Photovoltaic power generation prediction based on adaptive graph with time dimension transformation algorithm

Fengqi Li; Yajun Wang; Bo Jin

doi:10.1117/12.3011509

7 December 2023 Photovoltaic power generation prediction based on adaptive graph with time dimension transformation algorithm

Fengqi Li, Yajun Wang, Bo Jin

Author Affiliations +

Proceedings Volume 12941, International Conference on Algorithms, High Performance Computing, and Artificial Intelligence (AHPCAI 2023); 129414Y (2023) https://doi.org/10.1117/12.3011509
Event: Third International Conference on Algorithms, High Performance Computing, and Artificial Intelligence (AHPCAI 203), 2023, Yinchuan, China

Abstract

The solar power forecasting task plays a vital role in the field of photovoltaic (PV) power generation. To account for the complex correlations between PV power, meteorological data, and equipment data, a novel model called TimeRelationNet is proposed in this study. The TimeRelationNet model leverages a graph neural network to learn the relationship matrix among different features and construct a network of feature relationships. It combines the cross-attention mechanism with a two-dimensional modeling method for time series data. This enables the model to effectively capture the intricate temporal variations in PV power, transforming one-dimensional temporal sequences into two-dimensional tensors based on multiple periods. The model utilizes multi-scale convolution operations to extract PV data features from within and between different time periods more effectively and comprehensively. Additionally, the transformer is used to process the original data, and the transformed data is subjected to cross-attention mechanism, adaptive aggregation, and attention-based weighted fusion. The performance of the TimeRelationNet model is evaluated using data from a specific power station in Liaoning Province. Comparative analysis with three other forecasting models - Transformer, ESTformer, and LightTS - reveals that the TimeRelationNet model achieves the best predictive results, with a Mean Squared Error (MSE) of 0.089±0.002 and a mean absolute error (MAE) of 0.225±0.002.

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Fengqi Li, Yajun Wang, and Bo Jin "Photovoltaic power generation prediction based on adaptive graph with time dimension transformation algorithm", Proc. SPIE 12941, International Conference on Algorithms, High Performance Computing, and Artificial Intelligence (AHPCAI 2023), 129414Y (7 December 2023); https://doi.org/10.1117/12.3011509

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
7 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Photovoltaics

Data modeling

Feature extraction

Neural networks

Solar energy

Transformers

Modeling

Keywords/Phrases

Search In:

Publication Years