The Research Of Intelligent Control Algorithm Of Beam-pump Unit

Because of the complex geological condition, exhaustion phenomena will occur in most oil wells. The oil-providing-rules of these low-reserves oil wells are different, and change with time. Therefore, the poor management will result in low efficiency and function status's deterioration of oil pumping machine, which indirectly affect the economy benefit of oil pumping system. Furthermore, it's the main reason of the low efficiency that it can't adaptively adjust the function status of oil pumping system. As the widely used equipment in mechanical oil withdrawal field, beam-pump unit's total efficiency is lower than 25%. Thus, it has great and practical significance to development an intelligent control algorithm of beam-pump unit for the economy benefit.In the thesis, the FFT is used to deal with the electrical elements and rod-load signal of beam-pump unit to find the potential law between them. An algorithm to calculate the rod-load by BP neural network is first proposed in this thesis, which uses the motor torque signal and motor speed signal as the input. The simulation demonstrates that the output of the trained neural network has high goodness-of-fit. The change of rod-load can be well recovered by motor torque and motor speed signal. In addition, the algorithm of drawing dynamometer cards is introduced. It shows that the oil production prediction can be done by the area of dynamometer card. With the measurement and calculation of motor power consumption, three intelligent control strategies are adopted to improve the economy benefit of beam-pump unit system, which are the max of oil production, the highest efficiency, and oil production/power consumption = setting value.