Scholarlink

<p>This paper presents a comprehensive comparative analysis of different methods for determining the coefficient of consolidation (<i>c</i><sub><i>v</i></sub>), including the approach based on modelling the entire consolidation curve. The study was based on more than 420 datasets of paired <i>c</i><sub><i>v</i></sub> values from conducted laboratory experiments and archival research, covering seven methods. The primary objective was to assess the agreement of individual methods with computational approach and to identify any systematic differences. The analysis employed classical scatterplots, analysis of <i>c</i><sub><i>v</i></sub> ratios, and the difference approach with logarithmic transformation, including the determination of agreement limits using nonparametric and regression-based methods. The results showed that the Casagrande method (CM) exhibited the best agreement with model simulation (SYM) in terms of bias and slope of the limits of agreement, whereas the other methods systematically produced higher <i>c</i><sub><i>v</i></sub>. Based on the meta-analysis, a consistent trend of decreasing <i>c</i><sub><i>v</i></sub> values was observed in the order: TM &gt; IPM &gt; SRS1 &gt; ES &gt; HYP &gt; CM &gt; SYM. The obtained results emphasise the importance of using a logarithmic transformation and a nonparametric approach in assessing the compatibility of methods, especially in the context of heteroscedastic data.</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

Emerging trends in coefficient of consolidation determination: a statistical meta-analysis and nonparametric agreement assessment

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月8日

查看全文
  • Bartłomiej Szczepan Olek

摘要

This paper presents a comprehensive comparative analysis of different methods for determining the coefficient of consolidation (cv), including the approach based on modelling the entire consolidation curve. The study was based on more than 420 datasets of paired cv values from conducted laboratory experiments and archival research, covering seven methods. The primary objective was to assess the agreement of individual methods with computational approach and to identify any systematic differences. The analysis employed classical scatterplots, analysis of cv ratios, and the difference approach with logarithmic transformation, including the determination of agreement limits using nonparametric and regression-based methods. The results showed that the Casagrande method (CM) exhibited the best agreement with model simulation (SYM) in terms of bias and slope of the limits of agreement, whereas the other methods systematically produced higher cv. Based on the meta-analysis, a consistent trend of decreasing cv values was observed in the order: TM > IPM > SRS1 > ES > HYP > CM > SYM. The obtained results emphasise the importance of using a logarithmic transformation and a nonparametric approach in assessing the compatibility of methods, especially in the context of heteroscedastic data.