Leveraging IoT, Machine Learning and Cloud to Deploy at Scale Remote Monitoring Capabilities in Coal Seam Gas (CSG) Well Pads

Abstract

Abstract Coal seam gas (CSG) is a naturally occurring methane gas found in most coal seams and is similar to conventional natural gas. CSG is a major source of energy, accounting for about 15% of Australia's electricity generation. Also, CSG is an integral part of the gas industry in eastern Australia, particularly in Queensland. Origin Energy is the upstream operator for a leading joint venture in Australia and is responsible for the development of its CSG fields in the Surat and Bowen basins as well as the main transmission pipeline that transports the gas to the LNG facility on Curtis Island near Gladstone, Queensland. Origin Energy has been exploring ways to leverage modern technology solutions to increase field staff productivity, reduce operating costs per well, and improve asset reliability while also bringing more wells into production to meet growing customer demands. Field operators drive up to 100km per day and may have up to 20 well visits per day as part of a routine inspection that involves visual checks of the site to gain situational awareness. In recent years, with advancements in technologies such as edge compute (IoT), modern network connectivity options (satellite, narrow-band LTE) and Machine learning (no-code/low-code solutions) leveraging Cloud infrastructure early experiments are suggesting the implementation of safe, reliable, low-cost remote monitoring capabilities at scale is a possibility. This paper discusses how Amazon Web Services (AWS) collaborated with Origin Energy to deploy a trial focusing on remote monitoring of well pads using a camera solution. This trial captures images from the well pad at regular intervals, which when scaled out will help reduce the need for routine inspection of well pads by field operator at the current interval. This trial has also developed a data repository of all the images in the AWS Cloud and is designing a data processing pipeline to label images and use low-code/no-code ML to detect visual anomalies. In addition to the Origin Energy use cases (described in use case section 2.1 and 2.2) AWS has also helped an operator in North America (use case section 2.3) with building Machine learning models using Amazon SageMaker for anomaly detection. Hence, based on these projects it has become evident that using IoT, Machine Learning, and Cloud computing can accelerate innovation that drives down cost and increase operational safety of managing Oil & Gas assets remotely.