RESEARCH ON POWER QUALITY DISTURBANCE CLASSIFICATIONMETHOD BASED ON IMAGE FUSION AND MULTIMODAL FEATURE DRIVEN APPROACH

Zhang Zhenghao; Wu Kexin; Chen Peng; Xue Qiang; Wang Shuohe

doi:10.19912/j.0254-0096.tynxb.2025-0450

PDF(2623 KB)

Acta Energiae Solaris Sinica ›› 2026, Vol. 47 ›› Issue (7) : 736-746. DOI: 10.19912/j.0254-0096.tynxb.2025-0450

RESEARCH ON POWER QUALITY DISTURBANCE CLASSIFICATIONMETHOD BASED ON IMAGE FUSION AND MULTIMODAL FEATURE DRIVEN APPROACH

Zhang Zhenghao^1~3, Wu Kexin^1~3, Chen Peng^2~4, Xue Qiang^1~3, Wang Shuohe^1~3

Author information +

History +

Abstract

With the continuous development of new power systems integrated with photovoltaics, the types of power quality disturbances （PQDs） have become more complex. In response to the increasing proportion of composite disturbances, which makes it difficult for traditional methods to accurately classify complex PQDs under the guidance of human experience, a complex PQDs classification method based on image fusion and multimodal feature-driven approach is proposed. In terms of data processing, two-dimensional image encoding methods—recurrence plot （RP） and continuous wavelet transform （CWT）—are introduced to convert one-dimensional PQDs signals into two-dimensional images. RP-CWT feature images are generated through horizontal concatenation to achieve feature enhancement. At the model level, an ensemble model for multimodal fusion with parallel optimization of re-parameterized large kernel network （RepLKNet） and bidirectional gated recurrent unit （BiGRU） is established. This model captures high-dimensional features in the time, frequency, and spatial dimensions, enabling rapid and accurate identification of disturbance features in signals. Finally, tests based on simulation data and a PQDs experimental platform show that the proposed method can accurately classify complex PQDs.

Key words

power quality disturbances / multimodal fusion / continuous wavelet transform / recurrence plot / bidirectional gated recurrent unit / re-parameterized large kernel network

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

Zhang Zhenghao, Wu Kexin, Chen Peng, Xue Qiang, Wang Shuohe. RESEARCH ON POWER QUALITY DISTURBANCE CLASSIFICATIONMETHOD BASED ON IMAGE FUSION AND MULTIMODAL FEATURE DRIVEN APPROACH[J]. Acta Energiae Solaris Sinica. 2026, 47(7): 736-746 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0450

References

[1] 曾建斌, 杨洋, 曾晋, 等. 弱电网下光伏并网逆变器稳定性改进方法[J]. 太阳能学报, 2025, 46(2): 387-392.
Zeng J B, Yang Y, Zeng J, et al.Stability improvement method of photovoltaic grid-connected inverter under weak grid[J]. Acta Energiae Solaris Sinica, 2025, 46(2): 387-392.
[2] 张原飞, 王项南, 夏海南, 等. 基于功率区间划分的潮流能发电装置电能质量特性分析[J]. 太阳能学报, 2022, 43(4): 506-511.
Zhang Y F, Wang X N, Xia H N, et al.Analysis of power quality characteristics of tidal current power generation device based on power interval division[J]. Acta Energiae Solaris Sinica, 2022, 43(4): 506-511.
[3] 陈蓉, 杨勇, 樊明迪. 新型电力系统下电能质量信号的分数域降噪方法研究[J]. 电测与仪表, 2023, 60(9): 81-87.
Chen R, Yang Y, Fan M D.De-noising method of fractional domain for power quality signals in novel power system[J]. Electrical Measurement & Instrumentation, 2023, 60(9): 81-87.
[4] Ma J, Tang Q, He M J, et al.Complex PQD classification using time-frequency analysis and multiscale parallel attention residual network[J]. IEEE Transactions on Industrial Electronics, 2024, 71(8): 9658-9667.
[5] 陈诺, 吕干云, 叶加星. 基于SVM级联决策树的复合电能质量扰动识别[J]. 电气工程学报, 2023, 18(2): 149-156.
Chen N, Lyu G Y, Ye J X.Recognition of complex PQ disturbances based on SVM cascaded decision tree[J]. Journal of Electrical Engineering, 2023, 18(2): 149-156.
[6] 张彼德, 邱杰, 娄广鑫, 等. 基于CNN和Transformer的轻量化电能质量扰动识别模型[J]. 电力工程技术, 2025, 44(1): 69-78.
Zhang B D, Qiu J, Lou G X, et al.A lightweight power quality disturbance recognition model based on CNN and Transformer[J]. Electric Power Engineering Technology, 2025, 44(1): 69-78.
[7] 王义国, 林峰, 李琦, 等. 基于TCN-LSTM模型的电网电能质量扰动分类研究[J]. 电力系统保护与控制, 2024, 52(17): 161-167.
Wang Y G, Lin F, Li Q, et al.Classification of power quality disturbances in a power grid based on the TCN-LSTM model[J]. Power System Protection and Control, 2024, 52(17): 161-167.
[8] Zhang Y W, Zhang Y, Zhou X H.Classification of power quality disturbances using visual attention mechanism and feed-forward neural network[J]. Measurement, 2022, 188: 110390.
[9] 王昭卿, 常延朝, 陈建磊, 等. 基于多融合MVMD-ISVM的复杂电能质量扰动识别方法[J]. 供用电, 2024, 41(9): 70-77.
Wang Z Q, Chang Y C, Chen J L, et al.Identification method for complex power quality disturbances based on multi-fusion MVMD-ISVM[J]. Distribution & Utilization, 2024, 41(9): 70-77.
[10] Liu Y L, Yuan D, Gong Z, et al.Adaptive spectral trend based optimized EWT for monitoring the parameters of multiple power quality disturbances[J]. International Journal of Electrical Power & Energy Systems, 2023, 146: 108797.
[11] Singh G, Pal Y, Dahiya A K.Classification of power quality disturbances using linear discriminant analysis[J]. Applied Soft Computing, 2023, 138: 110181.
[12] Salles R S, Ribeiro P F.The use of deep learning and 2-D wavelet scalograms for power quality disturbances classification[J]. Electric Power Systems Research, 2023, 214: 108834.
[13] 龚正, 邹阳, 金涛, 等. 基于特征融合并行优化模型的电能质量扰动分类方法[J]. 中国电机工程学报, 2023, 43(3): 1017-1026.
Gong Z, Zou Y, Jin T, et al.Classification method of power quality disturbances based on optimized parallel model of features merging[J]. Proceedings of the CSEE, 2023, 43(3): 1017-1026.
[14] 贺才郡, 李开成, 杨王旺, 等. 基于双通道GAF和深度残差网络的电能质量复合扰动识别[J]. 电网技术, 2023, 47(1): 369-376.
He C J, Li K C, Yang W W, et al.Power quality compound disturbance identification based on dual channel GAF and depth residual network[J]. Power System Technology, 2023, 47(1): 369-376.
[15] 王继东, 王泽平, 张迪. 基于递归图和预训练迁移学习的电能质量扰动分类[J]. 南方电网技术, 2025, 19(2): 48-56, 114.
Wang J D, Wang Z P, Zhang D.Power quality disturbance classification based on recursive graph and pre-trained transfer learning[J]. Southern Power System Technology, 2025, 19(2): 48-56, 114.
[16] Jiang J Z, Wu H, Zhong C H, et al.A novel methodology for microgrid power quality disturbance classification using URPM-CWT and multi-channel feature fusion[J]. IEEE Access, 2024, 12: 35597-35611.
[17] Zhu H Z, Fang T.Compound fault diagnosis of rolling bearings with few-shot based on DCGAN-RepLKNet[J]. Measurement Science and Technology, 2024, 35(6): 066105.
[18] Ma S X, Wang L Y, Chen P P, et al.A mixed visual encoding model based on the larger-scale receptive field for human brain activity[J]. Brain Sciences, 2022, 12(12): 1633.
[19] Bano S, Khalid S, Tairan N M, et al.Summarization of scholarly articles using BERT and BiGRU: deep learning-based extractive approach[J]. Journal of King Saud University-Computer and Information Sciences, 2023, 35(9): 101739.
[20] Li X L, Zhang Y Y, Jin J Y, et al.A model of integrating convolution and BiGRU dual-channel mechanism for Chinese medical text classifications[J]. PLoS One, 2023, 18(3): e0282824.
[21] Lv H X, Chen J L, Pan T Y, et al.Attention mechanism in intelligent fault diagnosis of machinery: a review of technique and application[J]. Measurement, 2022, 199: 111594.
[22] Obeso A M, Benois-Pineau J, GarcíA VáZquez M S, et al. Visual vs internal attention mechanisms in deep neural networks for image classification and object detection[J]. Pattern Recognition, 2022, 123: 108411.