The onset of sand production in a hollow cylinder test is evaluated by three different sanding models, i.e. shear failure, cohesive tensile failure, and EPS (Equivalent Plastic Strain) failure. The comparison of the results with experimental results from a hollow cylinder test shows that the shear failure model provides the most conservative prediction, while the EPS can provide the closest results to those from the test.


Sand prediction has been an important issue for sand control, production enhancement, well completion and applied mechanics. Before the 70’s, most studies were associated with sand control. After that the interest in determining onset of sand production was increasing and a Mohr-Coulomb type shear failure was implicitly postulated as the sanding mechanism [1,2,3]. Ewy et al. [4] proposed a Lade model to overcome the possible problems of conservative or optimistic predictions by either the Mohr-Coloumb or the Drucker–Prager criteria. Bratli and Risnes [5,6,7] introduced a cohesive failure model for sanding and calculated a critical bottom hole pressure for sanding in a uniformly stressed cylindrical wellbore in an oil formation and such a study was later extended to gas formations [8,9,10,11]. Rather than focusing on critical stresses, Morita et al. [12,13,14] proposed the so-called Equivalent Plastic Strain (EPS) criterion for oil formations, stating that sanding occurs once a critical plastic strain is achieved. This model has also been applied to the gas formation by Wang et al. [15]. These three criteria, which reflect three different sanding mechanisms, are extensively applied to sanding problems in the field and consequently different critical sand pressures are generated. Using stress-oriented criteria is usually simple and requires less parameters but a criterion as such also have the drawback that they are too conservative and more importantly that stresses cannot be measured and monitored directly.

This content is only available via PDF.
You can access this article if you purchase or spend a download.