Green diesel blends design using decomposition-based optimization approach

Abstract

Petrodiesel-biofuel/biochemical blend (green diesel blend) is a promising solution in reducing environmental impact while improving the performance of petrodiesel. A systematic computer-aided approach can efficiently solve green diesel blends‟ design problems to replace the iterative, costly and time-consuming experimental trial-and-error approach. Other than the engine performance related fuel properties (density, kinematic viscosity, cetane number and higher heating value), the safety indicator: flash point is an important safety consideration for diesel fuel and it should be considered in the green diesel blend design to avoid fire accident. The aims of this study are to develop a systematic computer-aided tailor-made green diesel blend design algorithm and to improve the flash point prediction model, which is the Liaw model for B5 palm oil biodiesel (B5)-ester/ether/alcohol blends. The algorithm contains two main phases: the model-based design and the experimental validation. The optimum green diesel blend is computationally optimized in the model-based design phase. The accuracy of the Liaw model using UNIFAC type models is improved for B5-ester/ether/alcohol by regressing the UNIFAC group interaction parameters. The verified and improved Liaw models are embedded into the model-based design phase to optimize the green diesel blend. The physicochemical property, engine performance and emissions of the optimum blend obtained in the model-based design phase are experimentally validated in the experimental validation phase. The application of the developed design algorithm is illustrated by finding the right combination of the binary and ternary blends of B5-ester/ether/alcohol. The ideal Liaw model, the Liaw model using the original UNIFAC and the Liaw model using original UNIFAC with parameters set B (group interaction parameter between CH2 and OH are revised) are used to predict the flash points of the B5-ether, B5-ester and B5-alcohol blends. GB3 (B5-11.1 % by mass of diethyl succinate) and GT1 (B5-24.1% octanol-5.9 % diethyl succinate, by mass %) with or without the cetane enhancer: 2-ethylhexyl nitrate (2EHN) are the optimum binary and ternary blends with the high oxygen content obtained in the model-based design phase. Satisfactory experimental validation results were obtained in experimental validation phase and GT1A (GT1-0.17 % by mass of 2EHN) ) was identified to be the most promising green diesel blend owning to its lower emissions of nitrogen oxide (19.21 % lower than B5), un-burnt hydrocarbon (27.48 % lower than B5) and carbon monoxide (36.73 % lower than B5). Meanwhile, GT1A has comparative fuel efficiency to B5. The developed green diesel blend design algorithm serves as an improved model for solving green diesel blend design problem.