Cerfacs Enter the world of high performance ...

PhD Defense: Sreejith Nadakkal Appukuttan

  Wednesday 12 January 2022 at 14h00

  Phd Thesis       CERFACS, Toulouse, Jean-Claude ANDRÉ Meeting Room     

Abstract
There has been a tremendous increase in the production of ethylene over the last few decades, which has put tremendous focus on the steam cracking process and with an objective of improving its efficiency. This study, performed as a part of IMPROOF (Integrated guided Model PROcess Optimization of steam cracking Furnaces) project, is intended towards attaining that goal. Large Eddy Simulation (LES) is at a stage of being mature enough to be used in design and optimization of processes in industrial equipments. However, the application of LES to study large flow equipments is still a technical challenge due to the high computational cost arising from numerical stiffness. In this study, a novel, chimera-based, local time stepping scheme is developed to speed up explicit time integration based LES solvers and applied (for the first time) to study the reactive flow inside a steam cracking furnace. This new numerical technique is studied for its numerical properties using Global Spectral Analysis (GSA) and the impact of local time stepping on the accuracy and resolution of the baseline numerical scheme is analysed. The speed up obtained using this method is also ascertained with the help of canonical 2D and 3D non-reactive as well as reactive flow simulations. Numerical prediction of combustion inside a steam cracker comes with its own challenges. While detailed chemical mechanisms are ruled out from being used in simulations due to its high cost, simple global chemical mechanisms are not accurate enough to predict the flame structure and flame properties accurately. In this study, species transport equations are used with an analytically reduced chemical (ARC) mechanism. These chemical mechanisms are reduced from an up-to-date detailed mechanism using Directed Relational Graph with Error Propagation (DRGEP) technique and quasi-steady state (QSS) assumptions. The reduced mechanism is validated with respect to the detailed mechanism and experimental measurements is found to be in excellent agreement for all the flame properties of interest in this study. Radiative heat transfer is the predominant mode of heat transfer in steam cracking furnaces and hence cannot be avoided in realistic furnace simulations. In this study, the LES solver (along with the newly developed acceleration technique) is coupled with a radiative transfer equation (RTE) solver to carry out a coupled LES-RTE simulation. The approach is validated with experimental data from an axisymmetric jet diffusion flame and the experimental and numerical data is observed to be in good agreement with each other. Finally, all these methodologies are simultaneously applied to study the reactive flow occurring in the fire side of a steam cracking furnace. The LES acceleration technique speeds up the computations while ARC mechanism assists in predicting combustion reactions in an accurate manner. The radiative heat transfer effects are included by coupling the LES solver with the RTE solver as mentioned previously. Unsteady LES simulations of the combustion occurring inside the firebox is carried out. The computed and measured data for temperature and heat flux is found to be in close agreement with each other. LES of such a furnace demonstrates revealing information on the flame stabilization mechanism and the mean flame properties such as its shape and length and are discussed in this thesis. This study is intended to be a technology demonstrator by addressing three of the core challenges in the numerical modeling of steam cracking furnaces. By addressing these challenges, it is hoped that the petrochemical community is taking one step closer to using LES for their design and analysis processes in the near future.
Jury
Prof. Ronan Vicquelin
Laboratoire EM2C (CNRS)
Rapporteur
Prof. Tapan K. Sengupta
Prof. IIT Dhanbad, Inde
Rapporteur
Prof. Frédérique Battin-Leclerc
Prof. LRGP, France
Examinateur
Mr. Gilles Theis
Technical Manager, JZHC, Luxembourg
Examinateur
Dr. Eléonore Riber
Cerfacs Toulouse, France
Directrice de thèse
Dr. Bénédicte Cuenot
Cerfacs Toulouse, France
Directrice de thèse

 

CALENDAR

Monday

04

December

2023

Artificial Intelligence for computational physics

From Monday 4 December 2023 to Friday 8 December 2023

  Training    

Monday

11

December

2023

Multi-architecture parallelism using Kokkos/C++ library

From Monday 11 December 2023 at 14h00 to Wednesday 13 December 2023 at 17h00

  Training    

Thursday

21

December

2023

PhD Defense : Aurélien LINÉ : ” Modulation of European near-term climate change by multi-decadal internal variability “

Thursday 21 December 2023 at 15h00

  CERFACS - Toulouse - France     Organized by Nathalie BROUSSET    

ALL EVENTS