NETL experts are harnessing advanced computing capabilities including machine learning (ML) and artificial intelligence (AI) to create ways to analyze information from diverse data streams like historic records, information reported from citizen scientists, field data collections and other obscure sources to identify undocumented orphaned wells (UOWs), leading to remediation efforts that can reduce greenhouse gas emissions.

The generally accepted definition of UOWs describes them as idle wells for which the operator is unknown or insolvent. Many of these wells leak greenhouse gases into the atmosphere, which contributes to climate change.

Estimates are that as many as 800,000 UOWs dot the land across the United States, and the location and characteristics of those wells are not included in current digital records. Identifying those locations is critical to ensure that safe and effective plugging of the wells can be done to reduce risks to the environment, climate, and human health and safety.

The Appalachian Basin, with a long drilling history extending well before modern record-keeping practices were established, is expected to have a high number of UOWs. Further complicating the detective work is the fact that many properties changed owners and operators several times making it difficult to identify who is responsible for plugging and remediation efforts.

A tangle of information exists on the location of many of UOWs in the form of old land survey maps, drilling permits, historical images, production records, documented eyewitness accounts and other items. Stitching the pieces together to identify locations and understand characteristics of those wells will require the use of advanced computing techniques, and NETL experts are building new and adapting existing capabilities to address the issues.

AI uses computers to perform functions that traditionally require human intelligence because AI can process large amounts of data in ways that humans cannot. AI can recognize patterns, make decisions, and make judgments like humans. ML is a branch of AI and computer science that focuses on using data and algorithms to imitate the way that humans learn, which gradually improves accuracy. In ML, an algorithm is a procedure that runs on data to create an ML model that performs pattern recognition functions. Essentially, the algorithms learn from the data.

NETL’s Jennifer Bauer explained that “advanced computing techniques can be used to process and digitize maps and images. Using supervised machine learning to discern the marks and symbols that were used to represent wells on those maps and images can then be processed to extrapolate potential well locations. We can then compare the resurrected locations to modern records to determine if they are undocumented wells.”

On a parallel track, NETL experts are working to process well integrity testing records to extract information about how the wells were constructed, designed and performed over time. That information can then be used to prioritize plugging efforts so that the wells targeted are those that pose the most risks to the environment and human health and safety.

“Much of this research is just beginning,” Bauer said. “But unique capabilities and planned NETL facilities like the collaborative Geoscience, Environmental and Materials (GEMs) Computational facility planned for NETL in Albany (Oregon) will offer distributed computation analytical capabilities and expand connections to additional data sources and collaborators to support the development and use of advanced computing, including the application of multi-modal machine learning to identify UOWs. That’s a technique that uses multiple data sources and models that closely resembles ensemble learning.”

She said multi-modal machine learning models can be used to identify patterns and relationships that may not be apparent with traditional methods or to the human eye.

“The combination of data from different sources within the models can also offer a more complete picture of patterns of well drilling trends across the country, and how these locations relate to changes in the environment and population centers over time,” she said.

NETL’s high performance computing expertise is a key part of the Consortium Advancing Technology for Assessment of Lost Oil and Gas Wells (CATALOG), made up of Department of Energy national laboratories including Berkeley Lab, Lawrence Livermore National Laboratory, Los Alamos National Laboratory and Sandia National Laboratories. Creation of CATALOG was requested in the Bipartisan Infrastructure Law (BIL), which also provided funding to plug UOWs. BIL noted that addressing UOWs will help communities reduce methane emissions and eliminate other environmental impacts while creating jobs and advancing environmental justice.

NETL is a U.S. Department of Energy national laboratory that drives innovation and delivers technological solutions for an environmentally sustainable and prosperous energy future. By leveraging its world-class talent and research facilities, NETL is ensuring affordable, abundant and reliable energy that drives a robust economy and national security, while developing technologies to manage carbon across the full life cycle, enabling environmental sustainability for all Americans.