Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Topological data analysis for the turbulent premixed flames
This project aims to investigate the surface structure of turbulent premixed flames using topological data analysis.
Keywords: Topological data analysis; Direct numerical simulation; premixed turbulent combustion
For turbulent premixed flames, flame surfaces are typically wrinkled by the turbulent flow. Knowledge of the flame surface area increase by turbulence is essential for any modeling and prediction of premixed turbulent combustion. Thanks to the advance in computational resources, direct numerical simulations (DNS) have become a promising method to investigate the fundamental combustion processes since in DNS, all the scales including the smallest turbulent scales and the chemical scales are resolved. Therefore, DNS can give deeper insight into the evolution and structure of premixed flame surfaces, which may be characterized by the local curvature of the surfaces. A new idea of analyzing surface wrinkling is inspired by a decomposition of the entire flame surface into segments, on which the surface curvatures change monotonically. For this, the methods developed in the field of computer graphics and visualization are promising.
For turbulent premixed flames, flame surfaces are typically wrinkled by the turbulent flow. Knowledge of the flame surface area increase by turbulence is essential for any modeling and prediction of premixed turbulent combustion. Thanks to the advance in computational resources, direct numerical simulations (DNS) have become a promising method to investigate the fundamental combustion processes since in DNS, all the scales including the smallest turbulent scales and the chemical scales are resolved. Therefore, DNS can give deeper insight into the evolution and structure of premixed flame surfaces, which may be characterized by the local curvature of the surfaces. A new idea of analyzing surface wrinkling is inspired by a decomposition of the entire flame surface into segments, on which the surface curvatures change monotonically. For this, the methods developed in the field of computer graphics and visualization are promising.
In this thesis, we will use topological data analysis from computer graphics and visualization to develop post-processing methods that accurately decompose the flame surface into segments. After validation, these tools will be applied to existing state-of-the-art DNS data to gain a deeper understanding of combustion and flame propagation. In order to process the large DSN data, the post-processing tools should be able to run in parallel on the supercomputer.
In this thesis, we will use topological data analysis from computer graphics and visualization to develop post-processing methods that accurately decompose the flame surface into segments. After validation, these tools will be applied to existing state-of-the-art DNS data to gain a deeper understanding of combustion and flame propagation. In order to process the large DSN data, the post-processing tools should be able to run in parallel on the supercomputer.