December 2025

Conference Paper

Advancing Dynamic Modeling of Grid-Connected PV Inverter Using Bi-LSTM-Based AI Model

By:
Subedi, Sunil ; Gui, Yonghao ; Xue, Yaosuo
Page Number:
1-5
Book Title:
2025 IEEE Energy Conversion Conference Congress and Exposition (ECCE)
Publication Date:
December 2025
Publisher Location:
IEEE, Pennsylvania, United States of America
Conference Name:
2025 IEEE Energy Conversion Conference Congress and Exposition (ECCE)
Conference Location:
Philadelphia, Pennsylvania, United States of America
Conference Sponsor:
IEEE
View DOI Listing:
https://doi.org/10.1109/ECCE58356.2025.11260177

Abstract

Power electronic converters (PECs) are widely used in modern power systems to facilitate the interconnection between various AC or DC sources and loads. Because of the extensive integration, the power system has grown into a more dynamic system in which the dynamics of the PECs must be adequately modeled. The paper presents a new bidirectional long short-term memory (Bi-LSTM) method for evaluating grid-connected inverter-based resources (IBR) dynamics. The method is tested using real hardware data from a grid -connected commercial inverter in laboratory experiments. Results show the Bi-LSTM model accurately reproduces the detailed IBR model's dynamics, even when the internal structure is unknown and parameters are unknown, preventing the disclosure of the manufacturer's confidential data.