抽油机井云计算功图计产模式较边缘计算功图计产模式提高投资回报率  

The ROI of cloud computing mode is higher than that of edge computing mode in pumping well production calculate by dynamometer cards

作　　者：陆梅[1,2] 吴春生 刘天宇[1,2] 艾信[1,2] 田发国 吴利利[1,2] 辛宏 LU Mei;WU Chunsheng;LIU Tianyu;AI Xin;TIAN Faguo;WU Lili;XIN Hong(Oil&Gas Technology Research Institute of PetroChina Changqing Oilfield Branch,Xi’an,Shaanxi,710018,China;National Engineering Laboratory for Exploration and Development of Low-Permeability Oil&Gas Fields,Xi’an,Shaanxi,710018,China)

机构地区：[1]中国石油天然气股份有限公司长庆油田分公司油气工艺研究院,陕西西安710018 [2]低渗透油气田勘探开发国家工程实验室,陕西西安710018

出　　处：《石油钻采工艺》2024年第6期766-782,共17页Oil Drilling & Production Technology

基　　金：国家科技重大专项“鄂尔多斯盆地大型低渗透岩性地层油气藏开发示范工程”(编号:2016ZX05050)。

摘　　要：(目的意义)当前,随着油田数智转型工作的不断推进,相继出现基于边缘计算、云计算两种功图计产模式的油田数字化建设架构,如何选择最佳的技术路线,是摆在各大油田面前的难题。(方法过程)文章利用上述两种模式分别对长庆油田4类抽油机井开展了现场试验,并将两种模式的所对应的采样频次、处理效率、计产精度、投资回报率等方面进行了比较。(结果现象)分析结果显示,边缘计算功图计产模式具有更高的采样频次,更快分析处理速度,但投资回报率比较低。通过与人工单量结果对比显示,在云端功图采集张数达标的情况下,两种模式的计产结果与人工单量结果相对误差分别是2.37%、2.65%,基本保持一致;在云端功图采集张数不达标的情况下,对于出液平稳的抽油机井,两种模式的计产结果与人工单量结果相对误差分别是6.95%、3.94%,仍保持一致;在云端功图采集张数不达标的情况下,对于出液波动的抽油机井,边缘计算功图计产结果与人工单量结果相对误差4.37%,基本一致。但云计算功图计产结果与人工单量结果存在较大误差,相对误差18.42%。(结论建议)目前5G、光纤等传输技术已在各大油田规模应用,油田物联网建设基础较好,建议优选云计算功图计产模式,该模式计产结果准确且投资回报率相对较高。In order to solve the problem of how to choose the cloud computing dynamometer card production model and edge computing dynamometer card production model for pumping wells in the current oilfield,the paper uses the above two models to carry out field tests on four types of pumping wells in Changqing Oilfield,and compares the corresponding sampling frequency,processing efficiency,production accuracy,return on investment and other aspects of the two models.The analysis results show that the edge computing power map production planning mode has higher sampling frequency and faster analysis and processing speed,but the return on investment is relatively low.By comparing with the manual single quantity results,it is shown that the relative errors between the production calculation results of the two modes and the manual single quantity results are 2.37% and 2.65%,respectively,when the number of collected power maps in the cloud meets the standard,which is basically consistent;In the case where the number of collected power maps in the cloud does not meet the standard,the relative errors between the production calculation results of the two modes and the manual single quantity results are 6.95% and 3.94%,respectively,for a pumping well with stable liquid output,which still remain consistent;Under the condition that the number of collected cloud power maps is not up to the standard,for the pumping wells with fluid fluctuation,the relative error between the edge computing power map production results and the manual single quantity results is basically the same 4.37%.However,there is a significant error between the cloud computing power chart production results and the manual single quantity results,with a relative error of 18.42%.At present,transmission technologies such as 5G and fiber optics have been widely applied in major oil fields,and the foundation of IoT construction in oil fields is good.It is recommended to choose the cloud computing power diagram production calculation mode,which has accurate production resul

关 键 词：油井 采油 抽油机井 产液量 功图计产 云计算 边缘计算 物联网 

分 类 号：TE355[石油与天然气工程—油气田开发工程]