Prediction of biomass and nutrients removal behavior in membrane photobioreactor through using computational AI algorithms
Abstract
When microalgae are used to clean wastewater, the nutrients in the wastewater are eliminated, and the end result is microalgae biomass. Microalgae are used in wastewater treatment to simultaneously remove nutrients and produce microalgae biomass. Microalgae significantly alleviate pollution by removing nutrients, by providing microalgal biomass as aquaculture feed. Moreover, a revolutionary technique that combines membranes with photobioreactors (PBR) is the membrane photo bioreactor (MPBR). The goal of the present study was to compare Response surface methodology (RSM) and Artificial neural networks (ANN) in order to predict, forecast, and optimize the behavior of dry biomass, dissolved inorganic nitrogen (DIN), and dissolved inorganic phosphorus (DIP) in MPBR. The RSM was used to identify the optimum operating conditions. Hydraulic retention time (HRT) and cultivation time were the independent variables employed in the modelling. The HRT values were set at 1, 2, 4, and 6 days. Also, cultivation time values range from 1 to 19 days for municipal wastewater in China. The ANOVA results of the RSM model for dry biomass, DIN, and DIP showed good agreement with empirical data by obtaining F-value of 543.97, 160.34, and 47.47, respectively. According to RSM optimization results, the optimum mode for DIN and DIP achieved at an HRT value of 1.15 days and a Cultivation time value of 1.92 days. The margin of variation was less than 10%, indicating that the ANN model performed better than the RSM model. Additionally, the findings demonstrated that the ANN algorithm is more accurate than the RSM technique at predicting the behavior of DIN, DIP, and dry biomass in MPBR.