Non-destructive method of identifying suspended foulants in an anaerobic nanofiltration process using filter pad analysis

ABSTRACT Membrane autopsies provide valuable insight about foulant composition but require operation downtime, are destructive, costly, and infrequently conducted, limiting their practicality for routine monitoring. This study collected over 1.5 years of operation data to assess process performance and employed a non-destructive approach to identify foulants and their sizes throughout a nanofiltration process. Results were compared to historical autopsies from 2011 to 2023. Solids collected from 22-mm, 0.45-μm silver pads used to filter feed, interstage, and concentrate streams were analyzed using scanning electron microscopy, energy dispersive X-ray spectroscopy, and optical photothermal infrared spectroscopy. Unlike autopsies, this approach enables in-situ foulant characterization without destroying the membranes. Bayesian analysis was used to correlate the identified foulants with performance decline. Similar foulants were identified using the non-destructive method and the autopsies: organics, pyrite, colloidal aluminosilicates, calcium-, and iron-based compounds. Particle size data showed colloidal aluminosilicates formed via agglomeration within the membrane feed-concentrate channel. Bayesian analysis linked performance decline in the second stage to colloidal aluminosilicates and iron-related foulants. This research successfully demonstrated that the non-destructive technique can reliably identify foulants, enabling earlier detection, reducing downtime, and providing a predictive tool for diagnosing performance decline using Bayesian analysis.