Sub-Grid Modeling of Diesel Particulate Filtration Using the Lattice-Boltzmann Method
EMSL Project ID
11390
Abstract
Diesel engines offer a promising strategy to help meet short-term global energy demands while conserving petroleum resources. Emissions from diesel engines, especially submicron particulates, pose a growing health concern. There is general consensus among government and industry in America, Asia, and Europe that some form of Diesel Particulate Filter will be an essential aspect of future clean diesel engine technology. Practical DPF development over the past decade has been encumbered by a host of technical difficulties. The ideal DPF system would efficiently trap particulates while adding little back-pressure (resulting in small negative impact to fuel economy) and providing reliable operation through thousands of regeneration cycles. Many aspects of filter operation are still not well understood.A suite of lattice-Boltzmann models have been developed at PNNL to help fill the existing gap in fundamental understanding of DPF operation. Models covering a range of length scales examine the distribution of diesel soot throughout DPF devices, the deposition of soot particles within and upon porous substrates, and the diffusion of active gaseous components from surface catalytic sites into soot regions during filter regeneration. The lattice-Boltzmann method is uniquely suited to sub-grid models involving exhaust flow through the complex and irregular porous structure of DPF substrates.
As detail is added to the lattice-Boltzmann models, more computational resources are required. The LB method is inherently parallelizable, and the ability to run on 500 or more processors may greatly accelerate model development and the generation of useful results. Large parallel runs are currently limited to the computational resources available on Altix1, which will not be adequate for anticipated model use and development in the coming year.
Discrete particle DPF simulations carried out to date have tracked the flight and deposition of roughly a million soot particles to model the initial stages of DPF loading (please see attachment). Model runs to date have not included a sufficient number of iterations to adequately resolve the flow field evolution as soot deposits form. In addition, more iterations and particles will be necessary to model the later stages of soot cake formation. Insight gained from discrete particle simulations must then be incorporated into high-resolution regeneration models which predict the migration of active gaseous species within the porous filter substrate to soot deposits.
The existing discrete particle model will be extended to provide meaningful simulations of DPF loading. Model predictions will then be compared to experimental data for multiple substrate materials under a variety of operating conditions. Insight gained from the discrete particle deposition model will be used to develop a detailed regeneration model to help optimize catalyst performance. Results will be submitted to a peer reviewed journal.
Project Details
Project type
Capability Research
Start Date
2004-10-15
End Date
2006-11-08
Status
Closed
Released Data Link
Team
Principal Investigator
Related Publications
Dillon, H., et al., Optimizing the Advanced Ceramic Material for Diesel Particulate Filter Applications. SAE 2007 World Congress, 2007. 2007-01-1124.
Stewart ML, TR Gallant, DH Kim, GD Maupin, and A Zelenyuk. 2010. Fuel Efficient Diesel Particulate Filter (DPF) Modeling and Development . PNNL-19476, Pacific Northwest National Laboratory, Richland, WA.