Infill well placement optimization in two-dimensional heterogeneous reservoirs under waterflooding using upscaling wavelet transform

Abstract

Abstract Optimal location of infill wells is very important in the development strategies of any reservoir. It is even one of the most challenging tasks for waterflooding projects in heterogeneous reservoirs. The new infill wells have to be located based on optimum reservoir management practice taking into considerations several reservoir parameters. Usually, there are several potential locations of the new infill wells. Selecting and evaluating of all the possible candidate wells using fine scale models is neither computationally straightforward nor time efficient for geological models consisting of millions of grid cells. This paper presents a practical and cost effective method that minimizes the number of simulation runs and maximizes oil recovery by optimizing infill well location. The proposed method consists of upscaling the geological reservoir model from fine scale into a coarser scale while preserving the important characteristics of the fine scale model. The proposed approach uses wavelet transform as an upscaling technique. To illustrate its feasibility and effectiveness, the proposed method is applied to three synthetic cases of two-dimensional inverted five-spot heterogeneous reservoir models. Static Dykstra-Parsons coefficient of permeability variation is utilized to compute the degree of heterogeneity of each of the three models. Permeability distribution values are used to represent and construct the reservoir models. The permeability models are upscaled from the original fine grid model to coarser scale levels. A computationally efficient and extremely time saving three-dimensional streamline simulator is used to implement the proposed approach. This paper shows that the results obtained from the upscaled models are in excellent agreement with the fine scale models. More importantly, the results also illustrate that the proposed method extensively reduces the number of simulation runs required to find the optimum well location between 75% and 93% depending on the reservoir heterogeneity. The methodology presented in this paper provides a straightforward road map for upscaling several heterogeneous reservoir models using wavelet transform approach to determine the optimum infill well location.