A Dataset Augmentation Method for Non-convergent Power Flow Adjustment Behavior Based on Improved Conditional Generative Adversarial Network
Main Article Content
Abstract
This paper proposes a data augmentation method for generating non-convergent power flow adjustment behavior datasets based on an improved generative adversarial network. Firstly, a mechanistic analysis of power flow non-convergence and adjustment strategies is conducted, and metrics for evaluating the non-convergent power flow adjustment behavior dataset are proposed. Secondly, the fundamental principles and architectural features of the Conditional Generative Adversarial Network (CGAN) and Long Short-Term Memory (LSTM) network are introduced. Then, a data-driven augmentation method for the non-convergent power flow adjustment behavior dataset is presented. By incorporating an extreme value theory mechanism, the training capability of the CGAN on the training data is enhanced. Furthermore, LSTM is embedded into the GAN framework to strengthen the learning and capture capability of multi-dimensional features during the non-convergent power flow adjustment process. Finally, case studies are used to validate the effectiveness of the proposed method. The generated dataset of non-convergent power flow adjustment behavior can effectively characterize the complete closed-loop process of a power grid transitioning from an ill-conditioned state back to a solvable state, providing a crucial training and verification foundation for power flow optimization practices.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
References
S. Talebi and K. Zhou, “Graph neural networks for efficient AC power flow prediction in power grids,” Electr. Eng. Syst. Sci., 2025, doi: 10.48550/arX iv.2502.05702.
S. Wu, W. Hu, Z. Lu, et al., “Power system flow adjustment and sample generation based on deep reinforcement learning,” J. Mod. Power Syst. Clean Energy, vol. 8, no. 6, pp. 1115–1127, 2020.
T. Wang and Y. Tang, “Parallel deep reinforcement learning-based power flow state adjustment considering static stability constraint,” IET Gener. Transm. Distrib., vol. 14, no. 3, pp. 6276–6284, 2020.
S. N. Okhuegbe, A. A. Ademola, and Y. Liu, “Newton–Raphson AC power flow convergence based on deep learning initialization and homotopy continuation,” IEEE Trans. Ind. Appl., 2024.
X. Hu, J. Yang, Y. Gao, et al., “Adaptive power flow analysis for power system operation based on graph deep learning,” Int. J. Electr. Power Energy Syst., vol. 161, Art. no. 110166, 2024.
X. Zhang, D. Li, and X. Fu, “A novel Wasserstein generative adversarial network for stochastic wind power output scenario generation,” IET Renew. Power Gener., vol. 18, no. 16, pp. 3731–3742, 2024.
Y. Chen, X. Wang, and B. Zhang, “An unsupervised deep learning approach for scenario forecasts,” in Proc. 2018 Power Syst. Comput. Conf. (PSCC), Dublin, Ireland, 2018, pp. 1–7.
Y. Chen, Y. Wang, D. Kirschen, et al., “Model-free renewable scenario generation using generative adversarial networks,” IEEE Trans. Power Syst., vol. 33, no. 3, pp. 3265–3275, 2018.
Y. Zhang, Q. Ai, F. Xiao, et al., “Typical wind power scenario generation for multiple wind farms using conditional improved Wasserstein generative adversarial network,” Int. J. Electr. Power Energy Syst., vol. 114, Art. no. 105388, 2020, doi: 10.1016/j.ijepes.2019.105388.
Y. Ma, M. Ma, N. Wang, et al., “Mid-term scenario generation for wind power using GAN with temporal-correlation enhancement block,” E3S Web Conf., vol. 6, Art. no. 01003, 2020.
Y. Chen, P. Li, and B. Zhang, “Bayesian renewables scenario generation via deep generative networks,” in Proc. 52nd Annu. Conf. Inf. Sci. Syst. (CISS), Princeton, NJ, USA, Mar. 2018, pp. 1–6.
J. Liang and W. Tang, “Sequence generative adversarial networks for wind power scenario generation,” IEEE J. Sel. Areas Commun., vol. 38, no. 1, pp. 110–118, 2020.
R. Yu, W. Hu, H. Jiang, et al., “Research on scene generation method of wind and solar active power output based on k-medoids clustering and generative adversarial networks,” in Proc. 2021 11th Int. Conf. Power Energy Syst. (ICPES), Shanghai, China, 2021, pp. 690–695.
S. Narayanan, R. Kumar, S. Ramadass, et al., “Innovative hybrid approach for enhanced renewable energy generation forecasting using recurrent neural networks and generative adversarial networks,” J. Electr. Eng. Technol., vol. 19, pp. 4847–4864, 2024.
H. Muller, M. Rider, and C. Castro, “Artificial neural networks for load flow and external equivalents studies,” Electr. Power Syst. Res., vol. 80, no. 9, p. 10, 2024.
Y. Yang, J. Yu, Z. Yang, et al., “Fast calculation of probabilistic optimal power flow: A deep learning approach,” in Proc. 2024 IEEE Power Energy Soc. Gen. Meeting (PESGM), 2024, pp. 1–5.
M. Xiang, J. Yu, Z. Yang, et al., “Probabilistic power flow with topology changes based on deep neural network,” Int. J. Electr. Power Energy Syst., vol. 117, Art. no. 105650, 2024.
P. Li, B. Yu, N. Chang, et al., “An efficient solution method for quadratic convex relaxation unit commitment in AC power flow based on binary neighborhood dynamic expansion,” Proc. Chin. Soc. Electr. Eng., pp. 1– 19, accessed Nov. 13, 2025. [Online]. Available: https://link.cnki.net/urlid/11.2107.tm.20250616.0837.002
J. Guo, S. Fan, Z. Cai, et al., “Human-machine hybrid enhanced intelligence in large power grid control: Conceptual connotation, application framework, key technologies, and system validation,” Proc. Chin. Soc. Electr. Eng., vol. 44, no. 17, pp. 6787–6811, 2024, doi: 10.13334/j.025 8-8013.pcsee.241892.
H. R. Baghaee, M. Mirsalim, and G. Gharehpetian, “Power calculation using RBF neural networks to improve power sharing of hierarchical control scheme in multi-DER microgrids,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 4, no. 4, pp. 1217–1225, 2023.
Z. Ma, Y. Li, X. He, et al., “A graph learning model for static security analysis of power grids based on power flow embedding and minimum cut pooling,” Southern Power Syst. Technol., vol. 19, no. 1, pp. 63–73, 92, 2025, doi: 10.13648/j.cnki.issn1674-0629.2025.01.007.
M. Mirza and S. Osindero, “Conditional generative adversarial nets,” Computer Science, vol. 1411, no. 1784, pp. 2672–2680, 2014.
I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, et al., “Generative adversarial networks,” Adv. Neural Inf. Process. Syst., vol. 3, pp. 2672–2680, 2014.
S. Lee and I. H. T. Kim, “Exponentiated generalized Pareto distribution: Properties and applications towards extreme value theory,” Commun. Stat. Theory Methods, vol. 48, no. 8, pp. 2014–2038, 2019.
A. Al Mamun, M. Sohel, N. Mohammad, et al., “A comprehensive review of the load forecasting techniques using single and hybrid predictive models,” IEEE Access, vol. 8, pp. 134911–134939, 2020.
P. Pinson and R. Girard, “Evaluating the quality of scenarios of short-term wind power generation,” Appl. Energy, vol. 96, pp. 12–20, 2012.
X. Wang, Z. Hu, M. Zhang, et al., “Research on the construction of evaluation framework for short-term wind power scenario quality,” Power Syst. Technol., vol. 41, no. 5, pp. 1583–1589, 2017, doi: 10.13335/j.100 0-3673.pst.2016.1985.