Scholarlink

<p>Existing models for predicting the resilient modulus of subgrade soils can accurately characterize the dynamic stress zone. However, they suffer from two primary limitations: 1) an inability to predict modulus values within the no-dynamic stress zone, and 2) the inaccessibility of these modulus values through conventional dynamic triaxial tests. To address these shortcomings, this paper introduces a comprehensive methodology for determining the resilient modulus of subgrade soil under a wide range of stress states. The no-dynamic stress zone is conceptualized as comprising two distinct zones: the no-stress zone and the self-weight stress zone. First, an integrated testing framework encompassing the entire subgrade domain is proposed. For the dynamic action zone, a testing method accounting for the combined influence of dynamic and static stresses is employed. For the area devoid of dynamic action, a novel testing method based on the dynamic triaxial apparatus is presented to measure the soil modulus in the no-dynamic stress zone. This method determines the modulus by maintaining constant confining and contact stresses while progressively reducing the cyclic stress. Subsequently, utilizing Gene Expression Programming (GEP), a unified prediction model integrating both the traditional resilient modulus and the modulus under no-dynamic conditions is developed, incorporating four fitting parameters. The results indicate that under low cyclic stress amplitudes (0 – 10 kPa), the dynamic resilient modulus of the subgrade soil increases with increasing cyclic stress. In contrast, under relatively high cyclic stresses, the dynamic resilient modulus decreases as the cyclic stress increases. Compared to the classic NCHRP 1-28A, Uzan, Ni, and <i>K-θ</i> models, the prediction error under small cyclic stresses is reduced from 15.49%, 16.15%, 18.85%, and 22.77%, respectively, to 4.68% using the proposed model. Furthermore, the model maintains high prediction accuracy across a total of 176 working conditions tested on 11 soil samples.</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

Characteristics of subgrade soil resilient modulus considering wide-range stress state

  • 期刊论文

  • 发表日期：

    2026年5月14日

查看全文
  • Jun-hui Zhang,
  • Hua-yu Shang,
  • Hai-shan Fan

摘要

Existing models for predicting the resilient modulus of subgrade soils can accurately characterize the dynamic stress zone. However, they suffer from two primary limitations: 1) an inability to predict modulus values within the no-dynamic stress zone, and 2) the inaccessibility of these modulus values through conventional dynamic triaxial tests. To address these shortcomings, this paper introduces a comprehensive methodology for determining the resilient modulus of subgrade soil under a wide range of stress states. The no-dynamic stress zone is conceptualized as comprising two distinct zones: the no-stress zone and the self-weight stress zone. First, an integrated testing framework encompassing the entire subgrade domain is proposed. For the dynamic action zone, a testing method accounting for the combined influence of dynamic and static stresses is employed. For the area devoid of dynamic action, a novel testing method based on the dynamic triaxial apparatus is presented to measure the soil modulus in the no-dynamic stress zone. This method determines the modulus by maintaining constant confining and contact stresses while progressively reducing the cyclic stress. Subsequently, utilizing Gene Expression Programming (GEP), a unified prediction model integrating both the traditional resilient modulus and the modulus under no-dynamic conditions is developed, incorporating four fitting parameters. The results indicate that under low cyclic stress amplitudes (0 – 10 kPa), the dynamic resilient modulus of the subgrade soil increases with increasing cyclic stress. In contrast, under relatively high cyclic stresses, the dynamic resilient modulus decreases as the cyclic stress increases. Compared to the classic NCHRP 1-28A, Uzan, Ni, and K-θ models, the prediction error under small cyclic stresses is reduced from 15.49%, 16.15%, 18.85%, and 22.77%, respectively, to 4.68% using the proposed model. Furthermore, the model maintains high prediction accuracy across a total of 176 working conditions tested on 11 soil samples.