Optimum biodiesel production from waste vegetable oil using functionalized cow horn catalyst: A comparative evaluation of some expert systems

Abstract

This study investigated the production of biodiesel from waste vegetable oil using iron-doped solid catalyst produced from cow horn. The catalyst was characterised using scanning electron microscope (SEM) coupled with energy dispersive X-ray (EDX), X-ray diffraction (XRD), X-ray fluorescence (XRF), Fourier transform infrared (FTIR) spectroscopy, Brunauer, Emmett and Teller (BET) analysis and thermo-gravimetric analysis (TGA)/differential thermal analysis (DTA). The transesterification experiments were planned using a Box-Behnken design (BBD) while three expert systems namely response surface methodology (RSM), artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) were evaluated to assess their capacity to model and optimise the process. The results of characterisation of the catalyst indicated the suitability of the catalyst for the transesterification reaction which was attributed to the dominant elements which were calcium (Ca) and iron (Fe). The performance of the three optimisation tools was assessed using statistical indices and the results showed that ANFIS performed better than ANN and RSM in that order. This was demonstrated in the very high R2 value (R2 = 0.9999) and low error values (MSE = 0.0010, RMSE = 0.0059, SEP = 0.0074%, MAE = 0.0277 and AAD = 0.0325%). All three models predicted very high biodiesel yields (>98%) although ANFIS (yield = 99.30%) performed slightly better than ANN (yield = 99.10%). The properties of biodiesel produced at the optimised conditions were compared with ASTM D6751 and EN 14214 standards and they were found to be within the acceptable limits indicating suitability of the fuel.

Keywords

• Biodiesel
• Artificial intelligence
• Box-Behnken design
• Genetic algorithm
• ANOVA