Modeling of NF/RO membrane fouling and flux decline using real-time observations
Jongkwan Park,Kwanho Jeong,Sangsoo Baek,Sanghun Park,Mayzonee Ligaray,Tzyy Haur Chong,Kyung Hwa Cho
DOI: https://doi.org/10.1016/j.memsci.2019.01.031
IF: 9.5
2019-04-01
Journal of Membrane Science
Abstract:Many numerical models for membrane filtration have been developed to explain and predict fouling mechanisms. The models can simulate flux decline, transmembrane pressure increase, and fouling thickness based on theoretical equations. However, the simulated fouling layer thicknesses have not been validated by in-situ observations on membrane surfaces because the membrane system is operated under a sealed and pressurized condition. In this study, humic acid fouling layers on nanofiltration and reverse osmosis membranes were monitored in-situ and in real-time using optical coherence tomography (OCT). The OCT system detected fouling layer growth over time, and showed that the compact and thick fouling layer had an estimated thickness of 80 µm. When comparing the thickness of the fouling layer between OCT and scanning electron microscopy (SEM) images, the OCT images showed values that were approximately 8 times higher than those of the SEM images. By comparing the obtained fouling thickness values with the estimated results, two existing models (Faridirad model and pore blockage-cake filtration model) were validated in terms of root mean square error (RMSE), Akaike Information Criteria (AIC), and coefficient of determination (R2). Both models showed similar high R2 (≥0.97) and low RMSE (<10-5) values, but the pore blockage-cake filtration model had lower AIC values than the Faridirad model. This study highlighted that the in-situ and real-time monitoring of fouling thickness can provide significant information for developing an accurate and precise membrane model.
engineering, chemical,polymer science