Scholarlink

<p>Mesoscale FE modelling of concrete requires aggregate assemblies that rigorously conform to prescribed gradation curves. This study develops a strength-grade-adaptive aggregate generation framework under code-specified gradation constraints via a coupled multientropy–multifractal optimisation scheme. Shannon, Rényi, and Tsallis entropies are embedded into differential evolution to jointly control size distribution, shape descriptors, and packing configuration, while multifractal optimisation is imposed against target gradation curves to ensure multi-scale consistency. Target gradation sets for C30–C60 are constructed as benchmarks. Benchmark tests and parametric studies show consistently reduced gradation fitting errors compared with conventional approaches, alongside an ordered hierarchy of multifractal dimensions across the target gradation sets. Across 10 statistically independent realizations per grade (40 total), the simulations consistently satisfied the JGJ 55-2011 compliance criteria, with a gradation RMSE coefficient of variation below 3%. In addition, three replicate physical blended sieve tests were conducted for each strength grade, and the measured composite gradation curves were found to lie within the JGJ 55-2011 envelope. FE simulations demonstrate geometric compatibility of the generated assemblies with established mesoscale frameworks; fine aggregates maintain consistently near-spherical morphologies (87.1–87.5% with <i>ψ</i> &gt; 0.95), while coarse aggregates retain appropriate irregularity. The proposed method provides a reproducible optimisation-based tool for generating code-consistent aggregate assemblies for refined mesoscale concrete FE modelling.</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

Strength-grade-adaptive aggregate generation via multientropy-multifractal optimization for refined concrete FE modelling

  • 期刊论文

  • 发表日期：

    2026年6月15日

查看全文
  • Kangjie Chen,
  • Yan-Gang Zhao,
  • Si-Qi Lin

摘要

Mesoscale FE modelling of concrete requires aggregate assemblies that rigorously conform to prescribed gradation curves. This study develops a strength-grade-adaptive aggregate generation framework under code-specified gradation constraints via a coupled multientropy–multifractal optimisation scheme. Shannon, Rényi, and Tsallis entropies are embedded into differential evolution to jointly control size distribution, shape descriptors, and packing configuration, while multifractal optimisation is imposed against target gradation curves to ensure multi-scale consistency. Target gradation sets for C30–C60 are constructed as benchmarks. Benchmark tests and parametric studies show consistently reduced gradation fitting errors compared with conventional approaches, alongside an ordered hierarchy of multifractal dimensions across the target gradation sets. Across 10 statistically independent realizations per grade (40 total), the simulations consistently satisfied the JGJ 55-2011 compliance criteria, with a gradation RMSE coefficient of variation below 3%. In addition, three replicate physical blended sieve tests were conducted for each strength grade, and the measured composite gradation curves were found to lie within the JGJ 55-2011 envelope. FE simulations demonstrate geometric compatibility of the generated assemblies with established mesoscale frameworks; fine aggregates maintain consistently near-spherical morphologies (87.1–87.5% with ψ > 0.95), while coarse aggregates retain appropriate irregularity. The proposed method provides a reproducible optimisation-based tool for generating code-consistent aggregate assemblies for refined mesoscale concrete FE modelling.