Monte Carlo Techniques for Evaluating Producing Properties

Abstract When purchasing property, selling property or estimating future cash flow from a project, forecasts of future production rates are required. Whether those forecasts are made with detailed reservoir simulation or decline curve techniques, there will be uncertainty in the forecasts. Even if the reservoir description was exact and an accurate reservoir simulation could be made, the field development plans would be uncertain because of uncertainty in future product prices and expenses. This paper demonstrates how Monte Carlo techniques were used to generate a range of production profile distributions to estimate the economic value of a producing property. The basic methodology used decline curve equations for each well in the field to describe the future production rate of the wells. Hyperbolic decline equations were used for all wells. Gas-oil ratio was modeled by another empirical equation. The standard decline equations were modified to properly account for downtime while still honoring ultimate recovery. Downtime thus resulted in delayed production. However, a well did not always reach its expected ultimate recovery if it "failed" for mechanical reasons or if the field was abandoned for economic reasons. Well failure was modeled as a discrete distribution (i.e., the producing rate was multiplied by a value of zero or one). Another unique aspect of the spreadsheet presented here was the incorporation of mechanical and workover failure probabilities. The decline equations were modified to allow for changes in reserves and rate as a result of workovers. The expected change in reserves could have some probability of being negative if there was risk of damaging a well as a result of an unsuccessful workover. Also, a workover could have some chance of failure resulting in no change to the expected reserves. The parameters which were risked for each well were downtime, well life, well failure, workover reserves, workover incremental rate, workover failure, and workover expense. For the field, the "risked" variables were hydrocarbon prices and operating and abandonment expenses. All the uncertainty described here could not have been properly assessed without the use of Monte Carlo techniques. P. 221