Data-driven smart home energy management based on proximal policy optimization

• Main Author: Ma, Aoxiang
• Other Authors: Xu Yan
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Electrical and electronic engineering::Electric power::Production, transmission and distribution
• Online Access: https://hdl.handle.net/10356/166256
• Institution: Nanyang Technological University

As the economy and population continue to grow, the demand for energy in modern society is increasing at a rapid pace. Unfortunately, limited non-renewable resources and the slow scaling up of renewable energy technologies have created a shortage of supply. As a result, saving energy and achieving efficient use of energy has become an essential task. Residential buildings are a significant contributor to total energy consumption, making them a crucial area to focus on in achieving full control of energy consumption. 1)Model energy management algorithms of various devices with different functions, such as PV and EV, in the intelligent home energy management system. 2)A variety of hyperparameter combinations are used in modeling to establish multiple models, and the optimal hyperparameter combinations are selected by comparing the characteristic values and their specific performances in the system. 3)Modeling of smart home optimization problems and development of home energy management algorithm based on deep reinforcement learning. Minimize the cost of electricity while considering consumer comfort.