A quasi-Monte Carlo based flocculation model for fine-grained cohesive sediments in aquatic environments

The quasi-Monte Carlo (QMC) method was enhanced to solve the population balance model (PBM) including aggregation and fragmentation processes for simulating the temporal evolutions of characteristic sizes and floc size distributions (FSDs) of cohesive sediments. Ideal cases with analytical solutions were firstly adopted to validate this QMC model to illustrate selected pure aggregation, pure fragmentation, and combined aggregation and fragmentation systems. Two available laboratory data sets, one with suspended kaolinite and the other with a mixture of kaolinite and montmorillonite, were further used to monitor the FSDs of cohesive sediments in controlled shear conditions. The model results show reasonable agreements with both analytical solutions and laboratory experiments. Moreover, different QMC schemes were tested and compared with the standard Monte Carlo scheme and a Latin Hypercube Sampling scheme to optimize the model performance. It shows that all QMC schemes perform better in both accuracy and time consumption than standard Monte Carlo scheme. In particular, compared with other schemes, the QMC scheme using Halton sequence requires the least particle numbers in the simulated system to reach reasonable accuracy. In the sensitivity tests, we also show that the fractal dimension and the fragmentation distribution function have large impacts on the predicted FSDs. This study indicates a great advance in employing QMC schemes to solve PBM for simulating the flocculation of cohesive sediments. (C) 2021 Elsevier Ltd. All rights reserved.

Automated summary

The study demonstrates significant progress in using QMC schemes to solve PBM for simulating the flocculation of cohesive sediments, showing reasonable agreements with both analytical solutions and laboratory experiments. Different QMC schemes outperform the standard Monte Carlo scheme in accuracy and time consumption, with the QMC scheme using Halton sequence requiring the least particle numbers for reasonable accuracy. Sensitivity tests indicate that the fractal dimension and fragmentation distribution function have significant impacts on predicted FSDs.