The Montalvo field has produced 42 million barrels of oil from the 3000 foot thick non-marine fluvial Sespe formation. Pressure transient analysis suggests that individual reservoirs are less than 200 feet in width and have limited connectivity. Since detailed mapping of individual sands is virtually impossible, early development was focused on the bigger sand packages. This case study presents the methodology to pick stepout locations and evaluate bypassed reservoirs in a field that has produced for over 60 years.
With as many as 200 small-scale reservoirs at Montalvo, conventional structure and net pay maps are of limited use when picking locations. Venoco used a stochastic approach that identifies deposition trends with models that varied channel width, length, sinuosity, orientation and facies. Issues addressed include the lack of a common oil-water contact, significant variations in reservoir pressures, and transmissibility of faults. The geologic model was populated with over 1000 realizations then reduced to 133 by history matching fluid production and reservoir pressures in a dynamic model. An objective function was then used to pick the best 30 realizations for prediction runs. The authors concluded it was unrealistic to expect the model to predict recovery from individual wells, but it could identify trends and rank locations. Drilling results have confirmed this conclusion, with a wide variation in performance of the eight wells drilled since 2010 and an average estimated ultimate recovery (EUR) very close to predictions.
The methodology is especially relevant where operators hope to increase recovery in fields with a long production history, poor data resolution, multiple horizons and/or significant variation of reservoir quality. This work shows that it is possible to use reservoir modeling to identify oil reservoirs on a scale smaller than seismic resolution.