Research On Well Productivity Prediction Model Of Ultra-low Permeable Reservoir With Hydraulic Fracturing

This thesis analyzed formation characteristics of three typical ultra-low permeable reservoirs,Bakken formation in Williston Basin and Yanchang formation in Erdos Basin and Qingshankou formation in Songliao Basin.Characteristics of ultra-low permeable reservoir concluding reservoir rock extra-high heterogeneity,ultra-low permeability and porosity,abnormal formation pressure results to complexity of oil and water relationship,and it is difficult to predict productivity after hydraulic fracturing stimulation treatment.Due to the ultra-low permeability of these reservoir,this thesis took the threshold pressure gradient and stress sensitivity into consideration,which result in the fluid filtration nonlinearity.According to take use of rock tensile strength a criteria of fracture initial crack to describe the initial crack and extension of the fractrure during stimulation,and analyzed the factors that influence the perforation completion well on bottom hole pressure and recommended a reasonable value range of perforating parameters,like perforation azimuth,eye diameter and shooting density by the result of numerical simulation.Additionally,this thesis simulated well dynamic productivity in the initial stage by reservoir numerical simulation also.By adopting the prediction chart method of Agarwal type curve,build a model of productivity prediction with fracturing in ultra-low permeable reservoir,which considered threshold pressure and stress sensitivity,and drew the corresponding curve chart and programming corresponding software based on Browser/Server mode by using language VB.NET and JavaScript.The model in this thesis compensate the classical stimulation ratio prediction method of Mcguire & Sikova,which adapt to steady state and quasi steady state,and Agarwal type curve,which didn’t consider the effect of threshold pressure gradient and stress sensitivity,this computational method is very simple and convenient and the parameters needed easy to access.By collecting construction design and production data of four wells to validate the model just mentioned.Compared to the results,we have proved this model within ±15% relative error,which provide a productivity prediction method for the field practice.