aalto1 untyped-item.component.html
An Innovative Diffusion Model Based on Adaptive Time-Frequency Drive for High-Resolution Photovoltaic Data Reconstruction
Loading...
Access rights
openAccess
acceptedVersion
URL
Journal Title
Journal ISSN
Volume Title
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
This publication is imported from Aalto University research portal.
View publication in the Research portal (opens in new window)
View/Open full text file from the Research portal (opens in new window)
View publication in the Research portal (opens in new window)
View/Open full text file from the Research portal (opens in new window)
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Date
Major/Subject
Mcode
Degree programme
Language
en
Pages
12
Series
IEEE Transactions on Smart Grid
Abstract
Distributed photovoltaic (PV) generation is widely deployed on the user side, and acquiring and storing high–time‑resolution data is costly, whereas low‑resolution PV data cannot satisfy fine‑grained analysis requirements. Existing high‑resolution reconstruction methods often rely on preset frequency bands or global stationarity assumptions, which prevent them from responding in real time to nonstationary short‑term disturbances in PV power series. This paper proposes a Fourier transform–driven diffusion model for high‑resolution data reconstruction. Compared with convolutional neural networks, Transformer models, generative adversarial networks, and baseline diffusion models, the proposed approach attains smaller mean squared error (MSE), frequency component error (FCE), and peak–valley difference (PVD). Experiments show that on the DKASC dataset, the model achieves an MSE of 0.271, an FCE of 5.878, and a PVD of 0.345, which are improvements of approximately 6.2%, 7.1%, and 5.5% over the second‑best method. On the Photovoltaic output dataset (PVOD), the corresponding values are 1.298, 2.699, and 0.110, with reductions of about 1.7%, 0.7%, and 0.9%. These results indicate that explicitly leveraging frequency‑domain information reduces spectral errors and maximally preserves peak–valley details, yielding the best reconstruction performance.
Description
Publisher Copyright: © 2010-2012 IEEE.
Other note
Citation
Fu, X, Ma, Q, Li, Z, Li, J, Feng, R, Sun, H & Zhang, Y 2026, 'An Innovative Diffusion Model Based on Adaptive Time-Frequency Drive for High-Resolution Photovoltaic Data Reconstruction', IEEE Transactions on Smart Grid. https://doi.org/10.1109/TSG.2026.3655584