ºù«Ӱҵ

The inference of chemical reaction networks has always been a challenging task when numerous species can potentially interact with each other. Significant amount of effort is required to postulate a chemical reaction network, which is then not guaranteed that will be reliable or applicable to a wide range of temperatures. The aim of this work is to develop an optimisation procedure aimed towards the automatic generation of chemical reaction networks for fuel thermal degradation.

This can be done by adopting a number of machine learning approaches such as target factor analysis and s-systems in which chemical reaction networks can be inferred requiring only the input of measured species concentrations. The result will be the generation of reaction networks that can be fast to calculate and reliable for a wide range of temperatures, therefore offering a significant improvement over the current state of the art mechanism generation for aviation fuels.

This work will be beneficial for the advancement of chemical kinetic modelling of thermally degrading fuels in gas turbine fuel systems because it will deliver a reliable methodology that can be applied to any proposed fuel chemistry. This is especially crucial for today when several SAF formulations are under development but without currently being able to assess for their thermal stability.

For further information contact Dr Spiridon Siouris (s.siouris@sheffield.ac.uk).