Dynamic pore network simulator for modelling buoyancy-driven migration during depressurisation of heavy-oil systems

A number of vertically-oriented heavy oil depletion experiments have been conducted in recent years in an attempt to investigate the impact of gravitational forces on gas evolution during solution gas drive. Although some experimental results indirectly suggest the occurrence of gas migration during these tests (especially at slow depletion rates), a major limitation of such an interpretation is the difficulty in visualising the process in reservoir rock samples. In contrast, experimental observations using transparent glass models have proved invaluable in this context and provide a sound physical basis for modelling gravitational gas migration in gas-oil systems. The experimental observations often exhibit somewhat contradictory trends however - some studies showing dispersed gas migration, whilst others describe fingered, channelised flow - and, to date, there appears to have been little systematic effort towards modelling the wide range of behaviours seen in or inferred from laboratory tests. To this end, we present a new pore network simulator that is capable of modelling the time-dependent migration of growing gas structures. Multiple pore filling events are modelled dynamically with interface tracking allowing the full range of migratory behaviours to be reproduced, including braided migration and discontinuous dispersed flow. Simulation results are compared with experiments and are found to be in excellent agreement. Moreover, simulation results clearly show that a number of network and fluid parameters interact in a rather complex manner and as a consequence, the competition between capillarity and buoyancy produce different gas evolution patterns during pressure depletion. The implications of evolution regime on recovery from heavy oil systems undergoing depressurisation are extensively discussed. Copyright 2008, SPE/PS/CHOA International Thermal Operations and Heavy Oil Symposium.