Eficiência energética residencial por meio de inteligência artificial

Abstract

In view of the unbalanced consumption of electric energy, caused due to the technological and demographic advances that Brazil has been experiencing, there is a need to apply techniques aimed at saving this consumption. In Brazil alone in 2019, approximately 601.68 TWh of total energies were wasted. Based on such a survey of exacerbated consumption and waste, it is proposed in this conclusion work, the development of intelligent automation that, through control methods, provides energy savings in homes. For this automation, the artificial intelligence technique known as the Multiagent System is used, which establishes different tasks for agents to work in a similar way to a society, in favor of a single objective. This objective concerns the preservation of the user's comfort, through their preferences and action patterns, at the same time that energy efficiency techniques are applied to optimize consumption. Finally, a concept known as Machine Learning is also applied, which aims to bring to the intelligence, the learning capacity, which starts to study the history of actions that the individual performs in order to create a pattern of behavior to automate recurrent tasks exercised by him. The prototype developed and controlled by intelligence is based on microcontrollers, sensors and electronic components that together allow the virtualization of information from the environment. Then they become available to the control center and also have the ability to function as actuators in the middle, allowing the control of different parameters of the residence.