Advanced Machine Learning-Based High Demand Forecasting of Household Energy Consumption for Enhancing Grid Operations

This thesis explores forecasting techniques for household energy consumption to help Distribution System Operators (DSOs) manage high demand loads and ensure grid stability. It focuses on predicting when demand exceeds critical thresholds and for how long, enabling proactive energy management. The study analyzes how different data aggregation levels affect forecast accuracy and investigates methods to restore altered load signals for better predictions. By comparing forecasting models and evaluating their performance, the research aims to improve energy management, support automation in grid operations, and enhance data-driven decision-making for a more stable and efficient power distribution system.