Upscaling Transmissitivity in the near-well region for numerical simulation: A comparison on uncertainty propagation

Abstract

Upscaling transmissivity near the wellbore is expected to be useful for well performance prediction. This article ascertains the need of upscaling by comparing several numerical schemes and presents an approach to upscale transmissivity in the near-well region. This approach extends the Laplacian method with skin, which was successfully applied to the parallelepiped flow case, to radial flow case in the vicinity of wellbore through a nonuniform gridding technique. Several synthetic fields with different stochastic models are chosen to check the efficiency of this method. Both flow and transport simulations are carried out in finite heterogeneous confined aquifers to evaluate the results. It is demonstrated that the proposed method improves the ability of predicting well discharge or recharge and solute transport on the coarse scale in comparison with other schemes by examining the uncertainty propagation due to upscaling.