Synergistic Integration of Demand Side Management, Renewable Energy Sources, Battery, and Hydrogen Storage in Hybrid Energy Systems

Abstract

In recent years, there has been a significant increase in the demand for hybrid energy systems (HES). This surge is attributed to a combination of factors, including the pursuit of sustainable and resilient future energy solutions. HES integrates various energy resources to achieve synchronized energy output. However, HES faces notable challenges due to escalating energy consumption, the expenses associated with utilizing multiple sources, and increased emissions from non-renewable energy resources. On the other hand, when utilizing renewable energy sources (RES), the management of distributed energy resources (DER) plays a crucial role in optimizing the practical objectives of the grid. This thesis employs optimization techniques, such as the shuffled frog leaping algorithm (SFLA), to manage DER and implement demand response programs (DSP). The aim is to optimize the economic, technical, and environmental aspects of a smart micro-grid (SMG). Furthermore, this thesis adopts a hybrid approach that combines well-established techniques, namely the Non-Dominated Sorting Genetic Algorithm II and Multi-Objective Particle Swarm Optimization (Hybrid-NSGA-II-MOPSO), aims to optimize operational costs, reduce pollution, and address the challenge of achieving a high penetration of RES while minimizing the energy gap between initial demand and consumption. For prediction of uncertain behavior of RES before integration with the grid, cumulative distribution function (CDF) and probability distribution function (PDF) are used. The DER included consists of wind, solar, micro-turbine, diesel generator, and utility grid. The demand side management (DSM) strategy is designed for three types of loads, sheddable loads, non-sheddable loads, and shiftable loads. To establish a bi-directional communication link between the grid and consumers, distribution grid operator (DGO) is employed. For validation, this model is is compared with different individual optimization techniques like SFLA, MOPSO and NSGA-II as well as different constraints are considered. The results obtained shows the superiority of proposed SFLA and Hybrid-NSGA-II-MOPSO algorithms in terms of avoiding pre-mature convergence which is a common challenge in optimization, and achieving global optimum for the proposed objectives.