Infrastructure Systems

Machine Learning Techniques for Energy Forecasting and Optimization

Led by Zhen Ni, Ph.D.

PROJECT

This project aims to perform data visualization, computer programming, result analysis of artificial intelligence and machine learning for energy forecasting and management. Smart energy data includes electricity demand, load profile and so on. Students will obtain large-timescale residential load profiles from different locations and seasons (e.g., summer and winter, weekday and weekend), and use machine learning approaches (e.g., K-mean and self-organizing map) to analyze the demand patterns and predict energy consumption. Students will also use optimization approaches (e.g., genetic algorithm and evolution algorithm) for demand reduction, load allocation and analysis of rebound effect. Most of these approaches have built-in functions in python and Matlab environment. Students will obtain enhanced skills of scientific communication, multi-task management and oral presentation.

Dr. Ni will help students to build connections with relevant industry partners and government research laboratories. This project aligns with Dr. Ni’s NSF CyberTraining project to study the microgrid and smart community energy management. This project provides a pilot study for machine learning methods on load profiles clustering. The results provide the feasibility analysis of certain machine learning methods for energy management systems on advanced microgrid cyberinfrastructure.