Dynamic Mode Decomposition for Virtual Flow Metering

Abstract

Summary Accurate and consistent measurements of multiphase rates allow operators to make decisions for better reservoir characterization, production monitoring and reservoir management. The availability of such measurements is strongly related to the performance of monitoring equipment which is used to obtain oil, gas and water flow rates from each well. Currently, there are two industry established approaches, where the flow rates are measured by a test separator or a multiphase flow meter. On the other hand, the increased availability of cost-efficient downhole sensors raised dramatically the amount of data obtained from a wellbore. A virtual flow meter (VFM) is a computational model which enables estimation of multiphase production rates from the available raw data without measuring flow rates directly. VFM systems can be considered as a cheaper alternative to the conventional multiphase flow measurements, as they do not require additional hardware deployment. The goal of the present work is to demonstrate the predictive capabilities of VFM using the Dynamic Mode Decomposition (DMD). A synthetic example is considered, describing multiphase slug flow in a horizontal well with a riser, where downhole pressure and temperature measurements are used to predict topside rates of liquid and gas.