In this work, homogenisation of textile composite models is studied in order to increase the accuracy of wave dispersion predictions in such complex structures. A multiscale methodology is formed involving (i) calculation of the ultrasonic wave propagation through a detailed mesoscale finite element model of the textile and (ii) updating the mechanical properties of a semi‐analytical finite element (SAFE) model to match the accurate mesoscale predictions. The speeds of the first shear (SH₀) and Lamb wave modes (A₀ and S₀) are used to define the objective function to minimise in an inversion process based on genetic algorithms. The algorithm is tested on an orthotropic plate structure whose all nine unknown elastic moduli are successfully reconstructed and is then applied to three different numerical models. We demonstrate both numerically and experimentally that the proposed scheme provides more accurate dispersion characteristics compared to the ones obtained through the statically measured mechanical properties.

中文翻译：

---

多尺度波和有限元建模的超声兰姆波在织物复合材料中的均质化方案

在这项工作中，对纺织品复合材料模型的均质性进行了研究，以提高这种复杂结构中波浪色散预测的准确性。形成了多尺度方法，包括（i）通过详细的纺织品中尺度有限元模型计算超声波传播，以及（ii）更新半分析有限元（SAFE）模型的机械性能以匹配准确的中尺度预测。第一剪切速度（S H ₀）和兰姆波模式（A ₀和S ₀）用于定义目标函数，以在基于遗传算法的反演过程中将其最小化。该算法在正交各向异性板结构上进行了测试，该结构的所有九个未知弹性模量均已成功重构，然后应用于三个不同的数值模型。我们通过数值和实验证明，与通过静态测量的机械性能获得的色散特性相比，该方案可提供更准确的色散特性。