Study On The Parameter Detection And Intelligent Discrimination Method Of Oil Well Wax

Petroleum is an indispensable non-renewable energy in the world today.Therefore,it is very important to improve the efficiency of oil collection and to identify the failure of oil production equipment system in time.At present,most of the oil fields adopt sucker rod pumping system to collect oil,which has the advantages of simple operation,low cost and so on.In the process of mining,the wax deposition often occurs,so that the load of the pumping unit is increased and the current of the motor is increased,which seriously affect the oil extraction efficiency of sucker rod pumping system.At present,there are a lot of methods to identify the fault of wax deposition,but sometimes the recognition accuracy is not ideal.In this paper,the support vector machine(SVM)based on statistical learning theory is used to identify the fault of the sucker rod pumping system,which has strong generalization ability.The paper mainly studies the following:First,this paper introduces the background of fault diagnosis of sucker rod pumping system at home and abroad,expatiates on the structure and working principle of rod pumping system,studies the indicator diagram form and makes use of load sensor to extract the load parameters of oil well.The indicator diagram contains a lot of information,according to its image characteristics to understand the production status of oil wells,so choose the indicator diagram as a basis for oil well fault diagnosis.Second,after the ground indicator diagram is collected,the mathematical model of the system is established,and the ground indicator diagram is converted into a downhole pump dynamometer diagram,which is more conducive to understanding the working condition of the mine,and is calculated by MATLAB software.The pump dynamograph of MATLAB image processing,that specific types of images,the use of large law threshold segmentation,get the best threshold and then converted to binary images,the resulting image is processed by dilation,erosion,thinning and shrinkage in mathematical morphology,and ultimately obtain the desired image.Third,the extraction of the pump dynamograph characteristic parameters is carried out,and the extracted parameters should be distinguished,clustered and independent.Therefore,the theory of moment invariants extracted 7 invariant moments parameters of pump diagrams to describe various oil well failures,to provide data samples for classifier pattern identification.Fourth,the support vector machine is used to identify the fault of sucker rod pumping system,especially the wax deposition fault.The support vector machine is analyzed theoretically and simulated by MATLAB software.In order to get better recognition effect,the cross validation method,particle swarm optimization algorithm and genetic algorithm are used to optimize the parameters of SVM.Because different kernel function recognition effect is different,therefore,the recognition effect of different kernel function is compared,find the optimal parameters,to achieve good intelligent recognition effect.