Education, Science, Technology, Innovation and Life
Open Access
Sign In
  1. Home
  2. Journals
  3. Automation and Machine Learning
  4. Vol 7, Issue 2, 2026

High-Precision Photovoltaic Power Forecasting under Complex Weather Conditions

Download as PDF

DOI: 10.23977/autml.2026.070206 | Downloads: 0 | Views: 49

Author(s)

Shaoyi Sun 1, Chunyu Ma 1

Affiliation(s)

1 School of Electronic and Optical Engineering, Nanjing University of Science and Technology, Nanjing, China

Corresponding Author

Shaoyi Sun

ABSTRACT

To address the issues of strong fluctuations and significant nonlinear characteristics in photovoltaic (PV) power generation under complex weather conditions, as well as the insufficient accuracy of traditional prediction models, a PV power prediction model integrating Convolutional Neural Networks (CNN), Bidirectional Gated Recurrent Units (BiGRU), and an attention mechanism is proposed. First, CNN is used to extract local fluctuation features and multivariate coupling information from the PV power sequence. Then, BiGRU is used to learn the bidirectional dynamic dependencies in the time series. Finally, an attention mechanism is introduced to dynamically weight features at key time steps, enhancing the model's ability to perceive complex weather changes. Experiments are conducted based on real PV power plant data, and comparative analyses are performed using models such as BP, GA-BP, CNN, LSTM, GRU, and CNN-GRU. Experimental results show that the predicted curve of the proposed model has a high consistency with the actual power change trend. Furthermore, monthly error analysis and attention visualization results further verify the model's stability and effectiveness under complex weather scenarios. The research results indicate that the proposed model can effectively improve the accuracy of PV power prediction, providing technical support for new energy grid-connected operation and grid dispatch.

KEYWORDS

Photovoltaic Power Forecasting; Renewable Energy; Time Series Prediction; Intelligent Forecasting

CITE THIS PAPER

Shaoyi Sun, Chunyu Ma. High-Precision Photovoltaic Power Forecasting under Complex Weather Conditions. Automation and Machine Learning (2026). Vol. 7, No. 2, 48-59. DOI: http://dx.doi.org/10.23977/autml.2026.070206.

REFERENCES

[1] Zhao Z. Current research status and prospects of electrode boilers under the background of the "dual carbon" goals[J]. Energies, 2025, 18(4): 769.
[2] Yi Y. Life cycle assessment of energy storage technologies for new power systems under dual‐carbon target: a review[J]. Energy Technology, 2024, 12(5): 2301129.
[3] Saleh H M, Hassan A I. The challenges of sustainable energy transition: A focus on renewable energy[J]. Applied Chemical Engineering, 2024, 7(2): 2084.
[4] Elkelawy M, Saeed A M, Seleem H. Egypt's solar revolution: A Dual approach to clean energy with CSP and PV technologies[J]. Pharos Engineering Science Journal, 2024, 1(1): 39-50.
[5] Verma S, Kameswari Y L, Kumar S. A review on environmental parameters monitoring systems for power generation estimation from renewable energy systems[J]. BioNanoScience, 2024, 14(4): 3864-3888.
[6] Tripathi A K. Advancing solar PV panel power prediction: A comparative machine learning approach in fluctuating environmental conditions[J]. Case Studies in Thermal Engineering, 2024, 59: 104459.
[7] Meghana P. Analysis of neural network algorithm in comparison to multiple linear regression and random forest algorithm[C]//2024 ASU International Conference in Emerging Technologies for Sustainability and Intelligent Systems (ICETSIS). IEEE, 2024: 437-443.
[8] Lin C M, Lin Y S. TPTM-HANN-GA: a novel hyperparameter optimization framework integrating the Taguchi method, an artificial neural network, and a genetic algorithm for the precise prediction of cardiovascular disease risk[J]. Mathematics, 2024, 12(9): 1303.
[9] Sivadasan E T, Sundaram N M, Santhosh R. Deep Learning for Energy Forecasting Using Gated Recurrent Units and Long Short-Term Memory[J]. Journal of Intelligent Systems & Internet of Things, 2025, 14(1).
[10] Shantal M, Othman Z, Abu Bakar A. Missing data imputation using correlation coefficient and min-max normalization weighting[J]. Intelligent Data Analysis, 2025, 29(2): 372-384.
[11] Bai A. Sliding-window CNN+ channel-time attention transformer network trained with inertial measurement units and surface electromyography data for the prediction of muscle activation and motion dynamics leveraging IMU-only wearables for home-based shoulder rehabilitation[J]. Sensors, 2025, 25(4): 1275.
[12] Ullah K. Short-term load forecasting: A comprehensive review and simulation study with CNN-LSTM hybrids approach[J]. IEEE Access, 2024, 12: 111858-111881.
[13] Mo S. From global to local: A lightweight CNN approach for long-term time series forecasting[J]. Computers and Electrical Engineering, 2025, 123: 110192.
[14] Tran-Ngoc H. A novel framework using gated recurrent units and residual network for time-series data recovery in structural health monitoring[J]. European Journal of Environmental and Civil Engineering, 2026, 30(1): 1-28.
[15] Feng Y. Prediction of network security situation based on attention mechanism and convolutional neural network–gated recurrent unit[J]. Applied Sciences, 2024, 14(15): 6652.

Subscription
E-Mail Alert
Most popular papers
Call for Special Issue
Publication Ethics & OA Statement
Plagiarism Policy
Contact Us
Downloads: 5501
Visits: 282893

Sponsors, Associates, and Links

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.