Modeling fouling of hollow fiber membrane using response surface methodology

Hollow fiber membranes have been widely employed for potable water treatment, wastewater treatment and reclamation, and desalination pretreatment. They have advantages such as small footprint, ease of operation, and high removal of particles and pathogens. Nevertheless, membrane fouling is one of the most serious issues in operating hollow fiber membrane systems. Although there have been numerous researches for understanding and forecasting fouling, it is still challenging to predict fouling based on explicit mathematical models. A more practical approach is to apply statistical models, which allows more accurate fouling prediction. In this context, this study attempts the development of a statistical model to analyze fouling of hollow fiber membranes. A response surface experimental design was used to optimize and investigate the influence of process variables such as foulant types (kaolin, colloidal silica, NOM, and alginate), foulant concentration, and imposed flux on the fouling rate. The results obtained from the experiments were evaluated by multiple regression analysis method and empirical relationship between the response and independent variables. Empirical models were developed to understand the interactive correlation between the responses and process variables.