Angeles, Renzo (ExxonMobil Development Company) | Yin, Jichao (ExxonMobil Upstream Research Company) | Hecker, Michael T. (ExxonMobil Development Company) | Podust, Alex V. (ExxonMobil Development Company) | Dares, Keith M. (ExxonMobil Development Company) | Burdette, Jason A. (ExxonMobil Development Company) | Huang, Hao (ExxonMobil Upstream Research Company) | Spuskanyuk, Alexander (ExxonMobil Upstream Research Company)
Cased-hole fracpacks (CHFP) can deliver high-rate, low-skin completions by creating a highly conductive fracture that extends beyond the perforation tunnels, bypassing near wellbore damage and preventing formation sand production. While the industry has a long history of successful CHFP applications, well performance prediction for this type of completions has remained challenged by complex geometrical (fracture geometry and orientation with respect to arbitrarily deviated wellbores) and multi-physics factors (multiphase flow, turbulence). Most fracpack modeling tools are limited to analytical and simplified reservoir simulation models, which can lead to poor accuracy in quantifying near-wellbore effects, such as non-Darcy pressure drop, particularly important for high-rate gas wells.
In this paper, we propose a new mechanistic approach to incorporate the cased-hole fracpack completion with non-Darcy flow through explicitly meshed perforation tunnels, fractures and rock formation in real dimensions. The fracture is modeled by Enriched Finite Element Method (EFEM), which flexibly accounts for arbitrary fracture geometry and orientation while enabling multi-physics effects, impact of perforation/gravel packing damage and perforation-fracture communication uncertainty on deviated well productivities.
The proposed approach is validated using (1) analytical and numerical models, and (2) two Gulf of Mexico (GOM) CHFP wells, one vertical and one deviated; where skins measured from step-rate tests were history-matched to longitudinal and transverse fracture models. We also introduce the concept of fracture neighborhood width to account for perforation performance relative to its alignment with fracture opening and orientation. Finally, the new approach is used to predict the deliverability of a high-rate, high-pressure gas condensate well. Non-Darcy effects, condensate banking effects, perforation gravel packing, and geological model uncertainties are included in predicting the well production.
Hsu, Sheng-Yuan (ExxonMobil Upstream Research) | Searles, Kevin Howard (ExxonMobil Upstream Research) | Liang, Yueming (ExxonMobil Corporation) | Wang, Lei (ExxonMobil Upstream Research) | Dale, Bruce A. (ExxonMobil Upstream Research) | Grueschow, Eric Russell (ExxonMobil Upstream Research) | Spuskanyuk, Alexander (ExxonMobil Upstream Research) | Templeton, Elizabeth (ExxonMobil Upstream Research) | Smith, Richard James (Imperial Oil Resources Ltd.) | Lemoing, Daniel R.J. (ExxonMobil Qatar)
The Cold Lake heavy oil development is located in northeast Alberta, Canada. It began commercial operation in 1985 and uses a thermal recovery process called cyclic steam stimulation (CSS). During steaming cycles, the dilation and re-compaction that occur within the reservoir cause the overburden to deform much like the motion of flexing a thick telephone book. At weak overburden layers, shear slip plane(s) can form due to excessive shear stress overcoming the interlayer cohesion. Over multiple steaming/production cycles, the cyclic flexing and associated shear slip may lead to overburden casing fatigue failures.
In this paper, a multi-scale geomechanics modeling methodology is presented to predict the onset of failure due to CSS-related ultra low cyclic fatigue (ULCF). The modeling methodology consists of: (i) converting geological data into a representative finite element model of a single or multiple CSS pads, (ii) constructing a near-well submodel that includes thermal cement and casing, and (iii) constructing a detailed casing and connection submodel to predict the ULCF life of a pipe body or connection.
To predict the ULCF life of the casing and connection, an algorithm based on the concept of cyclic void growth is incorporated into the submodel. It provides the capability to predict the number of steam cycles to failure using the concepts of demand and capacity. This enables studying the effects of alternative steaming practices on overburden shear slip and casing/connection life.
Based on the learning from the multi-scale modeling, it is found that shear displacements on a shear slip plane can be superimposed using a single-well solution. By applying steaming and production scaling functions, the shear slip can be determined at any location and time. Integration of the single-well solution with ULCF algorithm has facilitated development of a new software tool that can be used to manage CSS operations in Cold Lake.