Scholarlink

<p>This study investigates the influence of climate on the International Roughness Index (IRI) of pavements with varying subgrade types using regression analysis and machine learning (ML) techniques. Pavement subgrade type, a critical geotechnical factor, significantly impacts pavement roughness and performance. Data from the Long-Term Pavement Performance (LTPP) program were analyzed, focusing on IRI values measured along the center lane (CLIRI) and the mean of the left and right wheel paths (MIRI). CLIRI was hypothesized to be primarily influenced by climate due to limited traffic exposure, while MIRI reflects combined effects of traffic and climate. Climate impacts were assessed using the Freezing Index (FI) and precipitation (PPT), while subgrade properties were characterized by the Plasticity Index (PI) and percent fines passing the No. 200 sieve (P200). Advanced ML algorithms, including XGBoost, CatBoost, Random Forest, and LightGBM, outperformed regression models in predicting IRI changes. Results indicate that FI and PPT have stronger correlations with CLIRI than MIRI, highlighting the more significant climatic impact on center lane roughness. Models tailored to subgrade types showed fine subgrade pavements were more accurately predicted and more climate-sensitive compared to coarse subgrades. Climate change analysis revealed that fine subgrade pavements may experience reduced roughness due to decreased FI, whereas coarse subgrades may see increased roughness from higher annual precipitation. These findings emphasize the need for subgrade-specific, climate-sensitive IRI models to enhance predictive accuracy and guide adaptive pavement management strategies, fostering resilient infrastructure under changing climatic conditions.</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

The Role of Subgrade Characteristics in Climate-Driven Pavement Roughness: A Machine Learning Approach

  • 期刊论文

  • 发表日期：

    2026年7月1日

查看全文
  • Eyad Abu Rish,
  • Rashid Bashir,
  • Mohammad Shafiee

摘要

This study investigates the influence of climate on the International Roughness Index (IRI) of pavements with varying subgrade types using regression analysis and machine learning (ML) techniques. Pavement subgrade type, a critical geotechnical factor, significantly impacts pavement roughness and performance. Data from the Long-Term Pavement Performance (LTPP) program were analyzed, focusing on IRI values measured along the center lane (CLIRI) and the mean of the left and right wheel paths (MIRI). CLIRI was hypothesized to be primarily influenced by climate due to limited traffic exposure, while MIRI reflects combined effects of traffic and climate. Climate impacts were assessed using the Freezing Index (FI) and precipitation (PPT), while subgrade properties were characterized by the Plasticity Index (PI) and percent fines passing the No. 200 sieve (P200). Advanced ML algorithms, including XGBoost, CatBoost, Random Forest, and LightGBM, outperformed regression models in predicting IRI changes. Results indicate that FI and PPT have stronger correlations with CLIRI than MIRI, highlighting the more significant climatic impact on center lane roughness. Models tailored to subgrade types showed fine subgrade pavements were more accurately predicted and more climate-sensitive compared to coarse subgrades. Climate change analysis revealed that fine subgrade pavements may experience reduced roughness due to decreased FI, whereas coarse subgrades may see increased roughness from higher annual precipitation. These findings emphasize the need for subgrade-specific, climate-sensitive IRI models to enhance predictive accuracy and guide adaptive pavement management strategies, fostering resilient infrastructure under changing climatic conditions.