Optimisation of Biodiesel Production Using Taguchi Model

Abstract

In this study, an initiative has been taken to use Taguchi model for optimisation of reaction conditions in biodiesel production. Taguchi model helps to predict responses precisely for absent reaction combinations. Parboiled rice bran and white rice bran are considered as raw material for biodiesel production. Drying method, soxhlet temperature, sample pH, number of extraction cycle, transesterification, reaction temperature, catalyst concentration and solvent to sample ratio for fatty acid methyl ester (FAME) production were studied for lipid extraction. In case of lipid extraction, main effect plot for S/N ratio presents an optimised condition with given data. Taguchi model obtained an optimum lipid yield of 14.92 and 11.07% from parboiled and white rice bran respectively, at 55–60 °C soxhlet cell temperature, 20 number of extraction cycle, soxhlet heater temperature 100 °C, sun dried sample and pH 2, in a soxhlet extractor. During FAME production, surface and contour plots between transesterification parameters provide visualized information through graphical output for all possible responses. Taguchi model described accuracy of statistical approach for FAME yield with R2: 98.5% and R2 (adjusted): 93.7%. Taguchi optimised method resulted maximum FAME yield at transesterification temperature 60 °C, 12% (wt%) sulphuric acid, 12.5:1 methanol to lipid mass ratio (solvent/sample ratio) and initial pH 2 were 13.69 and 10.13% for parboiled rice bran and white rice bran, respectively. FAME conversion of 91.75% (dry weight basis) was achieved with Taguchi optimised parameters. The statistical method resulted an increase of 8.5 and 4.3% in lipid and FAME yield, respectively as compared to previously performed manual optimization. Taguchi method used for biodiesel production resulted in a residual error of 1.46%. Prepared rice bran methyl ester has improved biodiesel properties, i.e., cold filter plugging point, iodine value, flash point.