Prediction and optimization of biodiesel production by using ANN and RSM

Abstract

This experimental work examined the prediction and optimization of biodiesel production from pomegranate seed oil using Artificial Neural Networks (ANN) and Response Surface Methodology (RSM) with central composite design and The transesterification method chosen for biodiesel production. The Central Composite Design (CCD) optimization conditions were methanol/oil molar ratio (3:1 to 11:1), catalyst rate (0.5 wt% to 1.50 wt%), temperature (50 ℃ to 70 ℃) and time (45 min to 105 min). The process factors were optimized by using CCD based on the RSM method and developed an ANN model to predict biodiesel yield. The optimum yield was found 95.68% with optimum process parameters as 8.01:1 methanol/oil molar ratio, 1.08 wt% catalyst rate, 70 ℃ temperature and 45 min time. The coefficient of determination (R2) acquired from the response surface methodology model is 0.9887 and is better when compared to the coefficient of determination (R2) of 0.9691 acquired from the Artificial neural network model. According to the results, using RSM and ANN models is beneficial for optimizing and predicting the biodiesel production process.