Electrical Grid Reliability - AMI Momentary Outage Analysis
Team Members: Brayden Bell, Brandon Barboza, Ben Bradley
Faculty Advisor: Dr. Hong
Background
Prototype and Results
Sponsor/Client: Jackson EMC
Methodology and System Design
Conclusions and Futrure Work
Problem Statement and Objectives
Jackson EMC is an electric cooperative serving over 240,000 meters across northeast Georgia. As the grid modernizes, Advanced Metering Infrastructure (AMI) smart meters now capture momentary outage data in the form of "click counts," which are brief interruptions too short to trigger a traditional outage ticket but significant enough to degrade service quality for customers.
Jackson EMC tracks reliability using three separate systems: AMI smart meter exports, an Outage Management System (OMS), and GIS circuit maps generated in Milsoft WindMil software. Historically, these systems have not been analyzed together, leaving valuable grid reliability insights untapped.
This project bridges that gap by delivering a unified Python/Flask web application that integrates all three data sources into a single interactive interface for Jackson EMC engineers.
Data Integration Pipeline: AMI click-count records are ingested from Excel exports via pandas. OMS outage records are correlated against AMI events by time and circuit location. GIS circuit line data is exported from Milsoft WindMil in MPT GIS format.
Coordinate Projection: A custom Lambert Conformal Conic reprojection converts WindMil GIS coordinates from EPSG:2240 (Georgia West State Plane) to WGS84 for web mapping compatibility. Douglas-Peucker polyline simplification is applied to reduce vertex count and optimize rendering performance.
Backend: Python / Flask A Flask web server manages data ingestion, in-memory DataFrame filtering via pandas, and serves processed data to the frontend. Threshold parameters are passed as URL query arguments to filter events dynamically without page reloads.
Frontend: Leaflet.js + Plotly An interactive Leaflet.js map renders circuit line overlays, recloser markers, substation rings, and a click-count heat map overlay using the leaflet.heat plugin. ESRI World Imagery tiles provide satellite basemap context. Plotly Express charts display outage trends over time alongside the map.
Tools & Technologies: Python, Flask, pandas, openpyxl, Plotly Express, Folium, Leaflet.js 1.9.4, Milsoft WindMil MPT GIS, ESRI World Imagery
The prototype provides a single browser-based interface for Jackson EMC engineers to explore momentary outage patterns spatially and over time. Key features include:
Interactive Heat Map: Click-count density visualized across Jackson EMC's service territory using the leaflet.heat plugin
Circuit Overlays: GIS feeder lines, recloser markers, and substation rings derived from WindMil GIS exports
Threshold Filtering: Adjustable controls isolate high-frequency event locations without reloading the page
Trend Charts: Plotly Express charts display outage frequency over time alongside the map view
Jackson EMC lacked a unified tool to correlate AMI click-count data with OMS outage records and GIS circuit geography. Without this integration, engineers could not easily identify which feeders or geographic zones experienced the most frequent momentary interruptions, making proactive maintenance difficult.
Project Objectives
This project demonstrates that integrating AMI, OMS, and GIS data into a unified web application is technically feasible and operationally valuable for an electric cooperative. The prototype gives Jackson EMC engineers a foundation for data-driven grid reliability decisions that was previously unavailable.
Future Work
Special thanks to Jeremy Tilson and Jackson EMC for their support and guidance through this project.