Intelligent Energy and Operation Management of AC, DC and Hybrid Microgrids Based on Evolutionary Techniques

摘要

Intelligent energy management systems (EMSs) play pivotal roles for microgrids (MGs). The integration of distributed generators (DGs), energy storage devices (ESDs), electrical vehicles (EVs), and flexible loads in a large-scale system of interconnected MGs needs a local management and control platforms to avoid probable integration issues. Most of the small scale aforementioned components are operated in low voltage (LV) power networks which require optimal strategies to achieve sufficient performance to operate interconnected to other MGs and upstream networks. Thus, an active energy management strategy is required for off and on grid modes of terrestrial and shipboard MGs to satisfy all demands and minimize unwanted outcomes.;Due to the high penetration of renewable energy sources (RESs) and their output fluctuations in MGs, stochastic analysis besides deterministic should be considered to reduce the uncertainty effects of RESs based on their probabilistic nature. Robust EMSs are needed to diminish prediction errors and improve the reliability of power supplies in hybrid and interconnected MGs. Nevertheless, qualified approach to fulfill EMS for LV MGs in the large-scale will be challenging in both grid operation modes. Optimization modules into tertiary management layer have considered as a potential solution in order to actualize control strategy of the terrestrial or ship MGs with different types of practical constraints. However, each of these methods not only yield benefits but also bring new challenges related to their shortage.;Different approaches have been considered to fulfill the EMS requirements of MGs. A large amount of literature focuses on the management strategy of MG in an off-line manner rather than multiple MGs which interact with each other and an upstream network in an on-line manner. In addition, the most commonly used optimization modules in EMSs do not meet the computational burden and convergence capability trade-off required for real-time applications. This report proposes a heuristic optimization approach for distributed control and management of hybrid MGs for real-time requirements. The Crow Search Algorithm (CSA) offers a superior method to move traditional non-linear optimization approaches since its fewer control parameters permit a rapid response compared to other search approaches. Moreover, a distributed fashion CSA (DCSA) is implemented to fulfill linear and non-linear solver requirements of real-time EMSs for hybrid power distribution system.