Novel Methods for Production Data Forecast Utilizing Machine Learning and Dynamic Mode Decomposition

Abstract

Multiphase flowrate measurements play an important role during the reservoir characterization and production optimization phase of reservoir management. Accurate multiphase flow rate measurement is an indispensable tool for production optimization from oil and gas fields. One of the industry's accepted solutions is the use of multiphase flow meters, which are expensive, have a limited operational envelope, and are exposed to erosion and failures. This can limit the applicability of physical metering devices due to frequent calibration, transportation issues, space, safety, security, and possible high costs. Virtual flow metering (VFM) is a method for estimating oil, gas and water flowrates produced from wells without measuring them directly. The method uses data available from the field, such as downhole pressure and temperature measurements as well as a choke position and ESP operational parameters, to estimate the flowrates by implementing hydrodynamic multiphase models, measurement data, and a reconciliation algorithm. In this paper, an overview of the conventional multiphase flow metering solutions is presented, which is followed by application of some advanced artificial intelligence and data analytics techniques for a specific case of multiphase production monitoring in a highly dynamic wellbore. The considered case refers to a typical scenario, where the measurements of oil, gas, and water flow rates are obtained in real time using a topside multiphase flow meter. Alternatively, the values of these multiphase rates are estimated using a data-driven dynamic flow model obtained using a dynamic mode decomposition technique. The results obtained with this method are compared with another VFM approach, where the rates are obtained using deep LSTM neural network.

Rigorous review of electrical submersible pump failure mechanisms and their mitigation measures

“…Recent technological advancements have allowed for the integration of advanced algorithms and machine learning to predict the occurrence of a failure and thus provide early warning. This can help provide the time needed for failure prevention and mitigation (Ujjwal et al (Jennings et al 2020) 2019; Amendola et al 2019;Gryzlov et al 2020;Teplyakov et al 2018). Artificial intelligence models that have been integrated with ESP include response surface model (Sanusi et al 2021), high performance random forests (Sneed 2017), decision trees, and gradient boosting (Wasfi et al 2019).…”

“…In order to increase recovery from oil reservoirs, several methods are implemented to support the reservoir pressure and also help mobilize the crude oil Graham et al 2017;Knight and Bebak 2000). Artificial lift involves utilization of a mechanical method to displace the crude oil from the wellbore to the surface (Cui et al 2016;Rushby and Denholm 2013;Stavale 2001;Wensheng et al 2012;Zhu et al 2016).…”