An effective criterion to prevent injection test numerical simulation from spurious oscillations

Injection/fall-off tests are one of the most promising alternatives to the conventional production/build-up sequence because they eliminate surface emissions and can significantly reduce testing costs. This kind of test is characterized by the presence of two mobile phases, the fluid originally in place (hydrocarbon) and the injected fluid (diesel, brine or nitrogen). The conventional analytical approach used to describe the transient pressure behavior is no longer suitable due to the variations in fluid saturations during the test. Although applicable in theory, the analytical approach often implies excessive simplifications of the real system behavior, such as piston-like displacement. Thus only numerical simulations can thoroughly describe the phenomena occurring during the injection process. However, the pressure and pressure derivative response calculated numerically often show non-physical oscillations during the radial flow phase, when the pressure derivative is expected to be horizontal. It was found that these spurious oscillations arise in convection-dominated problems and are associated with sharp saturation fronts. In this paper an effective methodology, based on an adaptive time-step calculation, is presented so as to avoid pressure oscillations. The proposed time-step selection is both computationally efficient and suitable to capture the physics of the system.