Visual Twin for Pipeline Leak Detection

Abstract

Abstract Objectives/Scope We describe a visual digital twin system to allow for both operation and training of a data-driven pipeline leak detection system. We show system design in terms of its data inputs and the software system which incorporates this data in real time. This system allows visualization of pipeline data and machine learning-driven leak detection in a pipeline sitting in a subsea context. The intended purpose of the system is to both train operators of the leak detection system in its use and also provide high situational awareness to those tasked with monitoring pipeline deployments. The visual digital twin system uses gaming engine technology to achieve high visual quality. We also construct a novel software system enhancement to incorporate live data streams into the gaming engine environment. This allows real-time driving of gaming engine visualization elements with which we may augment the gaming engine environment. In terms of visualization, we focus on addressing problems of large ranges of multiple scales and providing high situational awareness which minimize operator fatigue and cognitive load. We show how multiple camera views in combination with a convenient user interface can help to address these issues. We demonstrate a digital twin system for leak detection. We show its realtime operation in a gaming engine environment with the ability to instantaneously incorporate outside data sources into the visualizations. We demonstrate using simulated pipeline flow data from sensors such as pressure, temperature, etc. This is visualized in the context of a subsea pipeline on a sea floor. Given the large range of scales, we demonstrate how we can view both the full kilometer scale pipeline and smaller subsections in the context of specific sensor data streams. The overall system demonstrates a novel combination of advanced software systems which incorporates real-time data stream with visualization using a high-fidelity gaming engine. The data used represents a leak detection scenario where both operator training and situational awareness are key desired outcomes.