Data-driven multi-layered intelligent energy management system for domestic decentralized power distribution systems

Abstract

Energy management techniques have been immensely adopted in residential, industrial, and commercial sectors to meritoriously flatten the load profile of consumers. In this article, a cutting-edge intelligent energy management system is proposed to analytically monitor and regulate the load demand of a residential area to meritoriously address the load balancing and scheduling optimization problem with an aim to design an autonomous domestic complex, less reliant on the focal power grid. To do so, single- and multi- objective functions are defined for the presented double-layered framework. In the first layer, a day-ahead scheduling problem is evaluated to achieve the mutual revenue of electricity utilities and consumers, while in the second layer, a real-time coordination scheme for robust energy transactions among appliances, multiple microgrids, and the main power system is examined. Herein, a 0−1 multiple knapsack problem incorporated with the mixed-integer linear programming method is used to formulate the system mathematical model. To evaluate the intended system model, three heuristic algorithms (genetic algorithm, wind-driven optimization, and their hybrid) are modified. Furthermore, the Monte–Carlo simulations are performed to mitigate the impact of randomness and avoid pre-convergence. From the results of extensive simulations, it is corroborated that the proposed methodology is robust, reliable, and eco-social compared to prior schemes. This article reveals that the hybrid approach outperformed the other two strategies in terms of energy consumption, electricity cost, peak-to-average ratio, and user comfort level.