GOALI: Predicting performance and fouling of membrane filters

膜过滤器本质上是一种薄薄的多孔介质，用来去除通过介质的流体进料中的某些悬浮颗粒。膜过滤器在工业上得到了广泛的应用，仅在美国就代表着一个价值数十亿美元的行业。大型跨国公司，如W.L.Gore&Amp；Associates和Pall Corporation，生产大量基于膜的过滤产品，并对改进和优化他们的过滤器有着浓厚的兴趣。膜过滤的应用非常广泛，如净水、处理放射性污泥、生物技术工业中的各种净化工艺、净化空气或其他气体以及澄清啤酒。虽然基础应用可能差异很大(气体过滤与液体过滤；小颗粒与大颗粒去除；慢速与快速处理；刚性颗粒与可变形颗粒)，但广泛的工程挑战是相同的：以低功耗实现精细可控的分离。但是，膜特性(以及过滤器的行为和性能)在过滤器的使用寿命内远不是恒定的：从进料中去除的颗粒沉积在过滤器内部和过滤器上，污染过滤器并降低其性能。发生这种污染的过程是复杂的，并强烈依赖于几个因素，包括膜的内部结构、料液的流动动力学和料液中颗粒的类型(它们的形状、大小和化学作用影响膜如何去除它们)。在这个项目中，开发了新的数学模型来模拟膜过滤和污染，允许膜设计参数进行调整，以实现最佳性能。特别注意内部膜形态细节所起的作用，以便制定详细的设计指南。与工业伙伴(Pall公司的Anil Kumar博士)的合作最大限度地增加了我们的理论成果转化为工业实践的机会。研究生都参与了这个项目的工作。本项目研究膜过滤器中的流动和污染，这对工业应用具有重要的意义。与一名博士生合作，PI和她的合作者建立了新的预测数学模型，描述了两种实际重要的情况：(I)折叠滤筒内的流动和污染，以及(Ii)内部非均质膜的膜污染模型。在每种情况下，研究小组都建立了模型，这些模型可以解释进料溶液中的任意颗粒大小分布，也可以解释膜孔大小的分布。对这些场景的第一原理理论研究对于对这些系统进行基础理论和实验研究的其他人以及那些寻求扩大当前应用范围和改进制造工艺的人来说是有意义的。波尔公司的一位工业同事阿尼尔·库马尔博士参与了这个项目。作为一名实验者，他与团队分享现有数据，根据需要生成新数据来测试模型，并担任行业顾问。这种互动确保项目保持重点，并确保与应用程序相关的问题得到识别和解决。实验数据有助于确定模型中未知参数的合适范围，并检验不确定的建模假设。

Membrane filters -- essentially, thin sheets of porous medium that act to remove certain particles suspended in a fluid feed that passes through the medium -- are in widespread industrial use, and represent a multi-billion dollar industry in the US alone. Major multinational companies, such as W.L. Gore & Associates and Pall Corporation, manufacture a huge range of membrane-based filtration products and have a keen interest in improving and optimizing their filters. Membrane filtration is used in applications as diverse as water purification; treatment of radioactive sludge; various purification processes in the biotech industry; the cleaning of air or other gases; and beer clarification. While the underlying applications may vary dramatically (gas versus liquid filtration; small versus large particle removal; slow versus fast throughput; rigid versus deformable particles), the broad engineering challenge is the same: to achieve finely-controlled separation at low power consumption. But membrane characteristics (and hence the filter's behavior and performance) are far from constant over the filter lifetime: the particles removed from the feed are deposited within and on the filter, fouling it and degrading its performance. The processes by which this fouling occurs are complex and depend strongly on several factors, including the internal structure of the membrane, the flow dynamics of the feed solution, and the type of particles in the feed (their shape, size, and chemistry affect how they are removed by the membrane). In this project new mathematical models are developed to simulate membrane filtration and fouling, allowing membrane design parameters to be tuned for optimal performance. Particular attention is paid to the role played by the details of the internal membrane morphology, so that detailed design guidelines can be formulated. Collaboration with an industrial partner (Dr. Anil Kumar of Pall Corporation) maximizes the chances that our theoretical results translate into industrial practice. Graduate students are involved in the work of the project. This project studies flow and fouling in membrane filters, which are of significant interest for industrial applications. Working with a PhD student, the PI and her collaborator formulate new predictive mathematical models that describe two situations of practical importance: (i) Flow and fouling within pleated filter cartridges, and (ii) Membrane fouling models for internally heterogeneous membranes. In each scenario the team builds models that account for an arbitrary particle size distribution within the feed solution, and also for a distribution of membrane pore sizes. First-principles theoretical studies of these scenarios are of interest to others carrying out fundamental theoretical and experimental research on such systems, as well as to those seeking to extend the scope of current applications and improve manufacturing processes. An industrial colleague, Dr. Anil Kumar of Pall Corporation, collaborates on the project. An experimentalist, he shares his existing data with the team, generates new data as needed to test the models, and acts as industrial advisor. This interaction ensures that the project remains focused and that questions relevant to applications are identified and addressed. The experimental data help to determine appropriate ranges for unknown parameters in the models, and to test uncertain modeling assumptions.