Power-System Stability Classification Using Deep Neural Networks

Machine-learning classification of power-system stability using deep neural networks, Python, scikit-learn, Keras, and MATLAB/Simulink.

Overview

This research applied deep neural networks (DNNs) to power-system stability classification and investigated the influence of model architecture and dataset size on classification performance.

Research Activities

  • Prepared and analyzed power-system stability datasets
  • Developed DNN-based classification models
  • Investigated the effects of dataset size and hidden-layer configuration
  • Evaluated classification performance using machine-learning workflows
  • Used MATLAB/Simulink for supporting power-system analysis

Tools

Python · Keras · scikit-learn · MATLAB · Simulink · Neural networks

The work led to the conference paper presentation “Effect of Dataset Size and Hidden Layers on the Stability Classification of IEEE-14 Bus System Using Deep Neural Network,” which received Second Place in the Best Paper Award at ICEPE 2022.