Research On Identification Method Of Friction Coefficient Of Rod Pumping Well Based On Genetic Algorithm

At present,the main way of oil extraction in China is rod pump.This oil recovery method has the characteristics of simple and reliable equipment and easy maintenance,but its biggest problem is that the energy loss is serious due to the low efficiency of the system,and friction energy consumption is the biggest impact on the efficiency of the system.At the same time,the system friction coefficient is also an important basis to reflect the flow of crude oil and to formulate viscosity reduction measures for heavy oil wells.In addition,with the increase of oil recovery life,friction also becomes the main factor that affects the operating conditions of the pump and determines the stability of the system.In order to improve the efficiency of the system,reduce the waste of energy and ensure the safe and stable operation of the system,the study and determination of the friction coefficient of the system is particularly important.First of all,In view of the limitation that most of the current studies only focus on the viscous damping in vertical wells.A three-dimensional model of rod pumping system with rod,tubing and liquid-column coupling vibration is established in this paper,and the solution of the model is completed.At the same time,several common deep well downhole pump conditions were simulated,which laid a foundation for the subsequent identification of friction coefficient.Secondly,the feature vector extraction of indicator graph based on contour moment is completed.In order to solve the problem that the contour moments of discrete function images are not scale-invariant,a new combined contour moment is obtained by removing the scaling transformation factor from the combination of contour moments.At the same time,a feature extraction method is proposed to show that the integral contour moment of the work graph is combined with the low-order contour moment of the partition,which preserves the sufficient global and local feature information of the pump work graph and avoids the influence of the high order moment invariant moment of the graph on the extraction effect.Then,in view of the friction coefficient identification of three-dimensional coupled vibration model is characterized by many parameters and high precision.The intelligent optimization algorithm is studied and a hybrid genetic algorithm based on simulated annealing is proposed.The friction coefficient of the system is identified by combining it with the idea of system parameter identification.At the same time,the validity of the identification results is verified by comparing the predicted liquid production with the actual liquid production by the simulation pump power diagram.Finally,the performance of hybrid genetic algorithm,classical genetic algorithm and simulated annealing algorithm are compared and analyzed.The results show that hybrid genetic algorithm has better optimization ability and convergence speed.Finally,a mobile monitoring platform for friction coefficient of rod pumping wells is built,which mainly realizes the functions of user management,oil well management,state warning,indicator diagram display,friction coefficient on-line monitoring,historical record query and so on.The platform is applied to oilfield production,and the on-line identification of friction coefficient and dynamic query of identification results are realized.The results show that the platform meets the actual production requirements in oil field.At the same time,on-line monitoring of oil well information,including friction coefficient,greatly facilitates oil field management and provides support for refined production,energy saving and emission reduction.