History matching is a long-standing industry challenge and resource-intensive process. In this paper, we introduce a new method in the area of automated history matching algorithms, called Digitally Augmented Subsurface History Match (DASH-Sim), that focuses on calibrating the permeability property at the field level. The new approach captures the global displacement mechanism as observed in the field by capitalizing on time-lapse surveillance data.

Time-lapse surveillance data is used to derive an accurate representation of subsurface saturation as a function of time. We refer to this dynamic saturation distribution as the 4D saturation model. The DASH-Sim method minimizes the difference in saturation between the 4D model and the simulation model at all timesteps by adjusting the 3D grid-cell permeability values. A novel mathematical relationship is introduced that calculates the required permeability calibration in the simulation model as a function of the underlying rock quality and the difference in water saturation between the 4D saturation model and the simulation model. The algorithm considers the dynamic nature of saturation by successively updating the static permeability property at each timestep of the corresponding observed 4D saturation.

A synthetic case study is presented in which the new algorithm is applied. First, a truth case simulation model was run using a permeability distribution from which water saturations were recorded as a function of time. The simulated water saturation over time using the truth case permeability was treated as the observed 4D saturation model. Next, a significantly modified permeability property was replaced in the simulation model resulting in a new water saturation distribution over time. The DASH-sim algorithm was then used as an inverse history match calibration process to generate permeability modifications guided by the objective function to minimize mismatch in saturations between the observed 4D saturation model and the numerical simulation model. Results indicate a very good history match is achieved in an automated manner, reducing the turnaround time for such tasks by orders of magnitude.

The inverse problem of history matching subsurface reservoir models is a very challenging and time-consuming task. The DASH-Sim offers a new approach to automate this tedious task consistently by considering both surface-level production and subsurface time-lapse measured surveillance data.

You can access this article if you purchase or spend a download.