Research On The Establishment Of Production Rule Library Based On Sucker Rod Pumping Wells Faults

At present,most of the oil production operations around the world rely on manual lifting,and sucker rod pumping accounts for the vast majority of manual lifting.In this kind of method,such as pumps and other equipments,mostly work in harsh field environment where the geographical location is dispersed and the production environment is very harsh.In addition,the condition of the pumping wells is too complex,so the pumping wells are prone to various faults.Once the fault is not handled in time,it will cause economic losses or even other terrible accidents such as wells damaged seriously.Therefore,it is of great significance to analyze and classify the working conditions of oil wells quickly and accurately for improving efficiency,saving cost and ensuring safety in oil production industry.In this paper,the indicator diagram information and well parameter information are taken as the research objects,and the expert system based on production rule is used to classify the fault diagnosis of pumping wells,and the establishment of the production rule library is researched and improved especially.Firstly,the history of fault diagnosis methods for sucker rod pumping wells is introduced,and the methods relying on computer technology are explained in detail especially.At the same time,the principle of the complete production process of sucker rod pumping wells is analyzed.Based on the working process of sucker rod pumps,the formation process of indicator diagram and its important role in the field of fault diagnosis are discussed.Typical faults are also introduced,which lays a solid theoretical foundation for feature extraction at next.Secondly,the feature extraction method based on mechanism analysis is researched.This paper selects the method of transforming indicator diagram based on electrical parameters.The indicator diagram is obtained by calculating and transforming the electrical parameters directly.Based on the analysis of the principle of oil production process,the mechanism feature information is extracted after the mechanism analysis of the indicator diagram,and combined with the feature information of the parameters of the pumping well itself,a new kind of feature parameter is constructed.The characteristics of the indicator diagram and the parameters of the oil well under different working conditions are analyzed with expert experience,then each working condition are matched with different parameters.Thirdly,the production rule library with the hybrid reasoning method are constructed to diagnose and classify different working conditions.In the research of the method of building production rule library,decision tree is used to build production rule library under different conditions relying on the parameters matched in the preceding text.Based on rough set,the disadvantage of ID3 decision tree that can not deal with noise data is reduced and improved,which speeds up the construction of production rule library and also improves accuracy of its production rules.Relying on the field data of an oil field in Liaoning province,the simulation experiment has achieved good results.Finally,based on Visual Basic,the upper computer software platform of the real-time fault diagnosis system for sucker rod pumping wells is designed and developed,and based on JAVA,the matching APP for oilfield staff is developed.The functions of remote monitoring of oil well information and real-time working conditions are realized.At present,the upper computer software platform runs well in the industrial field,meets the actual needs of enterprises,achieves its expected goals,and APP will be launched soon,which can make the staff more convenient to apply the system.