27 October 2023 in IEEE Transactions on Power Systems
In this research work, three practical correction methods are proposed to mitigate the impact of defective input features in power system data measurement for machine learning (ML) applications. A well-trained ML tool may become ineffective due to defective input features, which may originate from measurement issues, such as monitor malfunction, cyberattack, communication failure, or others. It is crucial to correct defective input features to enable ML tools with desirable performances. This letter first introduces the mechanism of three correction methods, i.e., statistical-value-based method, minimal-error-based method, and DNN-based adaptive method. Then, the methods are validated via a deep neural network (DNN) case for power system stability enhancement. Validation results demonstrate that the adaptive method achieves the best performance, enabling the well-trained ML tool with a similar accuracy level to the case of no data defects. Although actual measurements may have various data issues challenging ML applications in power systems, the proposed methods show promise for addressing these challenges.
1 May 2020 in Electric Power Systems Research
13 December 2020 in IET Generation, Transmission & Distribution
AZAlejandro Zamora‐Mendez
This paper proposes a new approach to identify a data‐based power system linear model by means of frequency interpolations, aiming to obtain a suitable system representation for selective‐modal analysis purposes. The key idea behind the identification process is the Loewner‐based frequency interpolation carried out by the Loewner matrices. The proposed approach demonstrates that the Loewner‐based frequency interpolation is able to fit a linear model that can be used for small‐signal analysis studies, since it provides the state‐space representation, the frequency response, and modal information (frequency, damping, and mode‐shape). Then, a selective modal analysis is accomplished over two test cases (Kundur and New England–New York power grids) by employing the identified linear model provided by the Loewner‐based frequency interpolation method. The attained results confirm the outstanding performance of the proposal which is validated against the small‐signal analysis and compared with the eigensystem realisation algorithm, overcoming the absolute error of the model identified with the traditional eigensystem realisation algorithm approach by at least 37 times, and properly capturing the modal information in a frequency band of concern.