Low-pressure membrane filtration systems, such as microfiltration (MF) and ultrafiltration (UF), have received a great deal of attention in the past 15 years due to their ability to remove microbial pathogens, especially Cryptosporidium and Giardia. The major concern for the application of membrane technology is, however, how to ensure integrity of these barriers, since small defects in membranes could result in a significant reduction in pathogen removal efficiency. In order to ensure safe drinking water treatment, a number of environmental agencies request the membrane operators to conduct regular direct integrity tests to control the microbial log removal values (LRV) of the plants. Typically, test conditions must be selected to provide information on defects larger than 3 μm to ensure Cryptosporidium removal. In that context, the objective of this project was to develop and validate, both at bench-scale and full-scale, a model based on the equations proposed by USEPA and ASTM that uses the air flow rate throughout a defect during the air pressure test for predicting the microbial LRV. The project was conducted on a pressurised low-pressure membrane module. MS2-phages were used at bench-scale to validate the model and the selected assumptions with various calibrated defects carried out on the membrane fibres. The validity of the model was then evaluated at full-scale. A user-friendly tool using the Hagen Poiseuille Model proposed by the ASTM was developed to assist membrane operators in the integrity monitoring. The calibration of the model with full-scale tests resulted in adjusting some key-parameters representing air diffusion, flow regime and particles deposition. The numerical applications provide a very reliable result in predicting the pathogen removal efficiency and the equivalent number of broken fibres. This model could detect one complete broken fibre out of more than 700,000 fibres, which guaranteed more than 4 log of microorganism removal efficiency.

You do not currently have access to this content.