Influence of rock failure behaviour on predictions in sand production problems

Abstract

The mechanical strength of rock in terms of shear or compressive failure has been previously adopted as a criterion for sand production and when used solely has been proven to overestimate the process. On the other hand, ignoring the mechanical strength behaviour of the material increases the tendency for inaccurate estimations of the erosion process. In this work, an equally proportionated inclusion of the mechanical strength and erosion-based criteria in sanding predictions is proposed and assessed by numerical models. Several rock failure models and their influences on the sanding process have been analysed, including models such as the Drucker–Prager (DP), the Drucker–Prager hardening (DP hardening), the Mohr–Coulomb (MC) and the Mohr–Coulomb softening (MC softening). Modelling outcomes show distinct differences in rock response to operating and boundary conditions (e.g. flow rate and drawdown), and predictions of sand production. It was confirmed by modelling results that despite the low magnitude of stresses and strains developed at the well face and perforation regions, post-yield hardening behaviour increases the estimation of the amount and intensity of sand production. Also, incorporating a post-yield softening behaviour increases the magnitude of stresses and strains; however, this effect is observed to have a negligible impact on sand production. The role of void ratio has been recognised as a dominant factor, as its evolution significantly determines the pattern and intensity of sand production. A more cautious selection and rigorous coupling of rock strength models in sand production modelling is therefore essential if accuracy of predictions is to be improved.