well drilling technology and its continuous expansion,downhole fluid flow pattern judgment has also become the primary task in logging operations.The clustering algorithm is used to predict and analyze existing logging data and compare the results with experimental fluid flow.The overall prediction accuracy of the algorithm can reach more than 80%,which verifies the effectiveness of this method in flow pattern prediction and provides a new idea for downhole fluid flow pattern prediction.