System And Method Research On Online Fault Diagnosis For Beam Pumping Unit Condition

At present,beam-type pumping unit oil production system occupies a large proportion in the oil production.Pumping machine is usually with bad working environment,and system equipment is easy to damage,seriously affects the oil field efficiency and economic benefits.Therefore,it is very important to realize the automatic monitoring and scientific management of oil recovery system.The work diagram is one of the important basis for judging the working condition of pumping well.But the traditional merit map acquisition method of the sensor cost is high and easy to damage,the process of complex feedback information is slow.Therefore,based on the detailed analysis of the working characteristics and loss of the oil extraction system,the paper puts forward the method of indirectly measuring the performance diagram of the oil well through the electrical parameters,and uses the support vector machine technology to intelligently diagnose the pumping condition.Based on the conservation of energy,this paper analyzes the motion law of the beam pumping unit by using the complex vector method,establishes the function relation of the displacement of the beam pumping unit in the working process,and deduces the current and the relationship between the load of the rod and the relationship between the load and the displacement.The relationship between the components and the energy consumption of the beam pumping unit is analyzed theoretically.At the same time,the method of indirect measurement of the pumping diagram of the pumping unit is calculated and analyzed by Matlab software.In this paper,the invariant moment method is used to extract the seven invariant eigenvalues of the dynamometer,and the ratio of the measured load and the theoretical load ratio,the measured load of the lower stroke and the theoretical load ratio are combined with the oil pump efficiency.Based on the comprehensive characteristic parameters of the oil well condition,a sample library of characteristic parameters of the typical working condition of the beam pumping system was established.Based on this,the paper studies the classification method of oil well condition based on support vector machine,analyzes the influence of different kernel function and parameter combination of different penalty factor C and kernel width g on the accuracy of classification.The simulation experiments of invariant momentcharacteristic parameter samples and comprehensive characteristic parameter samples are carried out by LibSVM respectively.Theoretical analysis and simulation experiments show that the equivalent dynamometer of indirect measurement of electrical parameter method can effectively reflect the condition of oil well.Based on the support vector machine,the integrated characteristic parameters can be used to identify the oil well condition more accurately.