Treatment of carpet and textile industry effluents using Diplosphaera mucosa VSPA: A multiple input optimisation study using artificial neural network-genetic algorithms

• Published in: Bioresource technology Vol. 387; p. 129619
• Main Authors: Singh, Virendra, Mehra, Ravi, Ramesh, Kirtan Babu, Srivastava, Pradeep, Mishra, Abha
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.11.2023
• Subjects: algorithms; Artificial neural network; Biomass production; Bioremediation; Genetic algorithm; hybridization; light intensity; Microalgae; mucosa; neural networks; pollutants; Python; Response surface methodology; temperature; textile industry; wastewater; wastewater treatment; Microalgae; Bioremediation; Response surface methodology; Artificial neural network; Biomass production; Genetic algorithm
• ISSN: 0960-8524; 1873-2976; 1873-2976
• DOI: 10.1016/j.biortech.2023.129619

[Display omitted] •Global optimisation of industrial effluent treatment by Diplosphaera mucosa.•Hybridisation of RSM and ANN models with GA for multi-input optimisation.•Application of both MATLAB and Python for ANN model construction.•Two-way interaction of pH and N/P ratio has significant effect on algae growth.•ANN models generated by Python showed better accuracy. The wastewater treatment efficiency of Diplosphaera mucosa VSPA was enhanced by optimising five input parameters and increasing the biomass yield. pH, temperature, light intensity, wastewater percentage (pollutant concentration), and N/P ratio were optimised, and their effects were studied. Two competitive techniques, response surface methodology (RSM) and artificial neural network (ANN), were applied for constructing predictive models using experimental data generated according to central composite design. Both MATLAB and Python were used for constructing ANN models. ANN models predicted the experimental data with high accuracy and less error than RSM models. Generated models were hybridised with a genetic algorithm (GA) to determine the optimised values of input parameters leading to high biomass productivity. ANN-GA hybridisation approach performed in Python presented optimisation results with less error (0.45%), which were 7.8 pH, 28.8 °C temperature, 105.20 μmol m−2 s−1 light intensity, 93.10 wastewater % (COD) and 23.5 N/P ratio.