Unrisked stochastic estimates of offshore Gulf of Mexico reserves through hindcasting: How Monte Carlo and fuzzy arithmetic methods can improve initial estimates .

摘要

An investigation into several reserves estimate methods for the offshore Gulf of Mexico fields, Tiger Shoal and Starfak, was undertaken from the unique perspective of hindsight. Data for two fault segments from Tiger Shoal and three fault segments from Starfak were analyzed using three distinct methods for reserves determination. Monte Carlo simulation was used to model continuous sampling of determined distributions, and to incorporate dependency between variables. Microsoft Excel was used to discretely sample actual calculated values. And a relatively new method to the petroleum industry, Fuzzy Arithmetic, was used to model reserves possibilities for the fault segments. Simulation with the above methods led to the following conclusions: (1) Initial reserves estimates using the above methods were much closer to the observed ultimate production for the fault segments than those determined by Texaco; (2) There is a small effect on the final results when dependency between variables is taken into account; (3) There was little difference between discrete sampling as opposed to continuous sampling, both methods arrived at approximately the same answer; (4) The flexibility in Fuzzy Arithmetic makes it a promising and useful tool to add to the simulation arena.