Abstract
A reliable production forecast is a critical part of the planning and decision making of companies in the oil and gas industry. The forecasts form part of a company's business and strategic plans and form the basis of evaluating an existing asset, major capital project, or exploration prospect. It therefore follows that generating a reliable and representative production forecast is a key desire of any Oil and Gas company.
There are many factors, surface and subsurface, that affect the reliability and accuracy of production forecasts. All these factors are not single-valued and would generally have a band of uncertainty around them. The challenge therefore is how to generate production forecasts in the face of these uncertainties. Previous production forecasts have been generated using deterministic values for these uncertainties at their end points – 3 forecasts. This method however, does not test the possible interactions between uncertainties which would lead to multiple production forecasts. This method, although reasonable, may sometimes lead to erroneous decisions due to optimistic or pessimistic production forecasts.
This paper describes another methodology which is currently a best practice within Chevron Corporation. This method involves the assessment of the uncertainties using design of experiments and probabilistic analysis using Monte-Carlo simulation. A design of experiment workflow will be presented and the mechanics of incorporating historical data into the workflow will also be discussed. This method enables the generation of a truly probabilistic range of forecasts which can then be used in decision making. A practical application of this method is also detailed in this paper.
