High-Resolution Long-Term Load Forecasting for Large-Scale AI Data Centers and Electric Vehicles Using Weather Normalization and Scenario Analysis: A Jeju Island Case Study Based on the ERCOT Method
A modular, hourly-resolution framework for long-term load forecasting on isolated grids, addressing the limitations of conventional monthly models in capturing weather-driven volatility and structural load growth and providing a quantitative basis for storage sizing, demand-response design, and transmission expansion.