AmericanChemicalSociety.com

Background and Motivation

Ever since the introduction of diesel engines, many research and development activities have focused on improving their performance and minimizing their emissions. To control diesel particulate matter (PM) emissions, diesel particulate filter (DPF) systems have been developed to trap and oxidize the solid PM from diesel engine exhaust gases. Models for the transient PM loading and the pressure drop in the DPF channels during loading can be coupled with PM regeneration models to lead to a more efficient and safe regeneration in the DPF.

The goals of this project are to understand the fundamental transport processes and reaction kinetics in the DPF through modeling efforts and then to utilize the models to help diesel engines meet EPA or California PM emission standards, which can lead to improved environmental quality and human health.

Summary of Annual Progress and Impact of Research

In the initial stages of this project, parametric and sensitivity analyses were carried out using two different DPF regeneration models. This led to the completion of a MS Thesis by graduate student Di Huang in September 2008.

During the second year of this project, research has focused on three key areas:

- Modeling of PM filtration in the DPF

- Modeling the pressure drop across a loaded DPF

- Applications to real world driving conditions (using the urban dynamometer driving schedule)

To aid with this research, two models have been reproduced from the technical literature (E.A. Kladopoulou, S. L. Yang, J. H. Johnson and G. G. Parker SAE 2003-01-0842).

The first model is a filtration model which is used to calculate the filter collection efficiency. There are two types of filtration: “deep bed filtration” and “cake filtration”. Filter materials with large pore sizes are well known for filtering small particles by mechanisms of impaction and diffusion; this is called “deep bed filtration” and occurs within the pores of the DPF. During filtration, the collected particles can increase the collection efficiency by “cake filtration” when a deposit layer is built up on the DPF walls. The filtration efficiency increases rapidly during “deep bed filtration” to approach unity, and will remain so during “cake filtration”.  

A typical initial size for the filter pores is 30 mm. The modeling results show that the filtration efficiency increases sharply when the effective pore size of the substrate wall is reduced from 30μm (80% efficiency) to 20μm (99.9% efficiency) due to the collected PM. This initial process is “deep-bed filtration.” When the pore size is smaller than 20μm, the filtration is mainly “cake filtration” and a deposit layer builds up on the filter walls. The filter permeability used in our model is 3.0 x 10^-13 m². Our full filtration model includes an empirical model for PM deposit thickness with position and total PM deposit mass in the DPF.

The second model is the pressure drop model. Besides high filtration efficiency, a low pressure drop is another main requirement on the filter. Our modeling research first estimates the pressure drop across an unloaded DPF. The pressure drop model has also been linked to the filtration model for the loaded filter. This is a non-trivial extension of the model due to temperature gradients along the filter length. Thus, the gas density is varied such that a nonuniform PM deposit thickness arises which strongly impacts the DPF pressure drop and gas flow.

Our work has varied DPF design and operating parameters to perform a parametric study and sensitivity analysis on the DPF pressure drop. The results shows that the gas flow rate and deposit thickness have an equal impact on the pressure drop and that the pressure drop increases with increasing gas flow rate, initial deposit thickness, and substrate wall thickness. More importantly, the pressure drop and regeneration models have been linked together. In general, the pressure drop increases with increasing inlet temperature. However, when the inlet temperature is over 700K, the particulate which has been accumulated near the leading edge of the filter is burned off. This leads to a lower local pressure drop near the filter entrance. This is important to channel the flow appropriately to regenerate the entire filter. 

It is clear from the above discussion that the PM filtration and DPF pressure drop play important roles in reducing PM emissions. These models provide inputs to determine an appropriate timing for DPF regeneration. At present, we are working on integrating realistic inputs to the models which are based upon the UDDS (Urban Dynamometer Driving Schedule). A driving cycle is a series of data points representing the speed of a vehicle as well as the engine out temperature, engine speed, engine load, exhaust gas flow rate, and PM concentration versus time. This will enable calculation of the PM accumulated in the DPF versus time during the UDDS driving cycle and determine a proper way for regenerating the DPF in an efficient manner.

Impact on Career of PI and Students

Funding from the American Chemical Society has allowed the PI to pursue this important project in emissions abatement and to support a graduate student that otherwise could not be supported. The student has benefited from the ability to study a research problem and completed his MS in Chemical Engineering during the past year. A presentation was given at the 2008 AIChE Annual Meeting and the student and the PI have also submitted a journal article based upon the MS thesis research.