Rearch On Intelligent Adjustment Method Of Frequency Of Stroke Of Pumping Well Based On Fuzzy Neural Network

At present, most of oil fields in our country have entered the mid-late development, and a number of pumping wells always run in a state, which the pump efficiency is low or air pump. The consequences of that are reducing the power factor, wasting electric energy and others. In order to make the ability of liquid supply in wells and the swabbing in pumping units match, during actual production process, the measures of adjusting the frequency of stroke mainly have two kinds. One is the replacement of motor output shaft pulley mechanically. The other is to apply frequency conversion control technology to the oilfield devices, by adjusting the output frequency of frequency converter to change frequency of stroke. They both rely on the operators'rich experience. On the premise of guaranteeing liquid production, they observe the running situation of pumping units to adjust frequency of stroke. But artificial regulation has many drawbacks, so it is of much importance in theoretical and realistic meaning on how to control the beam pumping unit system reasonably to ensure that low production wells work steadily and continuously, and enhance the automation of the system.The fuzzy neural network combines with the characteristics of fuzzy logic and neural network, which has privilege in dealing with nonlinear, fuzzy and other questions. It is a algorithm that gathers study, imagination, recognition and information processing. Because oil field has problems such as nonlinearity and uncertainty, for improving the automation level of oil field, this paper adopts fuzzy neural network to learn operators'experience for frequency of stroke deciding, on the basis of careful study of production process of beam pumping units and manual operation experience. And then an intelligent adjustment system of frequency of stroke is designed. It can achieve the purpose of stable production and reducing consumption. At the same time, it ensures that the production of oilfield will not be affected by different operator's experience.Firstly, about the decision part of frequency of stroke, this paper expounds the basic principle and design method of fuzzy neural network. Then through comparing analysis of network structure, fuzzy neural network based on the zero order T-S fuzzy model is confirmed. Grey relation analysis (GRA) is studied to select reasonable variables as the inputs of network. And Deng relation degree is employed to analyse multiple sets of pumping unit operation parameters. At last, make sure the current, pump efficiency and submergence depth as the inputs of fuzzy neural network.Secondly, the particle swarm optimization (PSO) is studied on parameters optimization. And in the light of the characteristics of high particle dimension in this paper, an improved particle swarm optimization algorithm (AHPSO) is proposed on basis of standard particle swarm optimization. It can adaptively change inertia weight and learning factors according to the iteration. On the other hand, in order to prevent the loss of diversity of particles at the same time, the selection and mutation operator of genetic algorithm are introduced to improve the ability of algorithm to jump out of local optimum for global search. Through experimental comparison, the use of the improved particle swarm algorithm to optimize the network, accuracy and speed of iteration both have been increased significantly, and then the optimal fuzzy neural network is set up to decide frequency of stroke.Finally, about the control part of frequency of stroke. A closed-loop control system of frequency of stroke is established with PID control theory that is widely used in current industry fields, based on frequency conversion control technology in oilfield system. The frequency of stroke decided by fuzzy neural network according to the working condition of pumping well acts the given value. The simulation results show that the frequency of stroke can track the given value, and it is feasible that PID is used to control frequency of stroke in the oilfield pumping system.