Digital Twins for Resilient Grids: AI-Driven State Estimation and Sensor Placement Optimization

Context

Are you passionate about shaping the future of energy systems in developing countries? Are you ready to dive into the exciting world of artificial intelligence (AI) and anomaly detection? Look no further!

Project Overview
In the heart of the energy revolution, developing countries face a critical challenge: ensuring a reliable power supply for their businesses and citizens. But here’s the twist—the current energy system resembles a tangled web, struggling to reach its full potential. Imagine a chaotic dance of electrons, unlike anything you’ve seen in Europe’s well-organized grids.

Your Mission
Are you passionate about revolutionizing power distribution in emerging economies? Ready to tame the chaos of Ethiopia’s network by turning it into a sleek virtual replica? As a thesis candidate, your mission is crystal clear: develop a high-fidelity three-phase digital twin and build an organized data space for these complex energy systems, leveraging cutting-edge power-system modeling and AI-driven analytics!

Tasks

• Develop and implement a high-fidelity three-phase digital twin of the Harar distribution feeder.
• Extend the existing single-phase simulator to support unbalanced, three-phase power-flow modeling.
• Generate synthetic data sets covering a range of load profiles and fault scenarios for state estimation.
• Integrate and evaluate a weighted least-squares state estimation algorithm to detect and localize anomalies in the three-phase network.
• Accelerate and refine the genetic algorithm for sensor-placement optimization, targeting ≥ 50 % runtime reduction without degrading estimation accuracy.

Prerequisites

• Motivation and interest in solving technical problems
• Interest in electronics and Data Science
• Experience in programming (Python)
• Basic knowledge of ML/AI