Hydrogen combustion releases just clean water as its main product without generating any carbon dioxide (CO2). For this reason it represent a key technological point for the hydrogen economy which is the framework in which hydrogen is employed as a zero-carbon fuel for a plethora of applications requiring high energy densities as those typical of propulsion and power applications. The development of hydrogen based combustion devices is often hindered by the lack of reliable predictive numerical simulations tools which still need detailed data for model developments. The present research project aims at filling this gaps, providing the much needed data-driven modeling support to clean combustion in the hydrogen economy.
High pressure, lean premixed hydrogen-air turbulent combustion in fact rises as the most efficient combustion procedure, which is however not easily achieved since lean hydrogen flames are intrinsically unstable mainly due to the unbalanced competition between heat and mass diffusion of hydrogen. The project will investigate the interaction between turbulence and intrinsic instabilities with the following objectives: 1) generation of a large and detailed dataset using state of the art direct numerical simulations employing high order methods capable of taking full advantage of high performance computing infrastructures; 2) Big-data organization, mining and physical understanding of the interaction between intrinsic instability and turbulence; 3) Development of knowledge based combustion models that can be directly used "off the shelf" for the development of new hydrogen based combustion technologies.
The present research project will be devoted to the development and investigation of hydrogen flames by means of the generation of a large and detailed dataset by means of state of the art DNS. The most innovative part is represented by the development of unprecedented dataset featuring three dimensional DNS of high pressure hydrogen combustion which are not present in the current literature as described in the state of the art paragraph of the previous section. State of the art high order methods and high performance computing strategies will also be employed. New data organization strategies, data mining as well as physical understanding of the interaction between intrinsic instability and turbulence will be developed. In order to do that, ad hoc and innovative big-data tools will be developed to start the consistent development of knowledge based combustion models. The newly proposed idea is to create turbulent combustion models that can be directly used "off the shelf" for the supporting new hydrogen based combustion technologies by means of computational fluid dynamics (CFD).
In this framework, an other important innovative aspect of the present research project will be a particular attention that will be given to the sharing of the data following the open data European best practices (https://data.europa.eu/euodp/en/data/). In fact, in order to maximize the impact of the research project , as data is generated it will be stored on-line in a dedicated website or on a suitable database like the European Research Community on Flow, Turbulence and Combustion database (http://cfd.mace.manchester). I personally think that data sharing can be of paramount importance as datasets are extremely cost-intensive in terms of time and resources employed for such large scale simulations (1 CPU hour on a HPC infrastructure costs roughly ~0.05€, while a single state of the art DNS can cost up ~10^7 CPU hours). Moreover, performing these simulations as well as managing the data requires highly qualified researchers, making such costs nearly impossible to justify from both a scientific community and industrial point of view.