Transesterification of castor oil: neuro-fuzzy modelling, uncertainty quantification and optimization study

This study conducted experiments used for the development of both regression model with uncertainty analysis and adaptive neuro-fuzzy inference system (ANFIS) prediction yield biodiesel (YB) produced from castor oil in presence calcium oxide derived eggshell. Box Behnken design (BBD) was to develop experimental condition five different variables while YB response. Uncertainty determined Monte Carlo simulation (MCS). The optimized validated before generated data applied three ANFIS modelling techniques. Root mean square error (RMSE), coefficient correlation (R2) average percentage (APE) were determine accuracy models developed. result this shows that optimum (94.29%) achieved at a methanol ratio 11.48, catalyst loading 3.38 wt%, reaction time 1.84 h, temperature 60.2 °C, agitation 343.5 rpm. BBD, MCS agreed most important parameter investigation. considered technique (subtractive clustering) outperformed BBD model. novelty are determination transesterification thermally treated anthill, establishment use YB, influence on using statistical and, AI techniques validation predictions two methods MCS.