Development of an Adaptive Fuzzy Logic System for Energy Management in Residential Buildings

Heating, Ventilation, and Air Conditioning (HVAC) systems are the main target for energy and load management in residential buildings due to their high energy consumption. The role of Thermostats is to automatically control the HVAC systems while users accommodate their everyday schedules and preferences. The initiatives such as demand response (DR) programs, Time of Use (TOU) rates, and real-time pricing (RTP) are often applied by smart grids to encourage customers in order to reduce consumption during peak demand periods. However, it is often a hassle for residential users to manually reprogram their thermostats in response to smart grid initiatives and/or environmental conditions that vary over time. In this thesis, the research endeavors are dedicated to bring forward a novel autonomous and adaptable system for control of residential HVAC systems which results in an “Adaptive Smart Thermostat”. To do so, a “House Simulator” is developed in MATLAB-GUI with thoughtful consideration as a tool to facilitate the study of energy management for residential HVAC systems in smart grids. The simulator also assists in the implementation and verification of our proposed techniques under different scenarios such as RTP, various user schedules, and different environmental conditions. Furthermore, a new algorithm using rule-based fuzzy logic and wireless sensors capabilities for residential demand-side management is developed. The algorithm is augmented into existing programmable communicating thermostats (PCT) in order to enhance the learning capability of PCTs during participation in DR programs. The conducted results show that the PCT equipped with our approach, versus existing PCTs, performs better with respect to energy and cost saving, while maintaining user thermal comfort. The achieved results led us to develop a novel “Autonomous Smart Thermostat” that is the result of a synergy of supervised fuzzy logic learning, wireless sensor capabilities, and smart grid incentives. The results demonstrate that the developed thermostat autonomously adjusts the set point temperatures in ASHRAE thermal comfort-zone, while not ignoring the energy conservation aspects. However, in cases that the user overrides the decision(s) made by autonomous system, a novel “Adaptive Fuzzy Learning Model” utilizing wireless sensor capabilities is developed in order to learn and adapt to user new preference and schedule changes based on rulebased fuzzy logic learning. The results show that the developed system is adaptable, smart, and capable of intelligent zoning control while it improves energy management in residential buildings without jeopardizing thermal comfort.