Fault Research About Pumping Unit Gearbox Based On Neural Network

In Oil and Gas Field and Petroleum Exploration area, there will be a lot of trouble and faults in the process of using oil pumping unit for oil production. Transmission of pumping unit is the most important part of the whole pumping unit. Because the gear box of the oil pumping machine can provide a large transmission torque and its transmission ratio fixed characteristics, and the mechanical structure has good stability and is widely used in pumping unit transmission equipment. Transmission is the most important and most prone to failure in the actual operation of the pumping unit. So in order to guarantee the normal production work and need to pump unit gearbox real-time operating state detection, fight for fault detection and maintenance in the shortest time after occurrence of a fault, so as to shorten the maintenance time and improve work efficiency.As the pumping unit in the process of work is a very complex process, the work environment is also subject to serious interference from the external environment. Pumping unit in oil pumping and mechanical vibration generated by the vibration signal is very complex, in the vibration signal exists not only the normal operation of the vibration information at the same time, there are various mechanical structure vibration information. It is very difficult to extract the vibration information of the gear box which is needed in the mechanical structure and the vibration information generated by the gear box. The on pumping oil machine gearbox and its subsidiary machinery structure vibration characteristics based, diagnosis method using neural network combined with frequency analysis of vibration signal analysis processing finally, analysis the signal for neural network training and optimization, the fault information will be extracted. Results show that the faults of this method in the pumping unit gearbox were extracted with high accuracy and very reliable.