Energy-saving Algorithm For Oil Pumping Based On BP Neural Network

Oil is one of China's basic industries and occupies a very important position in the development of the national economy.With the continuous development of oil fields,most of China's onshore oil wells are already in a low-permeability or ultra-low-permeability state,the mismatch between the working condition of the pumping unit and the characteristics of the oil well has produced a phenomenon of "non-full pumping" or "empty pumping",resulting in serious waste of electrical energy.Therefore,reducing unnecessary electrical energy loss,increasing oil production,and then improving oilfield production efficiency are of great significance to enhance the international competitiveness of our oilfields.Based on the research and analysis of the influencing factors of oil well pumping process,this dissertation proposes a kind of energy-saving algorithm of oil well pumping which integrates BP neural network.Firstly,a single-well benefit model is established based on factors such as the oil yield,crude oil price,and electricity cost that affect the profit of the well,and an energy-saving algorithm for oil extraction is designed,including the pre-test oil extraction algorithm,the first and second test oil extraction algorithms,average power-liquid production algorithm,time-electricity price matching algorithm,continuous pumping energy-saving algorithm and optimal stopping time algorithm.Secondly,the dissertation uses these algorithms to collect relevant data as a neural network training set,and use the good self-learning ability of the BP neural network to design a pumping energy-saving algorithm that incorporates the BP neural network.Then,the dissertation uses Keil ?Vision5 as the algorithm development platform,and choose the best number of hidden layer neuron nodes and the activation function through experiments to build a neural network.Finally,the algorithm is implanted into the ARM Cortex-M3 development board,and an experimental environment is built to predict the optimal speed,pause time and liquid production.The experimental results show that the dissertation proposes energy-saving algorithm for oil pumping based on BP neural network to improve the matching degree between the working state of the pumping unit and the characteristics of the oil well,and reduces unnecessary electrical energy loss,and realizes high production and high efficiency of oilfield production.