Abstract
Around the globe, innovation with integrating information and communication technologies (ICT) with physical infrastructure is a top priority for governments in pursuing smart, green living to improve energy efficiency, protect the environment, improve the quality of life, and bolster economy competitiveness. Cities today faces multifarious challenges, among which energy efficiency of homes and residential dwellings is a key requirement. Achieving it successfully with the help of intelligent sensors and contextual systems would help build smart cities of the future. In a Smart home environment Home Energy Management plays a critical role in finding a suitable and reliable solution to curtail the peak demand and achieve energy conservation. In this paper, a new method named as Home Energy Management as a Service (HEMaaS) is proposed which is based on neural network based Q-learning algorithm. Although several attempts have been made in the past to address similar problems, the models developed do not cater to maximize the user convenience and robustness of the system. In this paper, authors have proposed an advanced Neural Fitted Q-learning method which is self-learning and adaptive. The proposed method provides an agile, flexible and energy efficient decision making system for home energy management. A typical Canadian residential dwelling model has been used in this paper to test the proposed method. Based on analysis, it was found that the proposed method offers a fast and viable solution to reduce the demand and conserve energy during peak period. It also helps reducing the carbon footprint of residential dwellings. Once adopted, city blocks with significant residential dwellings can significantly reduce the total energy consumption by reducing or shifting their energy demand during peak period. This would definitely help local power distribution companies to optimize their resources and keep the tariff low due to curtailment of peak demand.
Highlights
Smart cities in brief can be defined as a city which uses information and communication technologies (ICT) such as smart sensors, cognitive learning, and context awareness to make lives more comfortable, efficient, and sustainable [1]
We model the power states as a Markov Decision Process (MDP) and derive its solution using reinforcement learning (RL) based Neural Fitted Q-Iteration (NFQI) algorithm
Energy management in smart cities is an indispensable challenge to address due to rapid urbanization
Summary
Smart cities in brief can be defined as a city which uses information and communication technologies (ICT) such as smart sensors, cognitive learning, and context awareness to make lives more comfortable, efficient, and sustainable [1]. Peak load reduction is mostly valuable for utilities and most popular only in a purely market-driven energy management environment Under these circumstances, Demand Response (DR) [11,12] offers an opportunity for consumers to play a significant role in the operation of the electric grid by reducing or shifting their electricity consumption during peak periods in response to time-based rates or other forms of financial incentives. User interface: Using a node-red development framework [33] (Node-RED is a web-based programming tool for wiring together hardware devices, APIs and online services.) and message queue telemetry protocol secure broker, a user interface has been designed It incorporates intelligent energy management capability and provides user input options.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
