基于人工智能算法的水平井气水两相生产测井流型识别方法研究  

Research on Flow Pattern Identification Method of Logging in Horizontal Well-gas-water Two-phase Production Based on Artificial Intelligence Algorithm

作　　者：赵晓云 宋红伟[1,2] 王明星[1] 张鑫 李岩军 王枭煌 Zhao Xiaoyun;Song Hongwei;Wang Mingxing;Zhang Xin;Li Yanjun;Wang Xiaohuang(School of Geophysics and Oil Resources,Yangtze University,Hubei,430010;Research Offi ce of Yangtze University,Key Laboratory of Well Logging,China National Petroleum Corporation,Hubei,430010;Huabei Branch,CNPC China Limited,Hebei,062550)

机构地区：[1]长江大学地球物理与石油资源学院,湖北430010 [2]中国石油天然气集团公司测井重点实验室长江大学研究室,湖北430010 [3]中国石油集团测井有限公司华北分公司,河北062550

出　　处：《当代化工研究》2023年第8期173-175,共3页Modern Chemical Research

摘　　要：为探究水平井气水两相流动规律,识别气-水两相流流型。本文首先对水平井气-水两相进行数值模拟,分析不同倾角、流量配比下的两相流流型;然后利用包括CAT在内的生产测井仪器串,对不同流动条件下的气-水两相流进行动态实验,获取仪器的响应数据和气、水两相流动状态;最后利用实验得出的中心持率、井斜角度、涡轮转数,CAT各探头实测值等数据,引入由麻雀搜索算法优化的BP神经网络算法识别水平井气-水两相流型。该算法将BP算法对气水两相流型的识别精度从83.75%提高到91.66%,并加快了运算速度。为探索水平井气水两相流流型识别提供了一种新的思路。In order to investigate the two-phase gas-water flow pattern of horizontal wells and identify the gas-water two-phase flow pattern.In this paper,firstly,numerical simulation is conducted to analyze the two-phase flow pattern under different inclination angles and flow ratios of horizontal wells;Then,dynamic experiments are conducted on gas-water two-phase flow under different flow conditions by using production logging instrument strings including CAT to obtain instrument response data and gas-water two-phase flow status;finally,the data such as center holding rate,well slope angle,turbine rpm,and measured values of CAT probes are used to identify the gas-water two-phase flow pattern of horizontal wells.Finally,the BP neural network algorithm optimized by the sparrow search algorithm is introduced to identify the gas-water two-phase flow pattern of horizontal wells by using the experimental data such as center holding rate,well slope angle,turbine rotation,and the measured values of CAT probes.This algorithm improves the accuracy of BP algorithm for identifying gas-water two-phase flow pattern from 83.75% to 91.66%,and speeds up the computation speed.It provides a new idea for exploring the identification of gas-water two-phase flow pattern in horizontal wells.

关 键 词：气-水两相流流型 数值模拟 SSA算法 BP神经网络 

分 类 号：TE[石油与天然气工程]