Working Characteristic Analysis And On-line Fault Diagnosis Of Submersible Screw Pump

Submersible screw pump has the advantage of low energy consumption,high efficiency,small occupation area,and fundamentally solves the problems of rod breaking,rod and tubing wearing of ground-drive screw pump,the number of it is increasing year by year in the oilfield,however,due to the lack of the way of online real-time monitoring the woking condition of underground submersible screw pump unit,it is difficult to judge the working condition of the downhole unit timely and accurately,which has seriously affected the normal production and the life of the unit and greatly limits its further promotion in the oilfield,in order to ensure the safe and high efficient production of submersible screw pump,increase the efforts to be promoted and adapt to the needs of the digital oilfield construction,this paper carries out working characteristic analysis and on-line fault diagnosis of submersible screw pump.The paper first analyzes the characteristics of submersible screw pump,and on this basis,wavelet packet and RBF neural network are used to diagnose the fault,the main contents are as follows:(1)To submersible screw pump system as the research object,the motion characteristics,mechanical characteristics,pump displcement characteristics,etc. are analyzed,and the node analysis methord is used to establish the mathematical model between the submersible motor active power and pump power.(2)The ANSYS software is applied to the finite element analysis of submersible screw pump,and the influence law of different conditions on the pump performance is obtained.(3)Based on the GPRS remote online transmission network,according to the common fault type and diagnosis methods of ground-drive screw pump,combined with its own structure charcateristics and working characteristics of submersible screw pump system,the online fault diagnosis scheme based on submersible motor active power as research subject and combining wavelet packet with RBF neural network is put forword.(4)Wavelet is used to extract the submersible motor active power signal,the fault sample database of submersible screw pump is built,Matlab software is applied to building RBF neural network,and the training parameters of the network are determined,the training and testing of the network are completed.(5)Based on the platform of Visual Basic 6.0,the online fault diagnosis system software of submersible screw pump is developed,which realizes the function of human-computer interaction,online monitoring and fault diagnosis and provides a reliable guarantee for the safe and high efficient production of submersible screw pump.