SMARTDRAIN: An Intelligent Workflow for Well-Placement Optimisation by Integrating Streamlines Time-of-Flight and Exhaustive Search Algorithm

Abstract

Optimal well placement remains both highly challenging and significantly important in the E&P business since they impact field development decision making. Conventionally, well placement is performed manually based on well spacing, which may not capture the effect of reservoir geology effectively, especially in cases of high reservoir heterogeneities. Modern techniques tackle this problem by treating well locations as discrete optimisation problems through reservoir simulations, and thus apply heuristic algorithms to search for optimal well locations. However, these methods require considerable computational effort, which forestall any efforts at novel techniques in searching to for global optimal solutions. This paper presents an innovative well placement optimisation workflow to minimize the calculation time of simulation using drainage volume via streamlines time-of-flight. A reservoir simulation is run for a short period of time to acquire streamlines for all proposed well locations. The time-of-flight property, along streamlines, indicates the theoretical time required for a theoretical tracer particle to move along each streamline to a producer (pressure sink). The time-of-flight, together with reservoir properties, are then used to calculate the hydrocarbon drainage volume from each producer. In which, it is the key parameter to suggest that how much hydrocarbon can move to wells with a given production period. This workflow will search for optimal well locations to maximize the hydrocarbon drainage volume with a given number of wells. The approach translates reservoir simulation to numerical matrix union optimisation, which can be carried out at an extremely fast computational speed (less than a second for a single iteration). The expedited calculation efficiency allows exhaustive search algorithms to evaluate millions of possible well combinations and can, consequently, guarantee a global optimal solution. The workflow has been conceptually proven with a synthetic 2D simulation model, providing a pattern-like scheme to mimic the conventional approach. Furthermore, it has been successfully tested with field scale reservoir simulations. The algorithm demonstrates the advantages of optimized well-placement over conventional methods without much of an increased computational burden. The workflow is also designed to be automated with a simple user-interaction via MATLAB and MS-Excel; namely, the SMARTDRAIN package. This allows engineers/geologists to implement it as a generic workflow without requiring extensive knowledge in mathematical algorithms. With such calculation efficiency and improved optimal solution, this approach can be applied as a new well placement optimisation standard that would add competitive value in field development planning and optimisation.