Scholarlink

<p>Saskatchewan, with over 250,000&#xa0;km (160,000 mi) of roads, holds the most extensive road network per capita in Canada. The region utilizes over 26,500 culverts, which are crucial for managing water flow along these roadways. Many of these culverts, including some installed over a century ago, lack recorded installation dates, heightening risks associated with their failure. Culvert failures can lead to severe consequences, including infrastructure damage, environmental degradation, and safety hazards, impacting both water quality and aquatic habitats. This study evaluates three analytical methods-ordinal logistic regression, artificial neural networks, and a Fuzzy Inference System-to prioritize renewals of metal culverts, which are predominantly used in the province. By assessing 1000 metal culverts on Saskatchewan highways, these methods were tested to establish renewal priorities based on their condition. The effectiveness of each method was measured using metrics such as the area under the ROC curve, prediction accuracy, and a suite of precision, recall, and F1 scores. The results show that the artificial neural network, optimized by a genetic algorithm, significantly outperformed the other models, providing a robust framework for efficient culvert renewal prioritization.</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

Smart infrastructure renewal: prioritizing metal culverts using hybrid analytical methods

  • 期刊论文

  • 发表日期：

    2026年8月4日

查看全文
  • Pooneh Marzbani,
  • Golam Kabir,
  • Niamat Ullah Ibne Hossain

摘要

Saskatchewan, with over 250,000 km (160,000 mi) of roads, holds the most extensive road network per capita in Canada. The region utilizes over 26,500 culverts, which are crucial for managing water flow along these roadways. Many of these culverts, including some installed over a century ago, lack recorded installation dates, heightening risks associated with their failure. Culvert failures can lead to severe consequences, including infrastructure damage, environmental degradation, and safety hazards, impacting both water quality and aquatic habitats. This study evaluates three analytical methods-ordinal logistic regression, artificial neural networks, and a Fuzzy Inference System-to prioritize renewals of metal culverts, which are predominantly used in the province. By assessing 1000 metal culverts on Saskatchewan highways, these methods were tested to establish renewal priorities based on their condition. The effectiveness of each method was measured using metrics such as the area under the ROC curve, prediction accuracy, and a suite of precision, recall, and F1 scores. The results show that the artificial neural network, optimized by a genetic algorithm, significantly outperformed the other models, providing a robust framework for efficient culvert renewal prioritization.