Optimisation of Biodiesel Production from Waste Frying Oil Via Response Surface Methodology (RSM) and Artificial Neural Network (ANN)

This work was aimed at determining the suitability of using Waste Frying Oil (WFO) to produce biodiesel which is a very much needed energy product in our world today. This was achieved through a transesterification process in which time, methanol/oil ratio (MOR), catalyst weight and temperature were used to optimize the yield of biodiesel from waste frying oil. RSM and ANN have been used to carry out this process. RSM had an R2 value of 0.967 and ANN had an R2-value of 0.9000. This result does not suggest that ANN is less a predictive tool but this could be as a result of the number of neurons used or the number of training the ANN model was subjected to. The RSM model generated was optimized and the optimum yield obtained after 100 iterations was 88.694% at optimum conditions of MOR (7.720), Catalyst weight (1.393), Temperature (59.567), and Reaction time (92.707). The GC-MS results validates the biodiesel produced from WFO and shows that the biodiesel is typical of what is obtainable from similar research.