PCA as tool for intelligent ultrafiltration for reverse osmosis seawater desalination pretreatment

• Published in: Desalination Vol. 419; pp. 188 - 196
• Main Authors: Naessens, W., Maere, T., Gilabert-Oriol, G., Garcia-Molina, V., Nopens, I.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2017
• Subjects: Desalination; Fouling; Principal component analysis; Seawater; Ultrafiltration; Fouling; Ultrafiltration; Desalination; Principal component analysis; Seawater
• ISSN: 0011-9164; 1873-4464
• DOI: 10.1016/j.desal.2017.06.018

A novel fouling monitoring methodology based on principal component analysis (PCA) has been validated using transmembrane pressure (TMP) data of a pilot-scale pressurized ultrafiltration (UF) system operated with seawater. The evolution of membrane fouling was investigated to determine its relation to the used cleaning strategy on the one hand and the quality of the raw seawater on the other hand. The developed models showed that in terms of cleaning efficiency there are no significant differences between the standard and optimized backwashing protocols that were employed. This confirms the hypothesis of being able to use the optimized operation in a sustainable manner and benefit from lower cleaning frequencies. In addition, it has been demonstrated that the use of PCA as a monitoring technique to detect abnormal fouling behaviour is a robust tool. By using PCA, decisions on cleaning sequences or frequencies could be taken dynamically instead of running the system with fixed cycles. •A pilot-scale UF unit as pretreatment for RO seawater desalination was studied.•As for MBR, PCA visually represents the current process state and detect outliers.•Backwash (BW) with RO brine has no influence on fouling behaviour.•Optimized BW settings don't change fouling behaviour and save permeate and downtime.•Small datasets give different models, long datasets depict trends instead of noise.