Scholarlink

<p>Accurately forecasting the strength properties of concrete is critical, especially for design code compliance. While statistical and empirical models are commonly used; however, these methods demand intensive experimentation and may yield inaccuracies in intricate relationships between concrete characteristics, mix components, and curing conditions. To address these limitations, the current study utilized the gene expression programming (GEP) method to forecast the mechanical characteristics of concrete, such as modulus of elasticity (<i>E</i>), modulus of rupture (<i>MOR</i>), and dynamic modulus (<i>DM</i>). For this purpose, 440 values for <i>E</i> and <i>MOR</i> and 279 values for <i>DM</i> were compiled to develop robust and reliable predictive models. Multiple statistical metrics were utilized to validate the predictions of the model. The GEP model exhibited excellent prediction accuracy with a correlation coefficient (R) of 0.95, 0.96, and 0.94 for <i>MOR</i>,<i> E</i>, and <i>DM</i>, respectively. The comparison between GEP and linear regression models revealed revealed that machine learning-based models (GEP) exhibit superior accuracy than traditional regression models. Moreover, SHapley Additive exPlanations (SHAP) revealed that age significantly influenced all three considered strength characteristics of concrete. In addition, a graphical user interface (GUI) has been created to simplify the utilization of ML-based models for predicting the strength characteristics of concrete. The outcome of this study has the potential to be utilized for smartly predicting concrete’s strength properties, eliminating the need for expensive and time-consuming testing procedures.</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

Development of prediction models for strength properties of concrete using gene expression programming

  • 期刊论文

  • 发表日期：

    2025年2月24日

查看全文
  • Asad Ullah Khan,
  • Muhammad Faisal Javed,
  • Majid Khan

摘要

Accurately forecasting the strength properties of concrete is critical, especially for design code compliance. While statistical and empirical models are commonly used; however, these methods demand intensive experimentation and may yield inaccuracies in intricate relationships between concrete characteristics, mix components, and curing conditions. To address these limitations, the current study utilized the gene expression programming (GEP) method to forecast the mechanical characteristics of concrete, such as modulus of elasticity (E), modulus of rupture (MOR), and dynamic modulus (DM). For this purpose, 440 values for E and MOR and 279 values for DM were compiled to develop robust and reliable predictive models. Multiple statistical metrics were utilized to validate the predictions of the model. The GEP model exhibited excellent prediction accuracy with a correlation coefficient (R) of 0.95, 0.96, and 0.94 for MOR, E, and DM, respectively. The comparison between GEP and linear regression models revealed revealed that machine learning-based models (GEP) exhibit superior accuracy than traditional regression models. Moreover, SHapley Additive exPlanations (SHAP) revealed that age significantly influenced all three considered strength characteristics of concrete. In addition, a graphical user interface (GUI) has been created to simplify the utilization of ML-based models for predicting the strength characteristics of concrete. The outcome of this study has the potential to be utilized for smartly predicting concrete’s strength properties, eliminating the need for expensive and time-consuming testing procedures.