This work aims at the development of methods for fingerprinting filtration processes. These fingerprints can potentially be used to optimize the filtration performance of large scale dead-end hollow fiber ultrafiltration systems that are used nowadays in the production of drinking water. The developed methods can be split into two major parts: Filtration process monitoring and feedwater characterization. Filtration process monitoring is performed with the help of a miniturized (to bench scale) version of the large filtration plant. The small plant should contain similar membranes (material and configuration) as its large-scale counterpart. With this small setup, critical flux measurements can be performed. The value for the critical flux then serves as a performance indicator of the used feed solution with the applied membrane. With several model feed solutions, the applicability of the method was shown. Feedwater characterization focused mainly on the determination of particle size distributions with the help of flow-field flow fractionation coupled to multi angle light-scattering. The use of light scattering enabled not only the determination of particle size, but also the confrontation of the polymers and their interaction with the membrane. These tools are applied in the thesis for an extensive review of the filtration of sodium alginates. The filtration inspection tool showed varying filtration data for the sodium alginate, depending on the ionic environment. Qualitatively, filtration models were setup, which could be informed by flow-field flow fractionation analysis of the alginates. Using a second model component, a soil-derived humic acid, we found that there is a distinct influence of the axial coordinate on the deposition in dead-end ultrafiltration. Most of the material deposits in the last part of the hollow fibers. Based on these observations, a partial backwash concept is decribed. This concept proposes to break up a filtration module in two, where one part is used for filtration and never backwashed, while the other module is allowed to foul and is regularly backwashed. The method was tested for various feed solutions. For the method to be successful, the material has to have a similar charge as the membrane and needs to have a specific size.
|Award date||24 Apr 2018|
|Place of Publication||Enschede|
|Publication status||Published - 24 Apr 2008|