SOURCE:

Faculty: Engineering
Department: Chemical Engineering

CONTRIBUTORS:

Esonye, C.
Onukwuli. O.D.

ABSTRACT:

ABSTRACT


This research evaluated the transesterification of seed oils of Prunus amygdalus, Dyacrodes edulis and Chrysophyllum albidium as viable feedstocks for biodiesel using sodium hydroxide as catalyst. Solvent extraction method was used for the oil extraction and esterification was used to modify oil derived from Dyacrodes edulis. Oil and biodiesel samples characterizations were carried out using Association of Analytical Chemist (AOAC) and American Society for Testing and Materials (ASTM) respectively. Other characterization methods were Fourier transform infrared (FTIR) and gas chromatography-mass spectrophometry (GC-MS) techniques. Engine performance and emission characteristics of the methyl esters and their blends with petro-diesel were tested on Perkins 4:108 CI diesel engine. Exhaust emissions were recorded using Bacharach, PCA2 Qs1007 emission gas analyzer. Biodiesel production process was modeled and optimized using artificial neural networks (ANN) and response surface methodology. Engine performances and emission characteristics were optimized using response surface methodology-genetic algorithm (RSM-GA), ANN and Nelder-Mead’s (NM) simplex algorithm. Kinetics of transesterification of the seed oils in a batch reactor were conducted on both reversible and irreversible consecutive reaction mechanisms. The results showed that Prunus amygdalus gave the highest oil yield while esterified Dyacrodes edulis seed oil showed best fuel properties. Statistical quadratic regression models under the platform of Design Expert 7.0.0 version for RSM on central composite design (CCD) was best fitted models to describe the biodiesel production processes. ANN on the platform of MATLAB 2015 version gave better prediction than the RSM. RSM optimized and experimental validation yields were 91.09, 95.03 and 86.85%wt and 92.58, 94.55 and 85.91wt% respectively. The optimum conditions were: reaction temperature (50.03, 63.92 and 65.62˚C), reaction time (58.52, 58 and 62.04 minutes), catalyst concentration (2.04, 1.88 and 2.14wt%) and methanol to oil molar ratio (4.66:1, 6.86:1 and 5.88:1) for sweet almond seed oil methyl ester (SASOME), African pear seed oil methyl ester (APSOME) and African star apple seed oil methyl ester (ASASOME) respectively. Average emission impacts of the blends with respect to petro-diesel showed decrease in values except for NOX. All the engine performance characteristics except specific energies increased with load. Results showed that excellent correlation of results of experimental and predicted values were in the following order: NM>RSM>ANN for both SASOME and APSOME blends respectively. The kinetics of the reaction based on the proposed mechanism had the conversion of monoglycerides (Mg) to diglycerides (Dg), diglycerides (Dg) to triglycerides (Tg) and Tg to Dg at 55 °C, 60 °C and 65 °C respectively as the rate determining steps (RDS). Hydrolysis of Mg to Dg was the RDS at both 55 and 65oC while Dg to Mg was the RDS at 65oC for SASOME production. The conversion of Dg to Mg was RDS for ASASOME production at all the temperatures. Arrhenius activation energies for the SASOME, APSOME and ASASOME productions were in the range of 58.2-139.4, 41.0-104.55 and 97.71- 166J/mol. respectively. The overall results show that the seed oils are potential feedstocks for biodiesel production.