Building a simple multivariable filtration model to predict irreversible fouling when directly filtering municipal wastewater

Abstract

This paper describes a simple generic model designed to predict membrane fouling in municipal wastewater (MWW) treatment. The work was conducted using data from a direct membrane filtration demo-system (middle/long-term filtration periods of about 35 – 124 days) to calibrate the model. Two influents were treated by the demo-system: raw pre-treated MWW and primary settler supernatant from a full-scale MWW treatment plant. A resistance-in-series mathematical model structure was proposed considering fouling due to two different mechanisms: persistent cake layer formation (from suspended material) and pore blocking (from soluble and colloidal compounds). The proposed model represented transmembrane pressure dynamics at different operating solids concentrations (around 1, 2.6, 6 and 11 gL−1) using 7 model parameters, achieving 7–28 mbar differences between the experimental data and model predictions in all cases (calculated as the root mean square error). The model was also able to match the results from two different influents (raw MWW and primary settler supernatant) by modifying 3 of the 7 parameters while low uncertainties were obtained in long-term filtrations, demonstrating its robustness. This model thus provides a good potential to generate reasonable membrane fouling predictions while its simple and open structure makes it easy to implement with complementary materials. Further research will be carried out to enhance the model’s precision and validate its potential for optimizing filtration and fouling control processes.