Summary

This paper describes the Midcontinent Independent System Operator’s (MISO) dynamic reserve by using advanced analytics to manage operational uncertainty for bulk power system operations. Motivated by growing system variability from renewable integration, electrification, large Artificial Intelligence (AI)-driven loads, and extreme weather, MISO developed a dynamic reserve framework supported by Machine Learning (ML) models. The paper focuses on three key innovations including net uncertainty quantification across different operation timeframes, ML-based net uncertainty forecasting and dynamically setting reserve requirements in MISO’s energy and ancillary service co-optimized markets.

Central to this approach are AI/ML models that generate time-series forecasts of Net

Uncertainty, aggregating uncertainties from load and renewable forecast errors, generator outages, interchange variability, and transmission constraints. These forecasts are used to align reserve requirements with anticipated system risk in day-ahead and real-time markets.

Production experiences of Short-Term Reserve have demonstrated improved situational awareness, more efficient resource commitment, and enhanced reliability. Building on this success, MISO is extending the framework to Regulation Reserve and Ramp Capability

Products through tailored uncertainty models and real-time pattern recognition tools, all supported by a cloud-based Uncertainty Platform. These efforts highlight the value of AI enabled, adaptive uncertainty management in maintaining reliable and efficient electricity markets amid increasing system complexity.

Additional informations

Publication type Session Materials
Reference C2_10855_2026
Publication year
Publisher CIGRE
Country United States of America
Study committees
File size 372 KB
Price for non member 30 €
Price for member 30 €

Authors

GHESMATI Arezou - MISO-Energy, United States of America; TSAI Chen-Hao - MISO-Energy, United States of America; WANG Congcong - MISO-Energy, United States of America

Keywords

Uncertainty Management - Reliability and Efficiency in Grid Operation - Net Uncertainty Forecasts - Dynamic Reserve Requirements - Time-Series Forecasting - Classification

Dynamic Reserve: Data-Driven and Market-Based Uncertainty Management for Grid Operations