Summer Research Project Grant Report
Cross-disciplinary Domains: Artificial Intelligence, Deep Learning, Power Systems, Electrical Engineering
Research Project Objective
Accurate electricity demand forecasting is vital for ensuring reliable, cost-efficient, and sustainable power system operation. Yet the growing complexity of modern grids which driven by renewable energy integration, electric vehicle adoption, and shifting consumption patterns, has outpaced the capabilities of traditional forecasting methods. This project will develop advanced Artificial Intelligence models, especially Transformer-based time series networks, to deliver accurate load forecasts.
Leveraging existing data such as Australian National Electricity Market (AEMO), the developed model will capture nonlinear, dynamic trends that traditional tools cannot. Successful implementation will provide power utilities with a robust, data-driven forecasting tool to support smarter, more resilient, and sustainable energy systems.
Background
Accurate short-term load forecasting (STLF) is a cornerstone of modern power system operation, underpinning critical decisions in generation scheduling, reserve allocation, energy trading, and grid reliability management. In deregulated electricity markets such as the AEMO, forecast errors directly become such economic costs:
Over-forecasting leads to inefficient resource commitment and excess generation, while under-forecasting risks supply shortfalls and price spikes.
