Scholarlink

<p>Traditional electromagnetic induction heating deicing is limited by low magnetic energy conversion efficiency, severe magnetic flux leakage, and poor deicing performance. To improve magnetic energy utilization, this study proposes a novel asphalt pavement structure using industrial solid wastes, which consists of a magnetic conduction layer and an induction surface layer. The magnetic conduction layer captures and guides the magnetic field to the surface layer, where magnetic energy is converted into eddy current thermal energy. After verifying the pavement performance, key factors influencing heating and deicing were investigated. An interpretable machine learning model optimized by bionic heuristic strategies was established to predict the deicing rate, and the SHAP method was used to analyze feature contributions. The results show that the magnetic conduction layer significantly enhances magnetic energy conversion efficiency, with a maximum improvement of 108.9%, a 19.5&#xa0;°C increase in surface temperature, and a 1.96-fold higher deicing rate. The optimized BP model exhibits the best fitting and generalization ability. This work achieves efficient and low-energy deicing, providing a reference for intelligent and sustainable pavement operation and maintenance.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Asphalt Pavement Inductive Deicing Technology Integrating Efficient Magnetothermal Conversion with Explainable Intelligent Prediction

  • 期刊论文

  • 发表日期：

    2026年7月1日

查看全文
  • Chengxiang Wang,
  • Xiaoqian Guo,
  • Jiahui Peng,
  • Zhiquan Zhang,
  • Enhui Yang,
  • Dong Liang,
  • Yanjun Qiu

摘要

Traditional electromagnetic induction heating deicing is limited by low magnetic energy conversion efficiency, severe magnetic flux leakage, and poor deicing performance. To improve magnetic energy utilization, this study proposes a novel asphalt pavement structure using industrial solid wastes, which consists of a magnetic conduction layer and an induction surface layer. The magnetic conduction layer captures and guides the magnetic field to the surface layer, where magnetic energy is converted into eddy current thermal energy. After verifying the pavement performance, key factors influencing heating and deicing were investigated. An interpretable machine learning model optimized by bionic heuristic strategies was established to predict the deicing rate, and the SHAP method was used to analyze feature contributions. The results show that the magnetic conduction layer significantly enhances magnetic energy conversion efficiency, with a maximum improvement of 108.9%, a 19.5 °C increase in surface temperature, and a 1.96-fold higher deicing rate. The optimized BP model exhibits the best fitting and generalization ability. This work achieves efficient and low-energy deicing, providing a reference for intelligent and sustainable pavement operation and maintenance.