Dynamics of Gas Kick Migration in the Annulus While Drilling/Circulating

Abstract

Multiphase flow pattern during gas kick is made more complex while drilling or circulating out the kick (dynamic conditions). The additional pressure losses due to friction when circulating significantly change the flow pattern in the annulus. This pattern also evolves as the fluid migrates up the wellbore due to changing in-situ conditions and fluid properties. The effects of this scenario on flow pattern evolution have been investigated using water as the continuous phase, air, and carbon dioxide as the kicked fluid. Experiments were carried out in a 140 ft high tower lab fitted with pressure gauges and digital cameras for visualization. We triggered gas kicks at 80 psi and 90 psi injection pressures and, an average liquid flow rate of about 7-gpm and 15-gpm. The gas rate ranged from about 0.05 ft3/min to 0.5 ft3/min for both air and carbon dioxide injection. The gas injection time ranged from 30 to 500 seconds for air and carbon dioxide to simulate kicks of different gas-liquid mass ratios of about 0.1 ft3/ft3 to 0.3 ft3/ft3. We observed that the fluid distribution, pressure gradient in the annulus of a 2.875-in drill pipe and a 5.5-in outer casing is more complex when circulating drilling fluid than in a shut-in scenario. The effect of the initial kick pressure on the initial flow pattern observed (Taylor bubble) can be considered negligible. This is due to the fixed 2 in diameter of the gas injection line. The average gas-liquid flow rate and the duration of gas injection significantly affect the flow pattern observed after the “Initial Taylor bubble” (in space). The complexity of this flow behavior is more significant during carbon dioxide kicks than air kicks due to the solubility of carbon dioxide in water. The turbulence following the initial Taylor bubble increased with the average liquid flow rate. This is due to the additional momentum from liquid flow. Similarly, the duration of the kick did not affect the initial Taylor bubble observed but the “After Taylor bubble” flow significantly. These observations are more pronounced during carbon dioxide injection when the liquid is saturated, towards the surface as carbon dioxide begins to come out of solution. This study shows the need to account for the “After Taylor bubble” flow effects when modeling gas kick behavior. Incorporating the physics of this scenario will significantly improve gas kick models and blow-out mitigation in drilling time.