Multisegment production profile models — A tool for enhanced total value chain analysis Nils F. Haavardsson ⁎ , Arne Bang Huseby 1 University of Oslo, Department of Mathematics, Postbox 1053 Blinderen, NO-0316 Oslo, Norway Received 18 October 2006; received in revised form 24 January 2007; accepted 7 February 2007 Abstract In the development phase of an oil or gas field, it is crucial to have a satisfactory model for the production. Since the first attempts in the 1940's, many different models have been developed for this purpose. Such a model typically incorporates knowledge about the properties of the reservoir. When used in a total value chain analysis, however, also economic and strategic factors need to be taken into account. In order to do this, more flexible modelling tools are needed. In this paper we demonstrate how this can be done using hybrid system models. In such models the production is modelled using ordinary differential equations representing both the reservoir dynamics as well as strategic control variables. The approach also allows us to break the production model into a sequence of segments. Thus, it is possible to represent various discrete events affecting the production in different ways. The flexibility of the modelling framework makes it possible to obtain realistic approximations to real-life production profiles. As the calculations can be done very efficiently, uncertainty may be added to the framework using Monte Carlo simulation. The proposed framework constitutes an important building block in total value chain analysis, that may be incorporated in a full scale analysis of a project. In such an analysis revenues, costs and investments are modelled to obtain assessments of project profitability and different strategies. As the focus of the present paper is on production profile modelling, such a full scale analysis will not be done here. © 2007 Elsevier B.V. All rights reserved. Keywords: Production profile models; Decline curve analysis; Total value chain analysis; Hybrid systems; Monte Carlo simulation; Ordinary differential equations 1. Introduction Estimating reserves and predicting production in oil and gas reservoirs have been studied extensively over many years. Many different models and methods have been suggested. A popular technique is the decline curve analysis approach. This dates back to the pioneer paper by Arps (1945) where the exponential, hyperbolic and harmonic curves were introduced. More recent papers consider other types of decline curves, and attempt to model the relation between the parameters of the curves and geological quantities, see e.g., Li and Horne (2003, 2005, 2006), and Marhaendrajana and Blasingame (2001). The purpose of the present paper is to develop pro- duction models that can be used in the broader context of a total value chain analysis. In a total value chain analysis the reservoir geology may be described by structure models, sedimentary models and saturation models. Stochastic models, combined with reservoir simulation, are applied to estimate the quantitative measures Stock Tank Original Oil In Place, Original Gas In Place and the Journal of Petroleum Science and Engineering 58 (2007) 325 – 338 www.elsevier.com/locate/petrol ⁎ Corresponding author. Tel.: +47 22 85 58 84. E-mail addresses: nilsfri@math.uio.no (N.F. Haavardsson), arne@math.uio.no (A.B. Huseby). 1 Tel.: +47 22 85 58 60. 0920-4105/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.petrol.2007.02.003