Robust Well-Cost Estimation Using a Support Vector Machine Model

Abstract

fax 01-972-952-9435. AbstractThe preparation of cost estimates for a well is the final and most critical step in well planning. Traditionally, well costs are estimated by using a deterministic calculation procedure with time and cost inputs, or by using probabilistic methods. Because in many cases the estimate determines whether or not the well will be drilled, the credibility of the estimated cost is crucial.This paper describes the development of a support vector machine (SVM) -based application. In addition to estimating the cost, this application also indicates the percentage of accuracy in the prediction, thereby proving the credibility of the predicted cost estimate. The proposed methodology enables drilling industry personnel to estimate the risk of the project during well planning and during drilling. The analysis from such a model is based on the constraints of different drilling variables. Extensive simulations have been carried out and are reviewed in this paper. The paper also includes two case studies of well cost estimation where the SVM methodology successfully the predicted risk involved in the completion of the wells.