预测油层无机积垢的BP神经网络方法  

BP neural network method for predicting the inorganic scaling in the reservoir

作　　者：程翊珊 李治平[1,2] 许龙飞 史华 CHENG Yishan;LI Zhiping;XU Longfei;SHI Hua(School of Energy Resources, China University of Geosciences, Beijing 100083, China;Beijing Key Laboratory of Unconventional Natural Gas Geological Evaluation and Development Engineering, Beijing 100083, China;Research Institute of Oil and Gas Technology, PetroChina Changqing Oilfield Company, Xi’an 710018, China;National Engineering Laboratory for Exploration and Development of Low Permeability Oil and Gas Field, Xi’an 710018, China)

机构地区：[1]中国地质大学(北京)能源学院,北京100083 [2]非常规天然气地质评价及开发北京市重点实验室,北京100083 [3]中国石油长庆油田分公司油气工艺研究院,陕西西安710018 [4]低渗透油气田勘探开发国家工程实验室,陕西西安710018

出　　处：《大庆石油地质与开发》2021年第3期84-93,共10页Petroleum Geology & Oilfield Development in Daqing

基　　金：国家科技重大专项“致密油气藏数值模拟新方法与开发设计”(2017ZX05009-005)。

摘　　要：油田注水开采过程中注入水与地层水不配伍导致无机积垢阻塞储层和管道,预测油层无机积垢对稳定油井长期正常生产具有积极意义。以影响积垢因素作为输入变量,设计积垢等级概念作为输出变量,建立了预测油层无机积垢的BP神经网络模型,对实际结垢情况进行预测并控制变量分析环境因素对积垢结果的影响。结果表明:神经网络预测结果与实际实验浊度算得的积垢等级相符,配伍性良好的2种地层水混合时预测积垢等级均小于0.1,不配伍的2种地层水混合时预测积垢等级均大于10;神经网络对各影响因素的预测与积垢理论相符,积垢趋势与积垢离子浓度、温度、pH值正相关,积垢趋势与压力负相关。应用BP神经网络模型预测油层无机积垢具有预测精度高、计算速度快、需要资料少等特点,为油层无机积垢预测提供了一种高效快捷的方法。In the process of water-flooded oilfield exploitation,the inorganic scale caused by the incompatibility between the injected water and formation water will result in the blocking of reservoirs and pipelines,so the prediction of inorganic scaling in the reservoir is of great significance for the long-term normal production of oil wells.Taking the scaling influencing factors as the input variables and the designed concept of scale grade as the output variable,a BP neural network model for predicting the inorganic scaling in the reservoir was established,which can predict the actual scaling and control the variables to analyze the influences of environmental factors on the scaled results.The results show that the predicted results by the neural network are coincident with the actual experimental scaling grade calculated by the turbidity,and their predicted scale grades are both less than 0.1 when the two types of formation water with good compatibility are mixed,while they are higher than 10 when two types of the formation water without compatibility are mixed.The prediction of neural network for each influencing factor is consistent with the scaling theory,and the scale trend is positively correlated with the fouling ion concentration,temperature and pH value,while negatively correlated with the pressure.The inorganic scaling predication of the reservoir with BP neural network model is characterized by high precision,rapid calculation speed and less data,providing a high-efficiency and rapid prediction method for inorganic scaling prediction in the reservoir.

关 键 词：油层无机积垢 BP神经网络 影响因素 预测积垢等级 地层水 

分 类 号：TE358[石油与天然气工程—油气田开发工程]