Research On Water Injection Flow Control Strategy Based On Fuzzy Neural Network PID

Stratified water injection is widely used in China's oil extraction industry as an effective way to stabilize reservoir pressure,improve production rate,and ensure stable oilfield production.In the stratified water injection process,the accuracy of the flow control directly affects the water injection efficiency.Therefore,accurate control of the water injection flow is extremely important for achieving fine water injection management,improving water injection efficiency,and oilfield extraction.This paper first studies the stratified water injection system and the principle of flow control to determine the control requirements and control objectives.Based on the characteristics of oilfield water injection flow and the goal of flow control,the scheme is designed.The mathematical model of the water injection flow control system is established,and the correctness of the model is verified by simulation verification and actual verification analysis of the established mathematical model.Then,based on the characteristics of non-linear and time-varying flow control objects in oilfield water injection,a fuzzy neural network PID control algorithm suitable for downhole flow control is designed.By constructing a model of water injection flow control method in Simulink,the simulation experiment of flow control is carried out,and the model is compared with conventional PID,fuzzy PID and neural network PID control.The superiority of fuzzy neural network PID control in water injection flow control is verified.Finally,the fault of flow sensor is analyzed,and the fault diagnosis method is designed to deal with the fault of flow sensor.Simulation and experimental results show that the method can effectively verify the flow sensor data,detect possible failures and estimate normal analog signals after sensor failure,further improve the accuracy of water injection flow control,and ensure more efficient and stable operation of the water injection system.