On-line monitoring of membrane fouling based on an improved electrical measurement method.

This article focuses on the development and evaluation of an on-line monitoring method for membrane fouling in hollow fiber membrane (HFM) filtration systems using Electrical Resistance Tomography (ERT) sensors combined with advanced algorithms. The method employs a Particle Swarm Optimization with Simulated Annealing (PSO-SA) algorithm to optimize finite element models, and a deep denoising extreme learning machine with an auto-encoder (DDELM-AE) alongside the K-singular value decomposition (K-SVD) algorithm for improved ERT image reconstruction. Experimental results demonstrate that ERT pixel variations correspond closely with the "two-stage" fouling phenomenon, enabling real-time detection of fouling stages, fouling distribution, and membrane integrity. Additionally, permeate flux is accurately measured using a lifting wavelet-based cross-correlation approach, validating the system's capability for effective, non-invasive, and early-stage fouling monitoring in water treatment processes.