Pipeline leaks are an ongoing threat to the environment across the United States. Between 2010 and 2024, more than 9,300 pipeline incidents resulted in $9 billion in property damage, There were more than 163 fatalities, and the evacuation of over 51,000 people. As more pipelines are installed and the existing pipeline infrastructure ages, leak detection becomes increasingly important. Effective proactive pipeline leak management is the key to saving lives and dollars and protecting the environment.
These leaks are usually addressed as isolated failures rather than part of a larger problem. A series of pipeline leaks in North Dakota brought the problem to light and prompted a deeper examination of pipeline monitoring and leak detection. It was clear that conventional tools were insufficient to detect leaks before they could escalate, and these so called “one-offs” were endemic of a larger problem that was not going to away unless addressed.
Therefore, a first-of-its-kind public/private initiative was launched to apply emerging technology to the science of leak detection and resolution.
Summit Midstream's single 143-day leak cost $85 million in fines and cleanup. Between 1996 and 2016, North Dakota's pipelines spilled hazardous liquids at least 85 times (an average of four spills per year), causing more than $40 million in property damage. By 2017, a documentary and a wave of media coverage had branded North Dakota "The Spills State."
In 2013, a farmer near Tioga found crude oil bubbling up from a pipeline rupture, which ultimately resulted in a loss of 20,000 barrels of oil. In 2014, a leak in a produced-water pipeline went undetected for 143 days, releasing millions of gallons of wastewater polluting nearby waterways. Thanks to the growing number of leak incidents, North Dakota earned the not-so-nice nickname, “The Spill State.”
Rather than imposing new regulations, then Governor Doug Burgum (who is now Secretary of the Interior), challenged the industry to find technological solutions capable of identifying leaks early and reliably. The result was the formation of the Intelligent Pipeline Integrity Program (iPIPE), which was launched in 2018.
iPIPE brought together energy operators, regulators, and technology companies to test emerging leak-detection tools using western North Dakota as a proving ground. The program functioned like a real-world proving ground for new leak-detection solutions.
iPIPE's goal was to adopt a new approach to pipeline management, shifting the focus from reactive cleanup to proactive detection.
The conventional approach to pipeline pressure monitoring and leak detection uses supervisory control and data acquisition (SCADA) systems. SCADA systems are most effective when a pipeline rupture causes a sudden pressure drop. They are less effective at detecting slow leaks that fall below the pressure alert thresholds.
Maintaining remote monitoring systems makes leak detection more challenging. There are thousands of miles of pipeline crisscrossing rural landscapes, making frequent physical inspections difficult and costly. The sheer size of the problem is why many leaks remain undetected.
The financial impact of even small pipeline leaks adds up quickly. Even small spills can cost as much as $5,000 per gallon to clean up, including the cost of resources, environmental impact, and regulatory fines.
In pipeline management, the cost of detecting a leak early is insignificant compared to the cost of repair and cleanup.
One of the technologies tested through iPIPE was geospatial analytics using satellite images. Pictures of pipelines and the surrounding area, taken from orbit, are analyzed using artificial intelligence (AI) to identify environmental indicators of a potential leak.
Images from low-orbiting commercial satellites can capture images with a resolution of 30-50 cm, about the size of a dinner plate. Multispectral and hyperspectral analysis using AI can detect surface anomalies, such as gas leaks, and potential environmental threats, such as nearby excavation and construction.
The advantage of using AI-powered geospatial analytics is that satellite monitoring can cover an entire operational footprint, including pipelines, well pads, and compressor stations.
One operator implemented satellite monitoring across North Dakota’s Williston Basin following a major spill. Since then, geospatial imaging and AI have identified dozens of early-stage crude oil and production water leaks.
Satellite monitoring has also detected anomalies along pipeline pathways. For example, in several cases, spectral analysis showed vegetation stress, indicating an underground pipeline leak. These types of leaks are generally undetectable using ground inspection.
Even the smallest pipeline leaks can result in millions of dollars in cleanup costs, penalties, and damage to business reputation. Compared to the costs and consequences of a pipeline leak, the investment in early-detection technology is nominal. Regular geospatial surveys alert operators to potential leaks early, enabling them to investigate quickly and address problems before they escalate into large-scale leaks.
The success of the iPIPE initiative demonstrates that public policy pressure and industry innovation can result in new solutions to environmental challenges. Pipeline networks will continue to expand, and with that expansion will require better monitoring technology.
Technologies that enable continuous pipeline monitoring at scale offer a practical approach to leak detection. The experiment that began in North Dakota not only proved the value of AI-powered geospatial analytics for leak detection but also changed the approach to pipeline management from reactive to proactive. Rather than responding to catastrophic toxic spills, energy providers are identifying risks before they become environmental disasters.
Clearly, the iPIPE initiative marks a new approach to monitoring and managing large pipeline networks. Geospatial analytics provide a practical solution for managing pipeline infrastructure, even in remote areas.
Sean Donegan is CEO of Satelytics, a company that uses cloud-based, geospatial analytics to analyze multispectral and hyperspectral imagery to identify pipeline leaks and other environmental issues. Donegan has over 30 years of experience building technology and software companies.
