Global sensitivity analysis of a filtration model for submerged anaerobic membrane bioreactors (AnMBR)

• A global sensitivity analysis of a filtration model for AnMBRs was conducted.
• It was conducted using a revised version of the Morris screening method.
• Sensitivity results significantly simplified the input subset to be calibrated.
• Only one input factor could be removed from the model.

The results of a global sensitivity analysis of a filtration model for submerged anaerobic MBRs (AnMBRs) are assessed in this paper. This study aimed to (1) identify the less- (or non-) influential factors of the model in order to facilitate model calibration and (2) validate the modelling approach (i.e. to determine the need for each of the proposed factors to be included in the model). The sensitivity analysis was conducted using a revised version of the Morris screening method. The dynamic simulations were conducted using long-term data obtained from an AnMBR plant fitted with industrial-scale hollow-fibre membranes. Of the 14 factors in the model, six were identified as influential, i.e. those calibrated using off-line protocols. A dynamic calibration (based on optimisation algorithms) of these influential factors was conducted. The resulting estimated model factors accurately predicted membrane performance.