Data-informed modelling of aerosol resuspension under aerodynamic loads

将以结构化和系统化的方式建立测试颗粒和表面的实验数据，最终扩展到与真实的悬浮问题相关的复杂表面和颗粒。一个广泛的尺寸范围将被认为是环境颗粒，存款表面拓扑结构的影响和随时间变化的流动效果将被探讨。影响再悬浮过程的关键参数及其与其他变量的相关性将从精心控制的实验中确定，数据将用于与选定的一组数值模拟进行交叉验证。随后，数值研究可以用来阐明完整的再悬浮过程，而且可以扩展到不容易复制的实验方案。机器学习技术将用于从数据中学习，并根据获得的实验和计算数据开发简化模型。

Experimental data for test particles and surfaces will be built in a structured and systematic manner, eventually extending to complex surfaces and particles relevant to real suspension problems. A broad range of sizes will be considered up to environmental particles, the effects of deposit surface topology and time-varying flow effects will be explored. Key parameters influencing the resuspension process and their correlation to other variables will be identified from the carefully controlled experiments and the data will be used to cross-validate with a selected set of numerical simulations. Subsequently, the numerical studies can be used to elucidate the complete resuspension process and moreover be extended to scenarios not easily replicable by the experiments. Machine learning techniques will be employed to learn from the data and develop simplified models, based on the experimental and computational data obtained.