Modelling and enhancing the production of green diesel in biodiesel industries have been
hampered by the failure of the conventional approach to pursue space with continuous conver-
gence velocity, being entombed in local minima, and maintaining unwavering resolutions. The
study presented for the first time the optimization protocol for the development of biodiesel
production from tobacco seed oil (TSO) on the batch reactor aided by the unique Grey Wolf
Optimizer-Response Surface Methodology-Artificial Neural Network (GWO-RSM-ANN) tech-
niques. Lower calorific value (LCV), higher calorific value (HCV), and specific heat capacity (Cp)
correlations were postulated for tobacco seed oil methyl ester (TSOME/B100/TSOB) and diesel
blends. RSM, ANN, and GWO approaches were used to model TSOME’s main production yield.
The ASTM test methods were used to examine the significant basic properties of the fuel cate-
gories, while the LCV and HCV were detected using standard procedures. Maximum TSOME yield
(90.2%) was obtained at methanol/TSO molar ratio of 5.95, KOH content of 1.15 wt. %, and
methylic duration of 77.6 min. The ANN model configuration (3-15-1) that was developed
showed more adaptability and nonlinearity. The estimated coefficient of determination (R2) of
0.9999, mean average error (MAE) of 0.00035, and RMSE of 0.00105 for the GWO model
compared to those of R2 of 0.9825, MAE of 1.3145, and RMSE of 1.7087 for RSM model; and R2 of
0.9976, MAE of 0.2405, and RMSE of 0.6381 for ANN model vindicate the superiority of GWO
model over the RSM and ANN models. The major fuel properties agreed with the ranges of the
ASTMD6751 and EN 14214 specifications. The LCV, HCV, and Cp are also correlated with the
TSOME fraction through the linear equations. There were excellent correlations between the
analyzed and calculated values for the LCVs and HCVs. The maximum absolute error between the
measured and estimated LCV and HCV are 0.108% for 20%TSOME (20% TSOME +80% diesel
fuel), and 0.17% for pure diesel, respectively. The model and correlations can offer biodiesel and
automobile industries with database information
