Neem-castor seed oil esterification modelling: Comparison of RSM and ANFIS

Abstract

ven though the desirable fuel properties and enhanced engine characteristics of an IC engine have led to the adoption of hybrid second generational oily feedstock (HSGOF) over a mono-oily feedstock, the high FFA associated with the HSGOF remains a challenge. The Response Surface Methodology (RSM) and Adaptive-Network-based Fuzzy Inference System (ANFIS) were used to model the esterification of neem and seed crude oils (NCSO) with a high free fatty acid content (FFA). The RSM and ANFIS were used to model and optimize input: esterification variables such as NCSO ratio (20–40), catalyst dosage (1.0– 2.0 wt%), retention time (1.5–2.5 h), and NCSO/methanol molar ratio (4.5–5.5) to achieve the lowest FFA for the NCSO esterification process and produce composite biodiesel. Optimum acid value (AV) of 2.1 mg KOH/g was attained at NCSO ratio of 20, catalyst dosage of 2 wt%, retention time of 2.0 h, and molar ratio of 4.5. A contrast of the developed models was accomplished by statistical norms such as coefficient of determination (R2 ), root mean square error (RMSE) with standard Error of Prediction (SEP). The R2 of 0.82522, RMSE of 0.68099, and SEP of 0.21729 for the RSM model (RM) related to the R2 of 0.9982, RMSE of 0.02434, and 0.01992 for the ANFIS model (AM). The AM seems more consistent than the RM in pre-treating and rendering high NCSO for composite biodiesel production. The key fuel properties of the neem-castor oil methyl ester (NCSOME) certified with those of biodiesel international standards. Copyright 2023 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Confer- ence & Exposition on Mechanical, Material and Manufacturing Technology.