Dynamic stress analysis and prediction of distribution conductors under ice-wind loads based on LSTM-DKF
Authors:
Open Access
- 20 May 2026
- journal article
- research article
- Published by Frontiers Media SA in Frontiers in Energy Research
- Vol. 14, 1733953
Abstract
With the increase in extreme weather events, icing disasters pose a significant threat to the safe and stable operation of distribution lines. In this study, a finite element model for stress analysis of iced conductors is established, and the dynamic stress characteristics of three types iced conductors under fluctuating wind conditions are analyzed, with particular attention given to the effects of varying ice thicknesses and wind speeds on the maximum conductor stress. Results show that JKLYJ-10/70 exceeds its tensile limit at an ice thickness of 27 mm, JL/G1A-50/8 reaches its tensile capacity around 47 mm, while JL/G1A-120/20 remains within safe limits even with 50 mm of ice. To achieve accurate dynamic stress prediction, a hybrid model combining a Long Short-Term Memory network and Dynamic Kalman Filtering is proposed. Under the condition of 10 mm ice thickness and 15 m/s wind speed, the proposed method achieves the lowest root mean square error among all compared algorithms, with a value of only 17.02 N, and the model also exhibits the lowest mean absolute error with a value of 16.69 N. The power spectral density exhibits dominant peaks at 1.76 Hz and 2.3 Hz, with the proposed method providing the closest agreement. Relative errors at these frequencies are only 2.49% and 0.78%, indicating that the proposed method more effectively captures the dynamic stress response of iced conductors under fluctuating wind conditions.Keywords
This publication has 9 references indexed in Scilit:
- Research on equivalent bending stiffness of conductor near the tension section suspension point of overhead transmission linesElectric Power Systems Research, 2025
- Machine Learning Techniques for Fault Detection in Smart Distribution GridsEnergies, 2025
- Near real-time machine learning framework in distribution networks with low-carbon technologies using smart meter dataApplied Energy, 2025
- Time history analysis of wind vibration response of offshore photovoltaic structures based on davenport wind spectrumJournal of Physics: Conference Series, 2025
- Ice Cover Prediction for Transmission Lines Based on Feature Extraction and an Improved Transformer SchemeElectronics, 2024
- Deep learning modelling techniques: current progress, applications, advantages, and challengesArtificial Intelligence Review, 2023
- Electrical Model-Free Voltage Calculations Using Neural Networks and Smart Meter DataIEEE Transactions on Smart Grid, 2022
- Assessing the effects of DER on voltages using a smart meter-driven three-phase LV feeder modelElectric Power Systems Research, 2020
- Mechanical Calculation of Power Lines and Strands in Overhead Transmission LinesMATEC Web of Conferences, 2016