Scholarlink

<p>Based on the existing flume-experimental records of meandering compound channels and explicitly accounting for the hydraulic influence of the floodplain vegetation, a machine-learning approach, namely genetic programming, was employed to derive a compact yet highly accurate discharge prediction equation. The resulting equation was compared against existing predictors and was subsequently interrogated, together with straight-compound-channel data, to quantify the individual influences of geometric and resistance parameters. Across the entire dataset for vegetated meandering compound channels, the new formula delivers a pronounced improvement in predictive accuracy, giving the smallest mean squared error and the substantially enhanced correlation coefficient. The sensitivity analysis identifies the main-channel sinuosity (<i>s</i>) and bed slope (<i>S</i><sub>0</sub>) as the dominant controls. The discharge decreases monotonically with increasing <i>s</i>, the width ratio of the main channel to the entire channel (<i>β</i>) and vegetation density (<i>φ</i>), whereas it increases with <i>S</i><sub>0</sub>. When <i>s</i> is low, the percentage reduction in discharge attributable to the increase in <i>s</i> intensifies with relative flow depth (<i>D</i><sub>r</sub>). Once <i>s</i> exceeds 1.5, the nondimensional reduction becomes essentially depth-independent.</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

Prediction of discharge capacity in meandering compound channels covered with vegetation

  • 期刊论文

  • 发表日期：

    2026年7月31日

查看全文
  • Yi-sen Wang,
  • Zhong-hua Yang,
  • Wen-xin Huai,
  • Meng-yang Liu

摘要

Based on the existing flume-experimental records of meandering compound channels and explicitly accounting for the hydraulic influence of the floodplain vegetation, a machine-learning approach, namely genetic programming, was employed to derive a compact yet highly accurate discharge prediction equation. The resulting equation was compared against existing predictors and was subsequently interrogated, together with straight-compound-channel data, to quantify the individual influences of geometric and resistance parameters. Across the entire dataset for vegetated meandering compound channels, the new formula delivers a pronounced improvement in predictive accuracy, giving the smallest mean squared error and the substantially enhanced correlation coefficient. The sensitivity analysis identifies the main-channel sinuosity (s) and bed slope (S0) as the dominant controls. The discharge decreases monotonically with increasing s, the width ratio of the main channel to the entire channel (β) and vegetation density (φ), whereas it increases with S0. When s is low, the percentage reduction in discharge attributable to the increase in s intensifies with relative flow depth (Dr). Once s exceeds 1.5, the nondimensional reduction becomes essentially depth-independent.