Genetic Algorithm optimised Chemical Reactors network: A novel technique for alternative fuels emission prediction

摘要

Sustainability of the conventional jet fuels and climate change has attracted the aviation sector to diversity to alternative fuels. However, fuel diversification requires an assessment of the long term impact to engine performance and engine emissions through the combustion process, as alternative fuels are not as well understood as conventional jet fuel. A detailed experimental study on alternative fuels emissions across the entire aircraft fleet is impractical. Therefore a plausible method of computer modelling combined Genetic Algorithm and Chemical Reactors network was developed to predict alternative fuels gaseous emissions, namely, Carbon Monoxide, Nitrogen Oxides and Unburned Hydrocarbons in aircraft engines. To evaluate the feasibility and accuracy of the technique, exhaust emission measurements were performed on a re-commissioned Artouste Mk113 Auxiliary Power Unit, located at the University of Sheffield's Low Carbon Combustion Centre. The simulation produced results with good agreements with the experimental data. The optimised model was used to extrapolate emissions data from different blends of alternative fuels that did not operate during the campaign. The proposed technique showed that it can develop a data base of alternative fuels emissions and also act as a guideline for alternative fuels development.