Modeling of Deep Subsurface Petroleum Biodegradation and the Effective Prediction of Petroleum Fluid Properties in Exploration and Production Settings
Larter, Steve, Jennifer Adams, Haiping Huang, Barry Bennett, Dennis Coombe,
Most of the petroleum on earth has been biologically altered(biodegraded) affecting production characteristics yet this terminal petroleum system process remains poorly understood and quantified. Many of the key elements of the compositional changes involved in biodegradation have been elucidated and petroleum geochemists have developed effective schemes for ranking different oils in terms of degradation level but our understanding of most aspects of the process are empirical. In most deep reservoir settings degradation must be anaerobic with indications that the 80o C temperature limit of biodegradation may represent a fundamental boundary of life on(or IN) our planet. Degradation may even be isochemical on a reservoir scale in some deep settings with oxidants, nutrients and oil all locally provided for organisms to feast upon. Whereas most aspects of petroleum systems are now routinely and effectively deterministically modelled in exploration settings, biodegradation, to-date, lacks an effective approach though TTI-type biodegradation indices(Yu et al(2000) and kinetic models of biodegradation have appeared(Larter et al, 2000). Our studies of biodegradation rates in-reservoir indicate that reservoir charge and temperature history and in-reservoir oil mixing are first order controls on fluid properties with oil leg and water leg proportions and topology and water and reservoir chemistry as secondary factors in successful biodegradation related fluid-property prediction assessments. We describe a new generation of effective modeling approaches that integrate the full complexity of basin charge models with a complex systems approach to modeling biodegradation in a basin modeling or reservoir simulation environment. We suggest our understanding of biodegradation and our models are already more sophisticated than our charge models can support effectively and point to improved charge history assessment as the key bottleneck component to more effective fluid property prediction. We illustrate our work with examples.