Icing Thickness Prediction for Transmission Lines Using Attention-CNN-GRU with Real-Time Meteorological Observation Data

WANG Yanjun; HOU Hui; ZHENG Wenzhe; WANG Zhenguo

doi:10.16516/j.ceec.2026-109

English Version Journal Login

您当前的位置：

首页 >

文章列表页 >

Icing Thickness Prediction for Transmission Lines Using Attention-CNN-GRU with Real-Time Meteorological Observation Data

更新时间：2026-10-14

- Icing Thickness Prediction for Transmission Lines Using Attention-CNN-GRU with Real-Time Meteorological Observation Data
- Southern Energy Construction Vol. 13, Issue 3, Pages: 45-55(2026)
- 作者机构：
  
  1. 武汉理工大学新能源与电气工程学院,湖北,武汉,430070
  2. 国网浙江省电力有限公司电力科学研究院,浙江,杭州,310014
- 作者简介：
- 基金信息：
- DOI：10.16516/j.ceec.2026-109
  CLC： TM7;P49
- Published：2026
- 稿件说明：
移动端阅览
王艳君, 侯慧, 郑文哲, 等. 常规气象实况观测数据下的Attention-CNN-GRU架空输电线路覆冰厚度预测[J]. 南方能源建设, 2026, 13(3): 45-55.

.
王艳君, 侯慧, 郑文哲, 等. 常规气象实况观测数据下的Attention-CNN-GRU架空输电线路覆冰厚度预测[J]. 南方能源建设, 2026, 13(3): 45-55. DOI： 10.16516/j.ceec.2026-109.

. DOI： 10.16516/j.ceec.2026-109.

摘要

目的

现有架空输电线路覆冰厚度预测方法多依赖导线应力、拉力等力学监测数据，存在实际工程中专业数据难以获取的缺陷。在规避导线应力监测数据依赖的前提下，仅依托易获取的常规气象实况观测数据开展建模，挖掘常规气象实况观测数据与覆冰生长演化的非线性耦合关系，实现低成本、高精度的线路覆冰厚度预测，为电网冰灾防控提供支撑。

方法

以温度、湿度、风速、气压等常规气象实况观测数据作为输入，构建注意力机制-卷积神经网络-门控循环单元（Attention-Convolutional Neural Network-Gated Recurrent Unit，Attention-CNN-GRU）组合预测模型；利用卷积神经网络（Convolutional Neural Network，CNN）捕捉输入特征序列短时波动与突变细节特征，结合门控循环单元（Gated Recurrent Unit，GRU）刻画覆冰累积演变的长时序依赖规律，通过Attention自适应强化关键致冰因子权重；设置多组时间步长开展对比试验，完成不同时序尺度下模型性能的系统性验证。

结果

实测架空输电线路数据验证表明，Attention-CNN-GRU相较于基础GRU、长短期记忆网络（Long Short-Term Memory，LSTM）预测误差分别降低26.1%和57.3%，预测精度与稳定性显著提升；全程仅采用常规气象实况观测数据建模，一定程度上规避对导线应力专用传感设备的依赖，有效降低监测部署与运维成本。

结论

该模型在仅使用常规气象实况观测数据的条件下，仍能实现良好的时序预测效果，性能优于传统单一时序网络。该方法数据易得、工程推广性强，可支撑线路覆冰实时研判、分级防控与智能决策。

Abstract

Objective

Existing transmission line icing thickness prediction methods mostly rely on real-time mechanical monitoring data such as conductor stress and tension

hich are difficult to obtain in practical engineering. Without relying on conductor stress monitoring data

a prediction model is developed basd solely on real-time meteorological observation data to explore the nonlinear coupling relationship between meteorological parameters and icing growth and evolution

achieve low-cost and high-precision transmission line icing thickness prediction

and provide support for on-line early warning and risk prevention and control of power grid icing disasters.

Method

An Attention-convolutional neural network-gated recurrent unit （Attention-CNN-GRU） hybrid prediction model was constructed by taking conventional meteorological data such as temperature

humidity

wind speed and air pressure as the sole input. Convolutional neural network （CNN） was used to capture short-term fluctuations and mutation detail features of meteorological sequences

gated recurrent unit （GRU） was adopted to describe the long-term time series dependence of icing accumulation and evolution

and the Attention mechanism was applied to adaptively enhance the weight of key ice-inducing meteorological factors. Comparative experiments with multiple time steps were conducted to systematically verify the model performanceunder different time series scales.

Result

Verification based on measured transmission line data indicates that the Attention-CNN-GRU model reduces the prediction error by 26.1% and 57.3% respectively compared with the basic GRU and long short-term memory （LSTM） network

and the prediction accuracy and stability are significantly improved. The whole modeling process only adopts real-time meteorological observation data

completely getting rid of the dependence on special sensor equipment for conductor stress monitoring

and effectively cutting down the cost of monitoring deployment

operation and maintenance.

Conclusion

The model can achieve satisfactory time-series prediction performance using only real-time meteorological observation data and outperforms t

raditional single time-series networks. The proposed method features readily available data and strong engineering applicability

enabling support for real-time judgment

hierarchical prevention and control

as well as intelligent decision-making for transmission line icing.

关键词

Keywords

references

ZHOU H L, ZHANG L Q, DONG X G, et al. Prediction of ice accretion on Qinling transmission lines integrating image recognition and numerical simulation [J]. Southern energy construction, 2026, 13(1): 182-190. DOI: 10.16516/j.ceec.2025-073.

MA F Q, LI J L, JIA R, et al. A fine identification method for transmission line ice coating types based on bilateral feature-fused semantic segmentation [J]. Smart power, 2025, 53(10): 17-25. DOI: 10.20204/j.sp.2025.10003.

LI J S, WANG S, GUAN S P, et al. Research on response of conductor icing and de-icing and prediction model of de-icing height [J]. Electrical measurement instrumentation, 2025, 62(8): 75-84. DOI: 10.19753/j.issn1001-1390.2025.08.009.

LUO J B, YU C, XIE Y Y, et al. A review on risk assessment of power grid security and stability under natural disasters [J]. Power system protection and control, 2018, 46(6): 158-170. DOI: 10.7667/PSPC170296.

ZHANG Z C, WANG J. Prediction model for icing thickness of power transmission line based on dynamic tension and dip angle [J]. Guangdong electric power, 2015, 28(6): 82-86, 92. DOI: 10.3969/j.issn.1007-290X.2015.06.016.

ZHOU Z H, SHU Z Y, LEI M, et al. Research on transmission line ice weight estimation using image fusion [J]. Computer measurement control, 2026, 34(3): 216-222. DOI: 10.16526/j.cnki.11-4762/tp.2026.03.027.

LI J, WANG S Y. Prediction of ice load on power lines incorporating meteorological characteristics [J]. Zhejiang electric power, 2026, 45(4): 3-11. DOI: 10.19585/j.zjdl.202604001.

HU J L, LIU J, JIANG X L, et al. Monitoring method for equivalent icing thickness of conductor under complex load based on sag measurement [J]. High voltage engineering, 2022, 48(2): 584-593. DOI: 10.13336/j.1003-6520.hve.20210217.

HAO Y P, MAO D M, CHENG Y T, et al. Test method for minimum critical thickness of ice coating on OPGW based on BOTDR monitoring [J]. High voltage engineering, 2025, 51(7): 3191-3200. DOI: 10.13336/j.1003-6520.hve.20241167.

WANG Y M, TANG J P, AO T, et al. Identification of meteorological hazards and spatiotemporal distribution characteristics of power grids in complex terrain [J]. Southern energy construction, 2026(3): 82-91. DOI: 10.16516/j.ceec.2025-177.

XIONG W, XU H, XU L X, et al. Combined model of icing prediction of transmission lines based on RF-APJA-MKRVM considering time cumulative effect [J]. High voltage engineering, 2022, 48(3): 948-957. DOI: 10.13336/j.1003-6520.hve.20210151.

DOU W W, PANG Y H, ZHANG T T, et al. High-precision icing prediction technology for transmission lines based on LSTM [J]. Power system and clean energy, 2025, 41(6): 11-19. DOI: 10.3969/j.issn.1674-3814.2025.06.002.

GAO S, XU P H, CHEN Z H, et al. A short-term calgorithm based on improved LSTM neural network [J]. Southern energy construction, 2024, 11(1): 112-121. DOI: 10.16516/j.ceec.2024.1.12.

XI X Z, XING C, QIN R S, et al. Power quality disturbance recognition method based on multi-layer feature fusion attention network [J]. Smart power, 2022, 50(10): 37-44. DOI: 10.3969/j.issn.1673-7598.2022.10.007.

CAI G F, WANG Q B, CHEN Z Y, et al. Analysis on electrocution diagnosis based on discrete wavelet transform and GRU [J]. Southern energy construction, 2024, 11(4): 127-136. DOI: 10.16516/j.ceec.2024.4.13.

TIAN R, LYU Z Q, FAN Y H, et al. Qualitative classification of waste garments for textile recycling based on machine vision and attention mechanisms [J]. Waste management, 2024, 183: 74-86. DOI: 10.1016/j.wasman.2024.04.040.

YANG J, WANG L, SONG D R, et al. Diagnostic method of zero-point shifting of wind turbine yaw angle based on isolated forest and sparse Gaussian process regression [J]. Proceedings of the CSEE, 2021, 41(18): 6198-6211. DOI: 10.13334/j.0258-8013.pcsee.202224.

XU M J, LI X, LI Q W, et al. LNBi-GRU model for coal price prediction and pattern recognition analysis [J]. Applied energy, 2024, 365: 123302. DOI: 10.1016/j.apenergy.2024.123302.

JIA R, YANG G H, ZHENG H F, et al. Combined wind power prediction method based on CNN-LSTMGRU with adaptive weights [J]. Electric power, 2022, 55(5): 47-56,110. DOI: 10.11930/j.issn.1004-9649.202104023.

CHEN Z Y, LIU J B, LI C, et al. Ultra short-term power load forecasting based on combined LSTM-XGBoost model [J]. Power system technology, 2020, 44(2): 614-620. DOI: 10.13335/j.1000-3673.pst.2019.1566.

KANAI S, FUJIWARA Y, IWAMURA S, et al. Training of gated recurrent units constrained to locally stable to prevent the gradient exploding [J]. IEICE transactions on information and systems, 2019, J102-D(8): 530-541.

ZU X R, SONG R X. Short-term wind power prediction method based on wavelet packet decomposition and improved GRU [J]. Journal of physics: conference series, 2018, 1087(2): 022034. DOI: 10.1088/1742-6596/1087/2/022034.

WANG C, CHEN Q, GU X M, et al. Assessment method for health state of li-ion batteries based on sparrow search algorithm [J]. Southern energy construction, 2023, 10(6): 89-97. DOI: 10.16516/j.gedi.issn2095-8676.2023.06.010.

HOU H, WU W J, WEI R Z, et al. Attention mechanism based CNN-LSTM-XGBoost electric power meteorological hybrid forecasting model of typhoon rainstorm [J]. Smart power, 2024, 52(10): 96-102. DOI: 10.20204/j.sp.2024.10013.

GAO S, XU P H, CHEN Z H. Wind speed multi-mode ensemble forecasting for wind farms based on machine learning [J]. Southern energy construction, 2024, 11(1): 85-95. DOI: 10.16516/j.ceec.2024.1.09.

HOU H, WAN Y, WANG Z G, et al. Wind and photovoltaic generation prediction bi-level model based on uncertainty scenarios under typhoon disaster [J]. Electric power construction, 2025, 46(3): 146-154. DOI: 10.12204/j.issn.1000-7229.2025.03.012.

ZHU J N, WANG Z, LI Y T, et al. Multi-scale fusion prediction method for wind power cluster power based on spatio-temporal graph spectral clustering and time-series 2D-reshaping Transformer [J]. Electric power construction, 2026, 47(3): 39-50. DOI: 10.12204/j.issn.1000-7229.2026.03.004.

YI T, WANG P B, LIU Y, et al. Trading strategies of power generation entities in the voluntary emission reduction market [J]. Electric power construction, 2025, 46(12): 20-32. DOI: 10.12204/j.issn.1000-7229.2025.12.003.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Review of Safety Risk Early Warning Technology and Engineering Application for Lithium Battery Energy Storage Power Stations

A Review of Monitoring and Early Warning Technologies for Multi-Resource Hybrid Energy Storage Systems

Research on the Temperature Suppression Technology for Artificial Cavern Walls Based on Inflated Circular Airflow

Advances in Numerical Simulation for Deep Underground Energy Storage and Compressed Air Energy Storage: A Review

Multi-Time Scale Inertia Coordination and Transient Stability Optimization Control of High Proportion Wind-Solar Cluster Grid-Connected

Related Author

No data

Related Institution

No data

AI问答

⁰