Reservoir development optimization under uncertainty for infill well placement in brownfield redevelopment

Abstract

In an uncertain oil price environment, brownfield redevelopment is becoming an increasingly attractive option to manage production decline. One of the essential strategies in brownfield redevelopment is optimally placing the infill well to maximize recovery from the field and minimize operational expenditure. Optimal decisions are the ones that maximize the recovery and keep the cost down. Robust optimal decisions are the ones that would hold in the presence of geological uncertainty, such as when the realization of reservoir setting deviates from the base case. The decision optimized for the base case would still be optimal for the updated reservoir condition. Hence, optimization across multiple realizations is required. While attempts on the optimization across multiple realizations have been conducted in the literature, the computational cost involved in performing the task and resulting in the proper estimation of uncertainty remains a great challenge.This paper introduces a new workflow for robust and reliable well placement optimization under geological uncertainty. The proposed workflow combines multi-objective assisted history matching, Bayesian posterior inference, and well placement optimization in multi-objective setting across multiple geological models. It is applied to an industry-standard reservoir benchmark case study and compared with several uncertainty quantification approaches conventionally used for decision making.The proposed workflow provides robust and reliable optimal decisions in placing the infill well over multiple history match models. The workflow results in a good estimation of uncertainty in respect of the optimized decision. Increase in the number of flow simulations due to the optimization over the multiple realizations to cover the geological uncertainty is reduced to a manageable size.