The geometry of the cap rock of a petroleum reservoir is never exactly known so that the depth and lateral location of the spill point are uncertain. Therefore, the gross rock volume of a filled structure is uncertain due to uncertainty in cap rock geometry and the depth to the hydrocarbon contact determined by the spill point. A method for evaluating this uncertainty has been developed. First, the geometry of the cap rock is simulated using established Monte Carlo simulation techniques for surfaces based on Gaussian random field models. Secondly, a new fast algorithm finds locations of spill points and trapping areas of the simulated structures. Finally, gross rock volume of the traps can be calculated and volume distributions can be quantified in terms of histograms and quantities. The spill point detection algorithm is fast and reliably even on complex and irregular structures.
A synthetic example and an example from the North Sea illustrate the possibilities and properties of combining a fast spill point detector with Monte Carlo simulation of surfaces.
P. 723
