Duke Energy Methane Emissions Monitoring

Context

Duke Energy is one of the largest energy holding companies in the United States, and they had set an ambitious target: net-zero methane emissions from their natural gas operations by 2030. The problem was that their existing approach to emissions monitoring relied on calculated estimations rather than actual measurements. Pipelines, meters, and distribution infrastructure spread across vast geographic areas were historically difficult to monitor with any frequency, leaving the company with a fuzzy picture of where emissions were actually occurring.

While at Avanade, we partnered with Duke Energy, Accenture, and Microsoft to build something that had not existed before in the natural gas utility space: an end-to-end platform that combined satellite monitoring, ground-level sensing, AI-powered analytics, and field response workflows into a single system capable of detecting, measuring, locating, and remediating methane emissions in near-real time.

Challenge

Natural gas infrastructure is inherently distributed. Thousands of miles of pipeline, tens of thousands of meters and connection points, and countless potential emission sources spread across regions that cannot be physically inspected with any practical frequency. The traditional approach -- periodic manual surveys supplemented by engineering calculations -- meant that leaks could persist for weeks or months before detection. And without precise measurement data, it was impossible to prioritize remediation efforts based on actual impact.

The technical challenge was multi-layered. Satellite monitoring provides broad coverage but limited precision. Ground-level sensors provide precision but limited coverage. Neither data source alone tells the full story. We needed a platform that could ingest both, apply analytics to reconcile and enrich the data, and produce actionable intelligence that field workers could act on immediately.

Approach

We built the platform on Microsoft Azure, using Dynamics 365 for the operational workflow layer and a custom analytics pipeline for the emissions data.

The data architecture ingested feeds from two fundamentally different sources. Satellite monitoring provided periodic, wide-area scans that could identify potential emission events across the entire infrastructure footprint. Ground-level sensing networks provided continuous, localized measurement at fixed points throughout the distribution system. The AI layer reconciled these inputs -- correlating satellite anomalies with ground-truth measurements, filtering false positives, and producing confidence-scored emission events with pinpoint geolocation.

The field response workflow was where the platform delivered its most tangible value. When the system identified an emission event, it generated a work order with precise location data, estimated severity, and recommended remediation approach. Field workers could deploy with the right equipment to the right location immediately, rather than spending hours on manual leak surveys. The near-real-time feedback loop meant that remediation could be verified by the same sensing infrastructure that detected the original emission.

Results

The platform gave Duke Energy something they had never had: a measurement-based view of their actual methane emissions rather than a calculation-based estimate. This distinction matters enormously -- you cannot manage what you cannot measure, and you cannot credibly commit to net-zero targets based on engineering assumptions alone.

The potential impact extends well beyond Duke Energy. The platform architecture was designed to be applicable across the entire natural gas supply chain -- from production facilities to mid-stream pipelines to local utilities. As regulatory pressure and climate commitments accelerate across the energy sector, the ability to detect, measure, and prove remediation of methane emissions will shift from competitive advantage to operational requirement.

This engagement represented the kind of work that drew me to Avanade's partnership with Microsoft -- the combination of cloud infrastructure, AI, and operational technology applied to problems where the stakes are measured in environmental impact, not just business metrics.