Abstract Differences in the effective thermal conductivity (ETC) between measurements and models may be attributed to the limited ability to capture microstructural information of geomaterials. Today, computed tomography (CT) technology offers unprecedented access to such information, particularly for sands. Since a sand can be represented as a contact network made of nodes (particles) connected by edges (contacts), features (or variables) arising from the contact network can characterise particle connectivity. However, existing contact network features neglect the contribution of contact quality and of small gaps between neighbouring particles to heat transfer. To redress these issues, this paper introduces new weighted contact network features by considering contact area at each edge in the contact network. Additionally, thermal network features are proposed by considering small gaps as edges with/without being weighted by thermal conductance. All network features are calculated based on CT images of five real sands. The relationships between each feature and ETC are investigated. The results show that some network features that account for both the particle connectivity and contact quality can be used to predict ETC accurately. Advantages and limitations of this approach are also identified.

中文翻译：

---

砂中传热网络分析

摘要 测量值和模型之间有效热导率 (ETC) 的差异可能归因于捕获地质材料微观结构信息的能力有限。今天，计算机断层扫描 (CT) 技术提供了对此类信息的前所未有的访问，特别是对于沙子。由于沙子可以表示为由边缘（接触）连接的节点（粒子）组成的接触网络，因此接触网络产生的特征（或变量）可以表征粒子的连通性。然而，现有的接触网络特征忽略了接触质量和相邻颗粒之间的小间隙对传热的贡献。为了解决这些问题，本文通过考虑接触网络中每个边缘的接触区域，引入了新的加权接触网络特征。此外，热网络特征是通过将小间隙视为边缘而提出的，有/没有热导加权。所有网络特征均基于五个实砂的 CT 图像计算得出。研究了每个特征与 ETC 之间的关系。结果表明，一些同时考虑粒子连通性和接触质量的网络特征可用于准确预测 ETC。还确定了这种方法的优点和局限性。结果表明，一些同时考虑粒子连通性和接触质量的网络特征可用于准确预测 ETC。还确定了这种方法的优点和局限性。结果表明，一些同时考虑粒子连通性和接触质量的网络特征可用于准确预测 ETC。还确定了这种方法的优点和局限性。