Title: Biodiesel production from third-generation feedstock: process parameter modelling and optimisation using RSM-ANN approach
Authors: Aqueel Ahmad; Ashok Kumar Yadav; Shifa Hasan
Addresses: Department of Mechanical Engineering, Graphic Era (Deemed to be University), Dehradun, 248002, India ' Department of Mechanical Engineering, Raj Kumar Goel Institute of Technology, Ghaziabad, Uttar Pradesh 201003, India ' Department of Management Studies, Netaji Subhas University of Technology, Dwarka Sector 3, New Delhi 110078, India
Abstract: This study employed response surface methodology (RSM) coupled with a central composite design (CCD) approach to ascertain the optimal conditions for biodiesel production from Neochloris oleoabundans microalgae oil. Four key process variables, including the methanol-to-oil molar ratio, catalyst concentration, reaction time, and temperature, were investigated across five levels to develop an L30 orthogonal array for experimentation. An artificial neural network (ANN)-based prediction model was developed using the experimentally obtained data, yielding high accuracy with mean square error (MSE) values of 0.019, 2.4327, and 0.8269 and coefficient of determination (R2) values of 0.9996, 0.9796, and 0.9890 for training, validation, and testing sets, respectively, indicating robust predictive capability. The optimisation analysis reveals a biodiesel yield of 94.94% under optimised conditions: 6.92:1 molar ratio, 1.22% catalyst concentration, 64.36 min reaction time, and 56.46°C temperature. Experimental validation confirmed the reliability of the optimisation results, demonstrating a marginal error of 2%. [Received: November 25, 2022; Accepted: April 30, 2024]
Keywords: biodiesel production; Neochloris oleoabundans; response surface methodology; RSM; artificial neural network; ANN; sustainable fuel; central composite design; CCD; mean square error; MSE.
DOI: 10.1504/IJOGCT.2025.144530
International Journal of Oil, Gas and Coal Technology, 2025 Vol.37 No.2, pp.235 - 254
Accepted: 30 Apr 2024
Published online: 18 Feb 2025 *