grid-stability

Application of machine/deep learning models & algorithms in the energy sector

Source code for the paper "Uncertainty-Aware Stability Analysis of IBR-dominated Power System with Neural Networks"

Matlab code for running the NRPS, RAPS, FEPS and nonlinear Kuramoto models: Models for electric grids that capture the dynamics of synchronous machines and VSMs.

Comparative Analysis of Machine Learning Models for Predicting Smart Grid Stability

This project analyzes renewable energy production trends in France, specifically focusing on Wind and Solar power over the seasons and years.

Bank marketing prediction with Logistic Regression (SMOTE, L1/L2) and Electrical Grid Stability classification with SVM (Linear, RBF, Polynomial). Achieved 97.3% accuracy with tuned RBF kernel.

Source code, dataset generation, and trained models for the MASc thesis "Physics-Informed Graph Neural Networks for Power-Grid Stability" — a closed-loop framework combining spectral-sensitivity virtual inertia allocation, a physics-informed ST-GNN predictor, and a PPO-based RL controller, validated on the IEEE 118-bus system.

Behavioral protocol for engineered inertia at inverter-based grid boundaries. Defines state machine, droop control, and asymmetric layer separation for grid-interactive deployments.

Simulating success rates for a campaign encouraging consumers to shift their electric vehicle charging from peak to off-peak hours to help alleviate grid strain.