Digital Twin for Optimization of Wells with Intelligent Completion

Abstract

The evaluation and optimization of wells with intelligent completion, whether multilayer or multilateral, requires a deep understanding of inflow characteristics at each inflow control valve (ICV). Zonal testing is crucial for gathering fluids and reservoir data; nonetheless, it leads to production deferment, which is undesirable by most operators. In this paper, we present how a wells’ digital twins on an edge Internet of Things (IoT) device will provide real-time virtual measurements as well as ICV optimization opportunities to maximize oil production. We present a digital twin solution for a synthetic well equipped with an electric submersible pump (ESP) and two ICVs. The digital twin system is composed of two primary components together called the estimator. One aspect is a physics-based well model that accurately calculates pressure losses and flow characteristics, and the other is an iterative algorithm that employs real-time field data, encompassing production and ESP operational data to dynamically recalibrate and update the digital twin representation of the well. By capturing the well’s dynamic state, the digital twin enables the optimizer, an optimization workflow that suggests optimal ICV positions and ESP pump frequencies, aiming to maximize oil production while maintaining water-cut constraints within individual layers. To validate the approach, we rigorously tested various well scenarios. Our approach involved flow tests with different ICV positions. In addition, we conducted a comprehensive parametric analysis, considering temporal variations in water cuts and the productivity index (PI) of individual layers. Recognizing that factors such as pressure fluctuations, wellbore conditions, and reservoir dynamics significantly impact overall productivity and inflow characteristics, the estimator within the digital twin avatar of the well is automated to allow fine tuning of PI and/or water cut for each layer to recalibrate itself dynamically. The estimator adeptly captures transient changes in the well, and our results demonstrate that initial calibration efforts substantially enhance its accuracy over time. The optimizer, an extension of the digital twin model, is tested against operational constraints of oil and water production to recommend ICV positions with projected flow rates and water cuts for each layer. Our findings align closely with a physics-based simulator, validating the approach within a 10% error range. This entire digital twin workflow helps provide a consistent and reliable well monitoring mechanism for the well along with reliable recommendations for production optimization decisions. The novelty in this approach is to provide accurate real-time flowrate estimates of complex multilateral/multilayer wells with intelligent completions. The digital twin workflow provides virtual sensing that helps estimate downhole well conditions with great reliability for production management and optimization.