سامانه همایش های دانشگاه آزاد اسلامی واحد نجف آباد

صفحه اصلی / ششمین کنفرانس بین المللی بهینه سازی مصرف انرژی الکتریکی

Enhanced Frequency Regulation in Islanded Microgrids Using a Machine Learning-Assisted Linear-Quadratic Regulator

نویسندگان :

Soheil Ebrahimian¹ Hossein Shahinzadeh² Hamed Nafisi³ Mahtab Bagheri⁴ Majid Moazzami⁵ Zohreh Azani⁶

1- دانشگاه آزاد اسلامی واحد نجف آباد 2- دانشگاه صنعتی امیرکبیر (پلی تکنیک تهران) 3- دانشگاه صنعتی امیرکبیر (پلی تکنیک تهران) 4- دانشگاه آزاد اسلامی واحد نجف آباد 5- دانشگاه آزاد اسلامی واحد نجف آباد 6- دانشگاه صنعتی امیرکبیر (پلی تکنیک تهران)

کلمات کلیدی :

Islanded Microgrids،Frequency Regulation،Machine Learning،Linear-Quadratic Regulator،Renewable Energy،Uncertainty،Energy Storage

چکیده :

The stability of frequency in islanded microgrids is highly sensitive to fluctuations in generation resources and load demand. Ensuring effective frequency regulation in such autonomous systems is critical, particularly in the presence of renewable energy sources and uncertain operating conditions. This paper presents an enhanced frequency control strategy that integrates a Machine Learning-Assisted Linear-Quadratic Regulator (ML-LQR) to optimize the dispatch of generation units while mitigating frequency deviations. The proposed approach leverages a machine learning model to dynamically adjust the penalty coefficients of the LQR controller, improving its adaptability to varying system conditions. By incorporating predictive insights from historical data, the method enhances frequency regulation even under uncertain scenarios. The effectiveness of the proposed ML-LQR strategy is validated through simulations on an islanded microgrid under normal and disturbance conditions. The results demonstrate that integrating machine learning significantly improves frequency stability, reduces control stress, and optimally compensates for variations in generation and load conditions. Additionally, a comprehensive analysis of penalty coefficient variations highlights the adaptive capabilities of the proposed approach.