As data processing capabilities improve electronic device performance, it becomes necessary for Oil and Gas operators to understand how computing power can be harnessed at the extremities of their production network by deploying Edge Analytics solutions. This paper will discuss adapting the Industrial Internet of Things (IIoT) in the Upstream Automation domain, specifically in Artificial Lift assisted production, and will shed light on how Edge Analytics can be leveraged to deploy Machine Learning Models (MLMs) directly at the well-head.

With the advent of IIoT enabled Edge Analytics, and its ability to run Machine Learning based inference at the extremities of a production network, it has become essential to enable Operators and Subject Matter Experts to transfer their knowledge to Edge Computing Devices. This paper discusses the application of Edge Analytics enabled Augmented Intelligence for wells operated by Electric Submersible Pumps, where Machine Learning and Pattern Recognition Models help detect anomalous events in multivariate time-series data. These Models runs on Edge Computing Devices where identify newly discovered and well known ESP performance patterns that can be labelled by a Subject Matter Expert. Once these patterns are identified and tagged, the Models are retrained and pushed back to the Edge Computing Device, where they continue to detect and predict patterns in real-time.

With the advent of robust and powerful IIoT Edge Devices, it is now possible to deploy Machine Learning models at the boundaries of a production network, i.e. using Smart Nodes directly at the wellhead that runs analytics in near real-time. These Smart Nodes, when paired with Augmented Intelligence capabilities, allow subject matter experts to interact with Machine Learning models and help improve their accuracy over time. This, in turn, helps increase confidence in data-based results and enables operators to make informed decisions. This paper will define and discuss an end-to-end architecture on how Augmented Intelligence, in tandem with Edge Analytics, can be implemented in the upstream production environment. Results, methodologies and lessons learnt from an Edge Analytics solution deployed on Rod Pump wells will be discussed in this paper.