Intelligent Optimal Design of a Membrane Bioreactor Based on Flow Field Quantitative Analysis

Abstract

The application of intelligent algorithms in the optimal design of a membrane bioreactor (MBR) is helpful in improving reactor performance. Our study constructed a numerical flow field model of a MBR. Membrane thickness, porosity, and inlet velocity were used as independent input variables. Flow field uniformity and turbulent kinetic energy were used as characteristic parameters to evaluate the flow field effect of the reactor. Based on the numerical calculation of samples, the back-propagation (BP) neural network model was used for prediction. Genetic algorithms (GA), artificial bee colony algorithms, and particle swarm optimization (PSO) algorithms were used to optimize the BP neural network. PSO–BP was screened out by error analysis as the best intelligent prediction algorithm. The function model between the reactor parameters and the flow field effect was derived by multivariate nonlinear regression analysis in combination with the results of computational fluid dynamics and PSO–BP prediction. The following optimal design parameters were determined using GA: membrane structure thickness of 45.6 mm, porosity of 76%, and inlet velocity of 2 V. A feasible intelligent optimal design method of MBR was established by combining fluid dynamics and modern intelligent algorithms to provide a new idea and method for the optimal design of an MBR.