Smart grid is an electricity network where actions of all users are intelligently integrated in order to efficiently deliver sustainable, economical and secure electricity supplies. The employment of sensing, embedded processing and digital communication must make the modern electricity network observable, controllable, automated and fully integrated. Thus, in the smart grid scope there is the self-healing concept that primary aims the maximum restoration of loads affected by a fault, obeying to operational limits of the electrical network, minimizing the number of switching operations within the shortest time interval and without the human intervention. Current distribution networks have not the capability to perform the self-healing because their protection devices have not the necessary intelligence to operate in a suitable way. Therefore, in this research project are being proposed the development and analysis of a multi-agent system (MAS) to enable the self-healing on the smart grid. The proposed system comprises the technique of distributed artificial intelligence (DAI) where every protection device is an intelligent agent able to adapt its behavior according to environment conditions because it has some embedded artificial intelligence, e.g. a fuzzy controller. The analysis of the MAS must be performed through the assessment of reliability indexes of distribution networks, such as the system average interruption frequency index (SAIFI) and system average interruption duration index (SAIDI), that must be collected from agent-based modeling and simulation (ABMS) where the stochasticity of smart grid must be simulated using the Monte Carlo method and the agent interaction rules.
News published in Agência FAPESP Newsletter about the scholarship: