Design of Energy Management System for Smart Home

Abstract

In India, residential buildings account for nearly one-third of energy consumption related to Greenhouse Gas (GHG) emissions. The space heating, cooling, lighting, and Indoor Air Quality (IAQ) control systems significantly consume electricity in maintaining occupant's comfort. The residential sector's energy consumption has been increased steadily and occupied approximately 30-40% of overall energy consumption. At present, India's urban population is about 410 million, and by the year 2050, it is estimated to reach around 814 million. In the last few decades, India has improved its economy rapidly and is also cautious for high growth in the future. The research on energy consumption forecasting and management has highlighted the significance of residential building energy consumption forecast for enhanced decision-making regarding energy conservation plan. Nowadays, smart homes are introduced to save energy and provide comfort to the occupants. Smart homes are automated buildings that incorporate advanced automation systems to give the occupants advanced monitoring and control of their functions. The early design stage of a smart home requires an energy consumption prediction model to predict future energy consumption. On large-scale short-term, medium-term, and long-term energy consumption prediction models have been introduced so far, but one of the most significant drawbacks of the models that they have not include the daylight factor and relative humidity. Despite the importance of the efforts put forward, there is still some unexplored area in the residential level's long-term and precise energy consumption prediction model. Also, the energy consumption and comfort management are vital in designing the Energy Management System (EMS). Such type of researches would help to minimize energy consumption using the EMS. This research aims to design a Smart Home Energy Management System (SHEMS) that manages the Heating, Ventilation, air conditioning (HVAC), and lighting system's energy consumption while meeting occupants' indoor comfort requirements. To achieve this goal, we need to analyze the future energy consumption in a smart home. Therefore, a long term precise, data-driven based energy consumption prediction model for a smart home is proposed. Firstly, a 2BHK single-story multi-zone residential building is modeled in TRNSYS16 building simulation software. Energy consumption for maintaining indoor comfort depends on environmental parameters such as temperature and relative humidity. Hence, temperature and relative humidity are predicted by the year 2050 with the machine learning approach. The predicted temperature and relative humidity data have been fed as input to the TRNSYS16 for smart home energy consumption calculation. Considering the building structure's complicated design, user comfort parameters, including heating, cooling, ventilation, IAQ, and illumination level, the energy consumption prediction model has been developed up to the year 2050. The inclusion of the daylight factor predicts more accurately as compared to the conventional prediction model. Thus, the most accurate and precise model, with 95% coefficients bound, has been developed for energy consumption prediction. Further, this model can be integrated with various controlling techniques in energy conservation planning and management. Secondly, the EMS design for a smart home also requires optimizing environmental parameters to provide comfort as per occupant's preferences. The energy consumption and occupant's comfort level often conflict with each other in indoor environmental conditions. To resolve the conflict, a multiobjective problem has been formulated, and solution methodology has been proposed. The proposed solution methodology considers occupants' adaptations while making decisions on fixed temperature, illumination level, and CO2 value. The Crisscross Search Particle Swarm Optimization (CSPSO) with multiagent topology is incorporated to optimize the environmental parameters temperature, illumination level, CO2 concentration with Fuzzy Logic Controllers (FLCs). The CSPSO is applied to the environmental parameters for an optimal solution corresponding to set temperature, illumination level, and CO2 concentration. Furthermore, Artificial Intelligence (AI) has been used to optimize environmental parameters like temperature, illumination, and CO2 to maximize comfort and minimize energy consumption. To do so, a multivariable objective function has been formulated with set temperature, illumination and CO2 constraints. A trust-region reflective algorithm based on AI has been used to solve it. Trust region reflective algorithm successfully optimized the environmental parameters within the set limits. The energy consumption is minimized while maintaining occupants indoor comfort. Next, the performance evaluation has been done with the Genetic Algorithm (GA), Bat, NNA, PSO, and ABC optimization techniques are compared for thermal comfort. Bat optimization technique was superior to other utilized optimization techniques (GA, PSO, NNA, ABC). Optimization algorithms such as GA, Bat, Neural Network Algorithm (NNA), Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC), and machine learning approach with FLCs are utilized to reduce the energy consumption and maximize the thermal comfort, visual comfort, and IAQ comfort.