Numerical Modelling of Soot Formation in Turbulent Non-Premixed Flames

Olga Roditcheva

Research output: ThesisDoctoral Thesis (compilation)


This thesis describes an extended flamelet approach for modelling of soot formation in turbulent hydrocarbons diffusion flames. Detailed chemical kinetic mechanisms are used to compute species concentrations, temperature, density and soot formation rates as functions of flamelet coordinates in the space of mixture fraction and scalar dissipation rate. An ensemble average based on a presumed probability density function of the mixture fraction and scalar dissipation rate provides the mean density, species concentrations, mixture enthalpy, temperature and soot formation rates.

A semi-empirical two-equation model of finite rate soot formation is applied to methane/air turbulent diffusion flames at different pressures and to propane/air turbulent flames with different air temperatures. Soot number density and soot volume fraction are modelled by transport equations with empirical expressions for source terms, which represent soot formation and oxidation rates.

It is shown that the model gives reasonable results for the flames at atmospheric pressure. It, however, over-predicts the soot yield, if the model parameters calibrated for the atmospheric pressure flames are used. It is shown that a decrease of the surface growth rate constant in proportion to 1/p leads to good agreement with the experiment.

Models of soot oxidation due to different chemical species, such as O2, H2O, CO2, OH, O and H, are evaluated at the pressure range from 1 to 3 atm. It is found that oxidation of soot by hydroxyl radicals, OH, is the dominating factor in soot burnout.

The ability of an extended flamelet approach, with semi-empirical modelling of soot formation, to reproduce the effect of residence time in the process of soot formation is examined. A decrease in residence time is achieved by controlling fuel injection diameter while fuel mass flow rate is kept the same. It is shown that the semi-empirical soot model is able to predict a decrease in soot volume fraction when there is a decrease in the residence time.

The effect of air preheat on soot formation is investigated numerically for propane/air diffusion flames at two incoming air temperatures Ta=323 K and Ta=773 K. The air preheat leads to an increase in the flame temperature and in the concentration of soot precursor species, which leads to an increase in soot concentration in the flame. However, the increase in temperature in the post flame zone leads to an increase in the soot oxidation rate and the amount of soot in the exhaust gases is drastically reduced. To account for the temperature effect on the soot formation modifications to an existing semi-empirical model are proposed, which take into account bell-shape soot dependence on temperature.

A presumed probability density function and flamelet library approach is developed further in order to incorporate the influence of turbulence on soot formation. The numerical calculations are compared with experimental measurements and previous numerical simulations with a simple model for calculating mean source terms.
Original languageEnglish
Awarding Institution
  • Fluid Mechanics
  • [unknown], [unknown], Supervisor, External person
Award date2000 Dec 18
Print ISBNs91-628-4583-7
Publication statusPublished - 2000

Bibliographical note

Defence details

Date: 2000-12-18
Time: 10:15
Place: N/A

External reviewer(s)

Name: Kraft, Marcus
Title: [unknown]
Affiliation: Ph.D., Dept. of Chemical Engineering, University of Cambrige, UK.


Subject classification (UKÄ)

  • Fluid Mechanics and Acoustics


  • Motorer
  • Motors and propulsion systems
  • Teknik
  • Technological sciences
  • oxidation of soot.
  • soot
  • polutant formation
  • flamelet approach
  • Combustion
  • turbulence
  • framdrivningssystem
  • Thermal engineering
  • applied thermodynamics
  • Termisk teknik
  • termodynamik


Dive into the research topics of 'Numerical Modelling of Soot Formation in Turbulent Non-Premixed Flames'. Together they form a unique fingerprint.

Cite this