Soft-sensor Modeling And Control Simulation Of Water-mixed System Based On Neural Network

In recent years,soft sensing technology has become a focus of development.The reason for its high popularity is that the complexity of modern industry has increas ed,and with it,it has put forward higher requirements for control systems.The biggest advantage of soft sensing is to solve some problems that can not be solved online,so it will occupy a place in the control field and pay more and more attention.An oilfield is located in the alpine region of Northeast China,with an annual minimumt emperature of-39.2 degrees.High pour point,high viscosity "double high crude "commonly used oil well production technology is the combination of crude oil and water for gathering and transportation.Referring to the environment of the targe t oil field,a reasonable water collection and oi l gathering process is needed to stabilize production and increa se effi ciency.The process require sautomatic adjustment of water content in the well.Many factors wil l affect the water consumption.There are many shortcomings based on the single artificial adjustment method.In this paper,we first study the soft sensor modeling problem of water blending process,and build the model of water blending process based on neural network.Combined with water mixing model,the water mixing cont rol scheme was improved.Ba sed on the PID control strategy,the automatic control of water mixing process was realized.The opt imal control parameters were det ermined with the specific producti on process,and the real-time and acc urate aut omat ic regulation of the water content of the single well was realized.On the basis of simulation,the feasibility of the design method is verified,and the accurate prediction of water flow is realized.The scheme can be applied to the process of oil gat hering and water mixing in oil wells,and can opti mize the parameters of water mixing,thereby achieving the actual effect of reducing energy consumption.