A review on application of data-driven models in hydrocarbon production forecast

Abstract

The accurate estimation of production is the bottleneck technique that constraints the efficient development of oil and gas fields. However, such multivariate and asymmetric reservoir parameters and highly nonlinear fluid flow behavior stake a stringent claim for precise production forecast, which makes semi-analytical modeling and numerical simulation techniques expose challenges. Based on the applications of data modeling methods in the prediction of oil and gas production, this paper proposes the procedures of data-driven models for multivariate oil field data with small samples. In addition, the strengths, weaknesses and limitations of widely used data-driven models and their combination models are analyzed in detail, and the experiences and lessons in oil and gas production prediction are summarized based on the applications of data-driven models in oilfield cases. Furthermore, the data modeling method for flow equations with complex boundary and mechanism will be a challenge and future direction to make production predictions more quickly and accurately.