کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

Data-Driven Simultaneous Estimation of Power System Inertia and Damping under Measurement Noise

نویسندگان :

Saeed Aslani amin abad¹ Turaj Amraee² Moein Abedini³

1- دانشگاه تهران 2- دانشگاه خواجه نصیر الدین طوسی 3- دانشگاه تهران

کلمات کلیدی :

Inertia constant،damping coefficient،frequency stability،renewable energy،machine learning.

چکیده :

High penetration of renewable energy resources reduces effective inertia and alters damping, leading to faster and more challenging frequency dynamics in power systems. This paper presents a data-driven framework for the simultaneous estimation of the inertia constant and damping coefficient using system frequency response data. Datasets are generated through dynamic simulations under two information scenarios, and multiple machine learning models including Random Forest, Support Vector Regression, Gradient Boosting, and Convolutional Neural Networks are evaluated using raw frequency trajectories, extracted features, and hybrid inputs. Model performance is assessed using MAE, RMSE, and the coefficient of determination, with particular attention to the effects of input representation and measurement noise. All models are trained on clean data, while noise is introduced only during testing to evaluate generalization under unseen disturbances. The results show that raw frequency trajectories provide a strong and reliable baseline, whereas feature-based and hybrid representations offer complementary benefits depending on noise conditions and model type. Tree-based models, particularly Random Forest and Gradient Boosting, exhibit more stable performance than SVR and CNN under noisy and complex conditions.