微塑料
分离(统计)
环境科学
环境化学
化学
废物管理
色谱法
工程类
计算机科学
机器学习
作者
Te Bu,Diego Mesa,Arjun Kumar Pukkella,Pablo R. Brito‐Parada
标识
DOI:10.1016/j.cej.2024.153718
摘要
Microplastics, particularly those in the 5–20μm range, pose substantial risks in aquatic environments, yet effective removal techniques are limited. This study assesses two critical geometric parameters of a 10 mm mini-hydrocyclone: inlet radius and insertion angle. Along with the inlet velocity, they are optimised collectively to enhance the MPs' removal. Computational Fluid Dynamics (CFD) simulations using the Mixture model and Reynolds stress model were employed to track the particle-water-air interactions, revealing the flow dynamics and absence of an air core. The Response Surface Methodology (RSM) using a Box-Behnken design of experiments was adopted to assess the impact of the geometric and operational parameters and determine the optimal design. Recovery efficiency, concentration efficiency, and split rate were recorded during experiments and compared with CFD results, showing a good agreement between experiments and CFD simulations. The resulting novel design achieved an 88.53% MPs recovery and a concentration efficiency of 1.83, which represents an obvious improvement over the commercial design's MPs recovery efficiency of 79.97% and concentration efficiency of 1.68. Simulations reveal that the novel design shortens the path for MPs through the forced turbulence region and modifies the velocity and turbulence fields, enhancing the concentration gradient of MPs within the mini-hydrocyclone's chamber. Moreover, the novel design exhibits a more robust response to inlet velocity variations, a crucial advantage for its application in arrays. Experiments further confirm that a matrix composed of 10 novel mini-hydrocyclones achieves 10 percentage points higher recovery efficiency than the matrix with 10 commercial mini-hydrocyclones. These findings mark a significant step towards effective MP removal, addressing a critical environmental issue.
科研通智能强力驱动
Strongly Powered by AbleSci AI