1-Sentence-Summary: A multistage simulation model is able to predict coal particle ignition and verifies the flamelet approach for these type of calculations.
Authors: M. Vascellari, H. Xu, and C. Hasse
Read in: Three Minutes
Favourite quote from the paper:
The current authors develop a multistage simulation approach to investigate ignition of coal particles. First an Eulerian/Lagrangian model is used to simulate transient motion and heating of a single particle placed downstream in a co-flow burner. Temperature, devolatilization rate, and Reynolds number are recorded from this simulation.
The results from the first model are used as input into a 2D Eulerian simulation at meso-scale, that resolves the velocity and species profiles in the vicinity of the particle. Here, the domain extends 100 diameters in the upstream direction, 200 diameters in the downstream direction, and 75 diameters in the cross-stream direction. The boundaries contain a circular cutout of the particle through which devolatilized gases and heat are transferred.
The fully resolved simulation is able to capture transient ignition and the particle diffusion flame. However, the scalar dissipation rate is also recorded to feed into a third simulation. The third simulation involves a 1D flamelet model from the scalar dissipation and temperature fields. The purpose of this model is to compare to the fully resolved simulation and verify the ability of the laminar flamelet approach to predict coal particle ignition.
The case simulated in this work is based on the experiments of Molina and Shaddix, 2007. The coal particle has a 100 µm diameter and contains 36% volatile matter. The particle is released downstream in a co-flow burner. The inflow conditions for the simulation are based measurements from the experiment.
Three of the main findings from this paper are:
- The Eulerian/Lagrange approach overpredicts coal particle ignition time by 21%.
- The resolved simulation approach underpredicts coal particle ignition time by 14%
- The laminar flamelet model is able to reproduce the ignition results of the resolved model downstream from the particle when the correct scalar dissipation field is used.
The following sections outline the main findings in more detail. The interested reader is encouraged to view the complete article at the link provided below.
Finding #1: Ignition delay is overpredicted using the Eulerian/Lagrange modeling approach
The coal particle ignition time using the Eulerian/Lagrange approach is 34 ms. This is 24% longer than the experimental ignition time of Molina and Shaddix, 2007 (28 ms).
The authors propose that the Lagrangian approach is unable to capture the boundary layer surrounding the particle. Furthermore, ignition takes place in the boundary layer several diameters away from the particle surface. The computational mesh resolution in this simulation is limited by the particle diameter and the steep temperature gradients in the flow cannot be accurately predicted.
Finding #2: Ignition delay is underpredicted using the fully resolved approach
Using the Eulerian/Lagrange model results as boundary condition input for the meso-scale simulation allows the boundary layer, temperature, and species profiles in the vicinity of the particle to be resolved. Furthermore, much higher grid resolution can be used.
In the resolved simulation ignition takes place approximately 24 diameters downstream of the particle. The flame then propagates back towards and around the particle. Lastly, the particle burns as a quasi-steady diffusion flame. The ignition delay time is predicted as 24 ms which is 15% faster than the experimental results.
Finding #3: The laminar flamelet model is able to reproduce the downstream ignition results of the resolved model
The last simulation uses a 1D flamelet approach with the temperature field from the resolved model. If the scalar dissipation rate from the resolved model is used the flamelet ignition delay time is nearly identical to the resolved simulation. Species profiles between the resolved and flamelet approaches also agree well.
If the scalar dissipation rate in the flamelet model is assumed to be zero, the ignition delay is significantly underpredicted at 22 ms. If a standard profile for diffusion flames is used (from the “Turbulent Combustion” textbook of Peters) the ignition delay is significantly overpredicted. This demonstrates the importance of scalar dissipation in the laminar flamelet approach.
One limitation of the current flamelet approach is that ignition on the upstream side of the particle cannot be accurately captured. This ignition occurs due to the flame wrapping around the particle. This geometric effect is not be captured in the current flamelet model which can only describe local auto-ignition processes. However, after ignition takes place there is good agreement in temperature and species profiles with the fully resolved model.
My Personal Take-Aways From
“Flamelet Modeling of Coal Particle Ignition”
This paper presents a very interesting methodology for investigating particle combustion and flame propagation. Using multiple simulations to resolve large and small scale features is an important concept and should be used more frequently in dust explosion research moving forward. The paper gives several referenced of other studies completing numerical simualtion of coal combustion such as Du and Annamalai, 1994, Wendt et al., 2002, Higuera, 2009, and Zhu et al., 2011.
The paper also presents two modeling concepts which may be useful to dust flame and explosion research. The first is the CPD model (chemical percolation devolatilization) used to describe detailed devolatilization of coal particles. The CPD model used in the current article is based on the work of Grant et al., 1989 and Genetti et al., 1999 .
The second modeling concept is the laminar flamelet approach to coupling chemistry and the flow field. As shown here this model can be used to predict correct ignition times for coal particles under laminar conditions. However, the real strength in this approach is that it can be extended to non-premixed and turbulent flames which makes it attractive to use for CFD calculations.
Full Citation: [bibtex file=references.bib key=Vascellari2013]
[otw_shortcode_button href=”http://www.sciencedirect.com/science/article/pii/S154074891200260X” size=”medium” icon_position=”left” shape=”square”]> > Get The Article[/otw_shortcode_button]