PowerChain to PowerDAG: Reliable and Verifiable Agentic AI for Automating Distribution Grid Analysis
Badmus, Emmanuel
Badmus, Emmanuel
Citations
Altmetric:
License
License
Abstract
This project studies how agentic AI can automate distribution-grid analysis from natural-language requests. We present a progression from PowerChain to PowerDAG, two systems that generate and execute multi-step workflows using grid-analysis tools and utility data. PowerChain shows that large language models can organize power-system tasks as executable workflows. PowerDAG improves reliability by adding adaptive retrieval of relevant expert examples and just-in-time supervision that detects invalid actions before execution. Together, these systems support tasks such as power flow, hosting capacity analysis, data retrieval, and infeasibility diagnosis. The main result is that reliability improves when the agent receives the right examples at the right time and when unsafe or invalid tool calls are blocked before they affect the environment. This work outlines a path toward dependable AI assistants for real power engineering applications.
Description
Date
1/1/2026
Student Status
Graduate Student
Journal Title
Journal ISSN
Volume Title
Type of presentation
Poster
Collections
Research Projects
Organizational Units
Journal Issue
Citation
DOI
Advisor(s)
Department
Program/Major
Electrical Engineering
College/School
College of Engineering and Mathematical Sciences
Organization
Research Category
Engineering