Research And Application Of Clustering Based On Energy Consumption Data Of Rod Pumping Well System

With the rapid development of digital oilfield construction,the data volume of oil production increases dramatically.In the light of the current conditions of lack of systematic rod pumping energy consumption index,blindness of measure selection and heavy analysis task,this paper takes advantage of influencing factors of rod pumping well and establishes a complete energy consumption index system,with the help of energy conservation theory and pumping unit production knowledge.Clustering methods like K-means,K-medoids and DBSCAN algorithm are applied to build the general,ground,and downhole energy consumption model.Then best oil well clustering results are selected for characteristic analysis of each cluster,search of oil wells with high ground energy consumption or downhole energy consumption and their production features.The research shows that such clustering models are able to find high consumption wells quickly and efficiently.For wells of high ground energy consumption,the distribution characteristics of influencing factors like motor load rate,rate of pumping unit load and balance ratio can be summarized.For wells of high downhole energy consumption,that of pumping parameters can be found.Therefore it is of great significance to improve the accuracy and effectiveness of the measures and achieve energy saving of rod pumping well.