Data-Driven Modelling Of Low-Pressure Hybrid Membrane Filtration Using Multivariate Polynomial Regression

Abstract

Hybrid membrane filtration processes involve complex physical, chemical, and biological phenomena, thus their mechanistic modelling is overly challenging. In this study we use multivariate polynomials to model the fouling of an in-line flocculationâ¿¿submerged membrane filtration system. The performance of obtained models is comparable to that of artificial neural network (ANN) models, to suit the needs of process optimisation and plant control. Their additional advantages are rapid model construction, easy presentation, inspection, and use.