برمك بيگ زاده نوعي

Considering the growing importance of energy in various sectors, its proper management and distribution hold a significant position. This study focuses on decentralized energy management for smart microgrids, including distributed generation (DG) units, batteries, and demand-responsive loads. Given the variable nature of load consumption and electricity price fluctuations, optimal energy management in microgrids faces multiple challenges. Moreover, centralized control structures in large-scale systems increase computational burden and the complexity of control algorithms. In this research, a multi-agent decentralized framework is developed, where generation units, batteries, and consumers are modeled as independent intelligent agents capable of learning and interacting with their environment. The proposed method, based on reinforcement learning and model-free control, determines optimal policies by selecting a comprehensive cost function that encompasses network objectives and constraints for energy management. To enhance reliability and reduce decision-making risk, a robust optimization approach is employed, modeling and controlling uncertainties related to load demand and energy exchange prices. Additionally, leveraging demand response programs allows cost reduction for consumers and increased profitability for producers. The primary objective of this study is to maximize producers’ profit, minimize consumers’ costs, and reduce the microgrid’s dependency on the main grid. Simulations and evaluations are performed using real load consumption data and electricity market prices from Iran. Quantitative and qualitative results under different scenarios indicate that the proposed approach achieves an optimal balance between costs and profits, leading to improved microgrid performance while satisfying energy management constraints.

One of the most important control layers in microgrids is the secondary control layer. This layer is responsible for compensating voltage and frequency deviations and sharing active and reactive power in the microgrid. Control strategies are commonly classified into three categories: centralized, decentralized, and distributed. The existence of a communication infrastructure for exchanging information in centralized and distributed control strategies can be a vulnerability for cyber attacks on the network, which is one of the drawbacks of these two methods. The lack of a communication infrastructure in the decentralized control structure, which compensates for voltage and frequency deviations and thereby eliminates one of the network's weaknesses against cyber attacks, has made this structure advantageous. In this research, an optimized decentralized controller based on fuzzy logic is proposed for frequency recovery and active power sharing in a microgrid. The proposed optimized fuzzy logic-based controller is independent of the need for a communication infrastructure for information exchange, which is the main goal of this research. The design of this controller is based on a quadratic-linear cost function, and an optimal solution for it is provided using the LQR (Linear Quadratic Regulator) method. The weighting coefficients Q and R in the proposed controller are determined based on fuzzy logic, which ensures the selection of the best weighting coefficients and, consequently, better performance of the controller. Only a low-pass filter dynamics is used in the control loop to design this controller. This process does not require time-dependent protocols and is event-based. By employing this method, the microgrid frequency quickly returns to its nominal value after disturbances, thus achieving the secondary control objectives (frequency recovery and active power sharing) effectively.