Digitalized Holistic Sand Management: Sand Production Prediction Across the Sand Face and Risk Analysis from Sand Face to Surface Facilities from Discovery to Abandonment

Abstract

Abstract This paper highlights the advancements of Sand Management where machine learning (ML) algorithms and Artificial Intelligence (AI) logics have been incorporated in a digital portal at various levels to digitalize holistic Sand management along with visualization dashboards for seamless sand production and risk predictions. This paper showcases our approach with real field experience, right from predicting the acoustic parameters to volumetric sand prediction with changing reservoir conditions (depleting reservoir pressure, increasing water and gas rates etc.) and from complex PSD transport replication to dynamic metal loss and deposition predictions over a well's life from discovery to abandonment. The paper showcases our experiences, duly validated by actual field work on: how the Sonic parameters forming the basis for the rock mechanics module were predicted using offset well data and Data science algorithms with high degree of accuracy.The Rock Mechanics module generates time based Critical Drawdown Pressure profiles which are then translated to provide time based volumetric sand index to serve as an input for the Sand Risk module to predict sand transport risks such as Metal Loss and Deposition.Models accurately matched historical trends and this synchronization between the two modules enabled future volumetric sanding index prediction with quantified Metal loss and Deposition scenarios.How the complex behavior of sand PSD travel and deposition was transformed into a correlation with variables to twin any form of sand transport behavior. Various sensitivity analyses on reservoir parameters and drawdown were conducted to determine the sand production tendency, severity and transport risks. This digital sphere formed by the interaction of the two modules provides answers and workflows to all possible sand management queries related to: sand onset and sand severity matrixbean-up operation,selection of remedial sand control measure,drawdown limits and sand severity over field life.rate and amount of sand deposition in the sump,rate and amount of sand transported to surface,sand deposition and bridging tendency inside the well as well as in the surface flowlines,metal loss predictions for choke, bends and x-mas' trees. This enabled a pro-active and field specific approach in adoption of most appropriate sand management practice to prevent costly interventions, improve hydrocarbon recovery and reduce production deferment. The digital framework for the sand management concept bridges the gap across rock mechanics, well bore and surface departments bringing them all under one umbrella for managing sand. The paper showcases how the digital modules and strategies were rigorously validated via case studies across various regions and how the pioneering application of digitalization in Sand Management provides a broader perspective than conventional mechanical sand control measures