Development Of Operation Performance Analysis And Fault Diagnosis System For Oil Production Screw Pump

Oil production screw pump is suitable for high viscosity and atmospheric oil ratio of crude oil production;the number of applications in the oilfield is increasing year by year.However,the operation performance and fault type of oil production screw pump are difficult to determine in time and accurately when it works in thousands of meters deep oil well,which affects the stability of oil well liquid production.In order to ensure the safe and efficient production of oil production screw pump,this thesis develops the operation performance analysis and fault diagnosis system of oil production screw pump.The research contents are as follows:Firstly,the basic theory of oil production screw pump is analyzed to provide theoretical basis for its operation performance analysis.Secondly,taking GLB120 production screw pump as the research object,this thesis analyzes the influence of rotor speed,oil viscosity,gas oil ratio and other active parameters in the fault diagnosis of production screw pump on its operation performance.And these parameters are rarely considered by domestic and foreign scholars.On this basis,active parameters and passive parameters such as active power,dynamic liquid level,instantaneous flow rate,and wellhead back pressure are proposed as characteristic parameters of fault diagnosis.Then,the existing fault diagnosis methods were studied and compared,and the method of combining Wavelet Packet and PNN is proposed to be applied to the fault diagnosis of oil production screw pump,and the method is compared with BPNN and RBFNN through examples in terms of diagnosis speed and accuracy.Finally,based on the proposed fault diagnosis method,the fault diagnosis system of oil production screw pump is developed.The research shows that the active parameters and passive parameters are taken as the characteristic parameters of fault diagnosis,the fault diagnosis method of combining wavelet packet and probabilistic neural network can process 300 sets of test samples of surface driven screw pump,and the diagnosis speed and accuracy are relatively better;the diagnosis system is tested by using the historical data of submersible direct drive screw pump,and the system can accurately identify the fault types,proving the operability of the fault diagnosis system.The method enriches the fault diagnosis technology of oil production screw pump.