Scholarlink

<p>The health condition of expansion joints, which are critical components linking bridge structures, is crucial for the safety and stability of bridges. This study proposes a method for predicting the evolution intervals of expansion joints by integrating complex network topology features with an enhanced fuzzy information granulation technique, aiming to establish an intelligent framework for analyzing multiscale monitoring data. Methodologically, the study first develops a topological correlation model for the deformation behavior of expansion joints based on complex network theory, which uncovers the nonlinear interaction mechanisms among monitoring data collected from multiple locations. Subsequently, the fuzzy granulation technique is refined, and new parameters are introduced to establish an optimized dynamic membership function, enabling a reasonable partitioning of the feature space while preserving the integrity of the nonlinear characteristics of the original data. The methodology is validated through experiments using four sets of multi-source measurement data from expansion joints under various service environments (including time scales of 0.5 and 1 h). The results demonstrate that, compared to other benchmark models, the proposed method achieves a significant improvement in both the coverage probability and the interval width indices. The Prediction Interval Nominal Confidence (PICP) meets the Prediction Interval Nominal Confidence (PINC), and the Coverage Width-based Criterion (CWC) remains consistently below 0.24, representing a 41.7% average reduction compared to the benchmark models.</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

Interval prediction of bridge expansion joint evolution: based on complex network analysis and fuzzy information granulation

  • 期刊论文

  • 发表日期：

    2025年6月5日

查看全文
  • Guanjun Lv,
  • Liangchao Chen,
  • Pengchao Wang,
  • Qianlin Wang,
  • Jinghai Li,
  • Jianwen Zhang,
  • Ahmed Mebarki,
  • Zhan Dou

摘要

The health condition of expansion joints, which are critical components linking bridge structures, is crucial for the safety and stability of bridges. This study proposes a method for predicting the evolution intervals of expansion joints by integrating complex network topology features with an enhanced fuzzy information granulation technique, aiming to establish an intelligent framework for analyzing multiscale monitoring data. Methodologically, the study first develops a topological correlation model for the deformation behavior of expansion joints based on complex network theory, which uncovers the nonlinear interaction mechanisms among monitoring data collected from multiple locations. Subsequently, the fuzzy granulation technique is refined, and new parameters are introduced to establish an optimized dynamic membership function, enabling a reasonable partitioning of the feature space while preserving the integrity of the nonlinear characteristics of the original data. The methodology is validated through experiments using four sets of multi-source measurement data from expansion joints under various service environments (including time scales of 0.5 and 1 h). The results demonstrate that, compared to other benchmark models, the proposed method achieves a significant improvement in both the coverage probability and the interval width indices. The Prediction Interval Nominal Confidence (PICP) meets the Prediction Interval Nominal Confidence (PINC), and the Coverage Width-based Criterion (CWC) remains consistently below 0.24, representing a 41.7% average reduction compared to the benchmark models.