Motivation of the AI-FIE project
Decarbonising the transport sector, responsible for ~25% of Green House Gas (GHG) emissions, is increasingly urgent in the race to address climate change. The European Union (EU) has already taken bold steps in this direction with the launch of the European Green Deal, in which achieving net zero GHG emissions by 2050 is the main strategy goal. However, electrification of internal combustion engine (ICE) vehicles is a long-term process; a disappointing ~25% growth in liquid fossil fuel demand globally is foreseen, due to increased commercial activity, movement of people and products and building infrastructure. While electric vehicles offer a significant reduction in GHG emissions, their life-cycle emissions are far from zero, mainly due to the embedded emissions in the production of the battery pack and the non-renewable electricity production powering the vehicle; neither of which is measured at point of use. In an effort to mitigate the immediate and inevitable environmental impact, hydrogen-derived CO2-neutral synthetic fuels produced using renewable energy sources (e-fuels) represent a promising alternative. Use of e-fuels avoids the need for new vehicles/ infrastructures, as they can be utilised in existing ICE and jet engines. To produce e-fuels, electricity is used to split water into H2 and O2; the H2 is then combined with CO2 to produce hydrocarbons; e.g. paraffinic distillate fuels to replace kerosene and diesel and lighter components such as liquefied petroleum gas (LPG) or naphtha. The matrix of possible combinations of conventional and e-fuels, injected from the wide range of fuel injection equipment (FIE) available for various internal combustion engines (ICE) types and sizes operating over a wide range of pressures and temperatures (P-T), makes testing a long-term process. The motivation here is to develop a data-driven deep learning (DL), artificial intelligence (AI) algorithm that will significantly accelerate e-fuel and FIE development by predicting simultaneously the in-nozzle flow and its effect on the characteristics of vaporising liquid fuel sprays at time scales 3-4 orders of magnitude faster compared to today’s state-of-the-art experimentation and CFD simulations. Predictions will consider the e-fuel composition, FIE design and the varying P-T conditions realised in combustion systems. Central to this process is the method predicting the physical properties of the e-fuel. Despite that the existing fuel properly libraries are limited to simplified hydrocarbon components, Equations of State (EoS) using the Perturbed Chain Statistical Associating Fluid Theory (PC-SAFT) have been efficiently applied to simultaneous predictions of nozzle flows and sprays for a wide range of fuel compositions.