Power fluctuation mitigation strategy for microgrids based on an LSTM-based power forecasting method

Luo Zhao, Xinan Zhang, Xiuyan Peng

Research output: Contribution to journalArticlepeer-review

Abstract

With the rapid penetration of renewable generation systems and active loads, the stability and reliability of modern power systems face several challenges owing to power fluctuations caused by renewable intermittency and load uncertainty. Power fluctuations are more significant in islanded microgrids that possess low inertia. Therefore, this study proposes a novel cost-effective proactive control strategy to mitigate power fluctuations of an islanded microgrid. The proposed strategy produces an early acting control reference for generators based on an improved ultra-short-term power fluctuation forecasting algorithm to significantly increase the fluctuation compensation capacity of the generators. Moreover, the size, workload, and cost of the energy storage system reduce. A combined LSTM neural network structure is employed to achieve accurate power fluctuation forecasting. The effectiveness of the proposed method is verified on an islanded hybrid AC/DC microgrid simulation platform.
Original languageEnglish
Article number109370
JournalApplied Soft Computing
Volume127
DOIs
Publication statusPublished - Sep 2022

Fingerprint

Dive into the research topics of 'Power fluctuation mitigation strategy for microgrids based on an LSTM-based power forecasting method'. Together they form a unique fingerprint.

Cite this