Effect of Scale-Up on Mass Transfer and Flow Patterns in Liquid–Liquid Flows Using Experiments and Computations
Effect of Scale-Up on Mass Transfer and Flow Patterns in Liquid–Liquid Flows Using Experiments and Computations
复制标题
DOI：
10.1021/acs.iecr.3c02284
复制
发表时间：
2023-09
期刊：
影响因子：
--
通讯作者：
Arnav Mittal;S. Bhattacharyya;Matthew Marino;Tai-Ying Chen;Pierre Desir;M. Ierapetritou;D. Vlachos-
中科院分区：
文献类型：
--
作者：
Arnav Mittal;S. Bhattacharyya;Matthew Marino;Tai-Ying Chen;Pierre Desir;M. Ierapetritou;D. Vlachos-
液-液微通道具有高传质速率但低通量。为了提高生产率，可以通过增加直径来扩大规模。因此，预测流型和传质速率，同时占溶剂的影响，直径变化是至关重要的，然而，这个话题是目前缺乏文献。我们开发了随机森林和符号遗传回归机器学习（ML）模型来预测流型和传质速率，分别使用我们的实验和计算流体动力学（CFD）数据和文献挖掘数据的组合，同时考虑溶剂性质和通道直径的影响。这使得能够快速预测微通道到毫通道的有效放大。为了最大限度地减少CFD模拟的数量并最大限度地提高模型精度，我们采用了主动学习技术。此外，我们量化了基于混合数据构建的ML模型的不确定性。
Liquid–liquid microchannels have high mass transfer rates but low throughput. To increase productivity, they can be scaled up by increasing the diameter. Therefore, predicting flow patterns and mass transfer rates while accounting for solvent effects as the diameter varies is crucial; however, this topic is currently lacking in the literature. We develop random forest and symbolic genetic regression machine learning (ML) models to predict flow patterns and the mass transfer rate, respectively, using a combination of our experimental and computational fluid dynamics (CFD) data and literature-mined data, while accounting for the effects of solvent properties and channel diameter. This enables rapid prediction for efficient scale-up of microchannels to millichannels. To minimize the number of CFD simulations and maximize the model accuracy, we employ active learning techniques. Furthermore, we quantify the uncertainty of the ML models built on the hybrid data.