Sand Prediction: A Practical Finite-Element 3D Approach for Real Field Applications

Abstract

Summary Sand production is a critical issue in the oil and gas industry. During the production of a well, sand production may have negative consequences, such as risk of well failure, erosion of pipelines and surface facilities, and the need for sand separation and disposal. Knowing the conditions for the onset of sand production allows optimizing sand free production and, eventually, avoiding or delaying the use of sand-control methods. The aim of this work is to establish a reliable workflow for the estimation of the conditions for sand production in real field cases by means of finite-element modeling. The fundamental requirement is to set up a 3D coupled model that can be easily adjusted to the most complex conditions (e.g., stress anisotropy, deviated wells, and complex perforation patterns). The most suitable geometries and associated meshing strategies to describe the wellbore, the perforation tunnels, and the surrounding formation are analyzed. Further improvements with respect to previous approaches include the fact that the drilling and completion phases were also simulated to compute the correct stress distribution before the production, and that fluid flow and rock deformation are simulated in a fully coupled way to investigate accurately the effect of drawdown. Shear failure of reservoir rock, considered as an elastoplastic medium, is the main sand-production mechanisms analyzed, and the damage of the rock around the perforations is evaluated by analyzing the distribution of the equivalent plastic strain. Two real field cases are simulated, and the results of the finite-element models are consistent with the ones obtained by means of an analytical models and with field observations. Moreover, this numerical approach allows quantifying the spatial distribution and the severity of the damage of the rock around the perforations, facts that are either oversimplified or not considered at all in analytical models. For future applications, this model can be straightforwardly extended to more complex conditions and can also be improved to provide volumetric sand prediction.