A noteworthy caprock failure occurred on the Joslyn Steam Assisted Gravity Drainage (SAGD) project in 2006 that continues to have a significant impact on the approval process for future SAGD projects. Two major reports were released by Total E&P Canada Ltd. as operator and Alberta Energy Regulatory (AER), respectively. A number of potential mechanisms were postulated within those studies, but without a definitive resolution. Inclusion of a fractured medium in the assessment of caprock integrity has not been extensively studied for detection of failure modes. The objective of this paper is to explore the effects of existence of discontinuities (e.g. fractures or fissures) in caprock, loading conditions, and steam chamber evolution, on surface heave, joint normal and shear displacements. Different modes of failure under various scenarios are presented for fissured and non-fissured caprocks.

In this paper, a distinct element code was utilized to simulate the possible mechanisms of caprock failure during SAGD operation with various fracture sets in the Clearwater Formation caprock. Three-dimensional numerical models, including caprock and overburden, were simulated under different load conditions to evaluate the impact of steam injection pressure. The lower bound for maximum operating pressure (MOP) was based on the current AER formula and the upper bound was the injection pressure prior to caprock failure. Multiple realizations of fracture network in caprock were executed to reflect various geomechanical and geometrical properties of fractures. The results were compared with a previous study performed with the assumption of a continuum medium for a non-fissured caprock.

For upper bound MOP conditions, the computed maximum vertical displacements at the base of caprock for models assuming 1) no fractures, 2) low fracture intensity, and 3) high fracture intensity were 79, 74 and 68 cm, respectively. It was observed that an increase in fracture intensity results in a reduction in vertical displacement at the base of caprock as well as surface heave. These variations in behavior are significant and illustrate that the assumption of a non-fractured caprock (in caprock integrity studies) may lead to conservative estimates of steam containment and ultimately, underestimation of the risk for caprock failure.

At the base of caprock and under lower bound MOP conditions, a few local shear failure zones occurred above the pressurized zone, while for upper bound MOP conditions, larger zones of both shear and tensile failures were computed. It was also noted that the existence of fractures could cause local shear failure in the caprock, even below AER mandated values of MOP. Lastly, the findings of this study, including geomechanical simulations, uncertainties, and risk associated with evaluating caprock containment of SAGD operations were compared with previous studies.

The results offer significant insight into our geomechanical understanding of the process in order to avoid a potential caprock failure during thermal projects, as unfortunately was experienced in the Joslyn SAGD steam release incident.

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