The Role of Depositional Facies and Uncertainty Assessments in Hydrocarbon Estimates – An Example from the Daralingie Field, Cooper Basin, South Australia, Australia
Doctor of Philosophy
University of Adelaide
In this research a procedure was developed to assess and quantify uncertainties in hydrocarbon estimates related to depositional facies, petrophysical data and gross reservoir volumes. This procedure was applied to the Daralingie Field, which is a mature gas field in the Cooper Basin, South Australia. The aim was to investigate the reasons for an unexpectedly high hydrocarbon recovery factor.
This study was conducted in four phases: reservoir characterisation, stochastic geological modeling, hydrocarbon estimation and uncertainty assessment. The objective of the reservoir characterisation phase was to conduct an integrated reservoir study utilising all available data from Daralingie Field and the surrounding area to build a conceptual geological model. A geological description, based on core analysis, of the facies in the reservoir interval in the Daralingie Field is provided. These facies were matched to their petrophysical log signatures, so facies can be defined based on logs in uncored wells. Extensive work was performed to estimate reservoir geometry by using thickness-to-width ratios plots, net gross ratio plots and modern and ancient analogues. Based on this work, nine facies maps representing the Daralingie Field depositional model are presented. The final outcome of this phase was the building of a new conceptual geological model for Daralingie Field that contains all the available data at the time of the study.
The stochastic modeling phase aimed to generate 3-D petrophysical properties models based on the conceptual geological model created in the reservoir characterisation phase. Different stochastic modeling algorithms were used to generate a range of petrophysical properties. Object based modeling algorithms were used to generate facies-based models based on specific conceptual geological models. The porosity models were generated using a facies based geostatistical algorithm. Several stochastic models constrained by well logs and facies maps were produced.
The hydrocarbon estimates were calculated using different methods such as stochastic modeling and Monte Carlo simulation. The porosity models were generated using different facies percentages while keeping the same facies geometry. The aim of this was to evaluate the impact of facies proportions on hydrocarbon estimates. Cumulative production data were also used to validate volumetric calculations for each model. In the final phase, uncertainty assessments were done to define and quantify the key uncertainties in the hydrocarbon estimates, using a newly developed technique that merges stochastic models with Monte Carlo simulation.
The final results showed that hydrocarbon estimates are highly controlled by facies proportions and the mapped reservoir gross rock volume. The high recovery factors observed in the Daralingie Field were attributed to gas influx coming from deeper Pre-Permian rocks or coming laterally from surrounding fields through faults.
During the course of this work, a geologically driven volumetric (GDV) method has been developed to produce probabilistic hydrocarbon estimates. This method integrates the stochastic modeling method with Monte Carlo simulation to generate hydrocarbon estimates. The GDV method was applied to Daralingie Field and demonstrated advantages over Monte Carlo simulation. The GDV is geologically dependent with lower uncertainty, unlike Monte Carlo estimates that are geologically independent and have higher uncertainty. The GDV is an effective and powerful method to estimate probabilistic hydrocarbons