Probabilistic well time and cost forecasting models are widely used for estimating budgets for proposed well operations. These models are based on probabilistic simulations and require the input of the range of times for the proposed operations and the frequency of occurrence of problem time. The best source for this information is historical well data from operations in the same area and of a similar type.

Valuable well-operations time information is recorded using operation reporting systems and stored in relational databases. This information is difficult to extract and use in probabilistic well forecasts. The standard procedure is to manually extract the historical activities from offset well reports and to manually process them using spreadsheets, which can be time consuming.

This paper describes techniques for directly extracting the operation information and provides an approach to forecasting using probabilistic methods that makes it easy for the user to review the input data for planned operations in the Monte Carlo model. Methods for working out performance data for planned operations are also discussed, using the frequency of occurrence of unplanned operations, and a novel approach to statistical distributions that can be understood by operations engineers is presented.

The proposed methods reduce preparation time and improve the accuracy of the probabilistic well forecast.

You can access this article if you purchase or spend a download.