Fracture dimensions and process mechanisms that resulted from cyclic steam stimulation above fracturing pressure in German and Canadian tar sands are analyzed with numerical modeling. Horizontal fracture radii of 100 ft [30.5 m] and vertical fracture half-lengths of 250 ft [76.2 m] with fracture surface areas of 30,000 to 70,000 sq ft [2787 to 6503 m 2 ] are sufficient to reproduce steam injectivity into reservoirs containing highly viscous oil and negligible amounts of movable water saturation. Mechanisms that are important during fluid loading and unloading of induced fractures include thermal expansion of tar oil, countercurrent imbibition of water and oil caused by capillary pressure effects, and fracture compressibility.
Induced fractures (i.e., thin zones or microchannels of highly permeable, unconsolidated clastic material formed by pressure parting) provide an important heat-injection mechanism in tar sands because fracturing strongly influences the shape of the heated zone in the reservoir and controls the amount of tar oil that can be contacted by steam. Very little has been published concerning the nature of induced fractures in either the cyclic steam process or in the continuous steam-injection process or in the continuous steam-injection process. process. This paper describes the characteristics of hydraulically induced fractures in tar sands that have been used in a compositional thermal simulator to reproduce observed field bottomhole pressures (BHP's) and fluid rates during cyclic steam injection and production operations. The described features are offered as a possible and approximate set of characteristics that are consistent with a reasonable set of assumptions concerning the manner in which a loosely consolidated porous medium fails in response to induced stresses.
Because of the proprietary nature of the Canadian field test, the majority of discussion concerns the German experience. In these projects, however, the mode of analysis and results were similar.
The TDC STEAMFLOOD simulator was used to analyze the fractures that were hydraulically induced in the field thermal projects described in this paper. This simulator has been used widely to design and to evaluate field-scale projects that involve steam distillation, 6 steam additives, 7 projects that involve steam distillation, 6 steam additives, 7 and induced fractures.
The TDC STEAMFLOOD model is a fully implicit finite-difference-based program designed to run in either the isothermal or thermal mode and can treat up to four mass conservation equations and an energy balance. This model has three phases: aqueous, liquid hydrocarbon, and vapor. The four mass species are water, a nonvolatile heavy oil, and two volatile components. Only water and a dead-oil component were used during analysis of the German reservoir because steam distillation effects are insignificant in the heavy tar oil.
The model takes advantage of vectorized computer technology. It was used in two-dimensional (2D) vertical (cross-section) simulations with radial (r, z) coordinates for the horizontal fracture analyses and in three-dimensional (3D) simulations with cylindrical (r, 0, z) coordinates for the work with vertical fractures. Relative-permeability functions were changed between the injection and production stages of the cyclic steam process to model important imbibition and drainage wetting-phase hysteresis effects that have both an empiricals and experimental basis. Conoco recently reported the use of the TDC STEAMFLOOD model to history match and to optimize their fracture-assisted steamflood pilot activity in south Texas tar sands.