𝗘𝗻𝗵𝗮𝗻𝗰𝗶𝗻𝗴 𝗘𝗠𝗦 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗧𝗵𝗿𝗼𝘂𝗴𝗵 𝗘𝘅𝘁𝗲𝗿𝗻𝗮𝗹 𝗡𝗲𝘁𝘄𝗼𝗿𝗸 𝗠𝗼𝗱𝗲𝗹 𝗥𝗲𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗮𝘁 Duquesne Light Company (DLC)

𝗘𝗻𝗵𝗮𝗻𝗰𝗶𝗻𝗴 𝗘𝗠𝗦 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗧𝗵𝗿𝗼𝘂𝗴𝗵 𝗘𝘅𝘁𝗲𝗿𝗻𝗮𝗹 𝗡𝗲𝘁𝘄𝗼𝗿𝗸 𝗠𝗼𝗱𝗲𝗹 𝗥𝗲𝗱𝘂𝗰𝘁𝗶𝗼𝗻
𝗜𝗻𝘀𝗽𝗶𝗿𝗲𝗱 𝗯𝘆 𝘁𝗵𝗲 𝘀𝘂𝗰𝗰𝗲𝘀𝘀𝗳𝘂𝗹 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 𝗮𝘁 Duquesne Light Company (DLC)

The Opportunity: Simplifying Network Models to Improve Real-Time Reliability

Electric utilities rely on Energy Management Systems (EMS) to provide accurate, real-time insights into the condition of the power grid. A core component of this is the State Estimator (SE), which calculates system status using real-time data and a detailed network model. However, many utilities over-model external systems—those parts of the grid outside their ownership—which leads to:

• Slower EMS performance
• Unnecessary maintenance complexity
• Increased data dependency on external entities
• Reduced system observability and responsiveness

This presents an opportunity to streamline EMS performance by reducing the complexity of external network modeling, without sacrificing accuracy.

Our Approach: Targeted Model Reduction Based on Operational Impact.

Utilicast leverages a proven, practical approach inspired by DLC, which achieved a 53% reduction in modeled external elements and dramatically improved EMS performance.

Key Steps in the Approach:

1. Boundary Identification – We define “core,” “near-neighbor,” and “external” zones based on regulatory guidance (e.g., NERC TOP-001).
2. Impact Assessment Using Line Outage Distribution Factors (LODFs) – We simulate line outages to identify external components that meaningfully affect the internal system. Only significant components are retained.
3. Model Reduction & Equivalencing – External components with low impact are replaced by simplified equivalents, using real telemetry where applicable.
4. Validation & Monitoring – Automated validation and custom dashboards track estimation quality, model health, and convergence performance across the network.

𝗧𝗵𝗲 𝗩𝗮𝗹𝘂𝗲: 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲, 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆, 𝗮𝗻𝗱 𝗖𝗼𝗺𝗽𝗹𝗶𝗮𝗻𝗰𝗲
By simplifying external network models, utilities realize:

• Improved State Estimation Accuracy (6% increase in SE quality at DLC)
• Faster System Response (11% fewer iterations required)
• Reduced Operational Complexity
• Easier Model Maintenance
• Alignment with NERC Standards
• Enhanced Situational Awareness through targeted dashboards

Duquesne Light Company’s experience demonstrated that a streamlined model leads to a more reliable and resilient grid, with fewer errors, less volatility, and greater operational confidence.

Utilicaster, Kiamran Radjabli, and Chad Hirsh from DLC, co-wrote and presented a paper at EEEIC (European chapter of IEEE) in early July detailing all this.