Research Of The Injection And Production Optimization Of Reservoir Based On Robust Control

Injection development is widely used in oil industry of China as a method to achieve the goal of high and steady yield, injection and production optimization is one of the most important ways to achieve reservoir development maintained in optimal state. However it is common to see that there are always many uncertain factors such as uncertain disturbance of noise in outer circumstance(change of porosity and permeability) and uncertain disturbance of operation(aged equipment and error of operation) can lead to the fluctuate of technology parameters, resulting to the actual injection development deviate from the ideal plans, and further making the ratio of oil production. Therefore,new methods and strategies are needed to solve this water drive reservoir injection and production uncertainty optimization problem.This paper proposed the method of Robust control based on oil reservoir injection and production which take the uncertain factors in actual oil reservoir production process as the basic, and established the model of water-driving oil reservoir production which consider net present value(NPV)of production as the object function,set injection capacity of water injection reservoirs and bottom pressure of production reservoirs as the controlled variables, taking saturation as state variables, and take principle of conservation of matter and actual circumstances of oil reservoir into account regard oil reservoir impressure influent differential system and upper and lower bound of flow of each well as constraint conditions to establish the optimization mathematical model of water-oil reservoir injection and production. Considering the instability of actual oil reservoir system,this paper further established the Robust Optimization Model of water-driving oil reservoir to enhance the stability of the whole system and reduce the cost of optimization,to generate random disturbance samples which can reflect characters of oil reservoir to approximate the actual oil reservoir circumstances and take the mathematical expectation mean value of NPV as the object function of Robust Optimization Model. Then we choose the Sequential Quadratic Programming(SQP) which have the character of ideal global convergence to turn problem of non-linear restriction into unconstrained problem to get it solved,together with adjoint method to get the gradient value of object functions. This is one of the most effective methods to get gradient value,we also use Broyden Fletcher Goldfarb Shanno(BFGS) to achieve the update the Hessian Matrix. Finally we obtained the oil reservoir parameter which can make the object function reach the maximum level.The paper also carried simulation experiment based on the method we described before onto a concept model of twice injection and production with two dimensional non-flow boundaries to approve the feasibility and effectiveness of this algorithm by analyzing the result of simulation experiment. In the end,we carried on water-driving oil reservoir injection and production based on realities of some oil field blocks,and the result shows that the oil reservoir injection and production based on Robust Controlling can not only improve the effect of water injection development,but also shows the more excellent characters of robustness and stability when the oil reservoir production system is influenced by uncertain factors than the foundation schemes.