首页|基于功图故障特征分析的液量计量方法研究

基于功图故障特征分析的液量计量方法研究

扫码查看
为解决现有原油产液量计量方法精度低、成本高、工艺流程复杂等问题,对一种基于功图故障种类与强弱特征和BP神经网络分析的组合计算油井产量方法进行研究.该方法通过马田系统操作从地面示功图中提取特征向量,并融合自适应算法,对示功图故障种类及强弱进行识别与区分;建立了产液量与不同故障类型及强弱之间的映射关系以此代替泵效,计算映射关系并建立BP神经网络以此来优化关系,得到基于故障种类与强弱的计量产量的模型.给出该计量模型的具体实现,并选择2020年的新疆公司某作业区50井的产量及示功图作为数据集进行计算验证.上述模型计算结果表明,计算的石油产量和实际的石油产量作对比,其平均相对误差为3.26%,平均绝对误差为0.5%,满足油田生产利益最大化.
Research on Measurement Method Based on Fault Types and Strength Characteristics of Power Diagram
In order to solve the problems of low precision,high cost and complicated process flow of the existing crude oil liquid production measurement methods,it proposes a combined calculation method of oil well production based on the fault types and strong and weak characteristics of power diagram and BP neural network analysis.In this method,the feature vectors are ex-tracted from the ground indicator diagram through the operation of the horse field system,and the adaptive algorithm is inte-grated to identify and distinguish the fault types and strength of the indicator diagram.The mapping relationship between liquid production volume and different fault types and strength was established to replace pump efficiency.The mapping relationship was calculated and BP neural network was established to optimize the relationship,and a production measurement model based on fault types and strength was obtained.The concrete realization of the measurement model is given,and the output and indicator diagram of well50in an operation area of Xinjiang Company in 2020are selected as the data set for calculation and verification.The above model calculation results show that the average relative error and absolute error between the calculated production and the real oil well production are 3.26%and0.5%respectively,which is less than the industry standard of 10%on the whole,satis-fying the maximization of oilfield production benefits.

Calculation of Oil Well ProductionIndicator DiagramFault IdentificationFeature AnalysisBP Neural Network

唐丽雯、樊军、赵新语、何方

展开 >

新疆大学机械工程学院,新疆 乌鲁木齐 830049

油井产液量计算 示功图 故障识别 特征分析 BP神经网络

国家自然科学基金地区科学基金项目

11462021

2024

机械设计与制造
辽宁省机械研究院

机械设计与制造

CSTPCD北大核心
影响因子:0.511
ISSN:1001-3997
年,卷(期):2024.403(9)
  • 5