• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Temperature excursions during regeneration of diesel particulate filters

Dissertation
Author: Kai Chen
Abstract:
A major technological challenge in the operation of diesel particulate filters (DPFs) is prevention of occasional melting of the ceramic filters during regeneration. The dynamic features of the combustion of particulate matter (PM) on a single layer diesel particulate filter were studied using IR imaging. The experiments showed that at stationary feed conditions the soot combustion may proceed in three different modes: either by a single moving hot zone or several hot zones generated at different ignition points or uniform combustion all over the surface. The maximum temperature of the moving fronts was much higher than those attained during uniform combustion. The highest temperature attained under stationary (constant feed) combustion is too low to cause the filter melting (melting temperature ∼1250°C). We conjecture that high temperature excursions are a counter-intuitive response to a rapid deceleration which decreases the exhaust gas temperature and flow-rate and increases the oxygen concentration. The experiments showed that a step-change of the feed temperature led to a transient temperature exceeding the highest attained under stationary operation with the initial feed temperature. A simultaneous step-change of the feed temperature, flow rate and oxygen concentration in the feed led to a transient temperature that exceeded the highest attained for stationary operation under either the initial or final operation conditions. It was also higher than those generated either by a step change of any single feed input or by a step-change of any pair of feed inputs. The temperature rise depended in a complex way on several factors, such as the direction of movement of the propagating front, the location of the front when the feed was step-changed, and whether the step-change was done either before or after formation of a moving temperature front. The simulations provided insight about the dependence of the amplitude of the temperature rise on the step change of the operating variables. The understanding generated by the simulations should help develop operation and control protocols that circumvent or at least decrease the probability of the occurrence of the destructive melting of the DPF.

Table of Contents Acknowledgements iv Abstract vi Table of Contents viii List of Figures xi List of Tables xx Nomenclature xxi Chapter 1 Introduction 1 1.1 Introduction 1 1.2 Diesel Particulate Filter 4 1.3 Regeneration Methods 10 1.4 Literature Review 15 1.5 Research Objective 21 Chapter 2 Experimental System 25 2.1 Introduction 25 2.2 Gas Flow System 25 2.3 Gas Preheating System 28 2.4 Reactor System 28 2.4.1 Diesel Particulate Filter 28 2.4.2 Reactor 29 viii

2.4.3 PM Deposition 31 2.5 Data Acquisition and Analysis System 33 2.5.1 Infrared Camera 33 2.5.2 Thermagram IR Image Processing Software 41 2.5.3 USB Data Acquisition System 42 2.6 Experimental Procedure 43 Chapter 3 Behavior Features of Soot Combustion in Diesel Particulate Filter 46 3.1 Introduction 46 3.2 Effect of Soot Loading on the Regeneration Process 47 3.3 Effect of Oxygen Concentration on the Regeneration Process 54 3.4 Effect of Filtration Velocity on the Regeneration Process 65 3.5 Discussion and Concluding Remarks 72 Chapter 4 Wrong-Way Behavior During Soot Combustion in Diesel Particulate Filter 80 4.1 Introduction 80 4.2 Wrong-Way Behavior of Diesel Particulate Filter: I. Impact of Feed Temperature82 4.3 Wrong-Way Behavior of Diesel Particulate Filter: II. Impact of Filtration Velocity 91 4.4 Temperature Excursions during Regeneration of Diesel Particulate Filter: I. Impact of Simultaneous Change of Two Input Variables 94 IX

4.5 Temperature Excursions during Regeneration of Diesel Particulate Filter: II. Impact of Simultaneous Change of Three Input Variables 104 4.6 Discussion and Concluding Remarks 116 Chapter 5 Temperature Distribution Within a Soot Layer During DPF Regeneration 122 5.1 Introduction 122 5.2 Overview of the Experimental System 123 5.3 Experiment Results 125 5.4 Discussions and Concluding Remarks 135 Chapter 6 Modeling of Soot Regeneration in the Diesel Particulate Filter 138 6.1 Introduction 138 6.2 Mathematical Model 139 6.3 Influence of Feed Temperature Reduction 146 6.4 Influence of Feed Rate Perturbation 159 6.5 Temperature Excursion during Transient Feed Conditions 168 6.6 Discussion and Concluding Remarks 184 Chapter 7 Conclusions 187 References 195 x

List of Figures Figure 1.1 The schematic of DPF channels 5 Figure 1.2 The commercial available DPF 5 Figure 1.3 Scheme of particle deposition mechanisms on collecting bodies 8 Figure 1.4 Crossing section SEM image of PM trapped inside filter pores and soot cake formed on the surface of the filter wall 9 Figure 1.5 Continuous regeneration trap filter system 12 Figure 1.6 Schematic of a single stage reactor plasma regeneration system 14 Figure 2.1 Experimental setup of the gas flow system 26 Figure 2.2 Commercial DPF on top of it is a single and multiple layers DPF used in the experiments 29 Figure 2.3 A schematic of the reactor 30 Figure 2.4 A schematic of the experimental system 30 Figure 2.5 The image of the PVA250 system 33 Figure 2.6 A schematic representation of a general thermo graphic measurement situation 1: Surroundings; 2: Object; 3: Atmosphere; 4: Camera 34 Figure 2.7 The image of the Merlin camera and the lenses 36 Figure 2.8 Infrared Camera calibration curve for planar DPF 41 Figure 3.1 Thermal images of PM regeneration at O2 = 20 vol.%, v = 5 cm3/(cm2-s) and T, = 550°C. (a) Single ignition point, PM = 10 g/L; (b) Two ignition points, PM = 8 g/L; (c) Uniform combustion, PM = 5 g/L. Black arrow is direction of flow 49 xi

Figure 3.2 Temperature image during single point ignition: (a) localization of ignition; (b) 22 sec after ignition. Soot loading of 10 g/L; v = 5 cm3/ (cm2-s) 52 Figure 3.3 Temperature profile along single layer DPF following ignition at a single point at t=22 sec; Soot loading of 10 g/L; 02 = 20 vol.%; v = 5 cm3/ (cm2-s) 53 Figure 3.4 Dependence of ignition points on PM layer thickness; T, = 550°C; O2 = 20 vol.%; v = 5 cm3/ (cm2-s) 54 Figure 3.5 Thermal images of soot regeneration for soot loading of 10 g/L. Feed oxygen concentration (a) 5 vol.%; (b) 10 vol.%; (c) 12.5 vol.% and (d) 15 vol.% 55 Figure 3.6 Dependence of the maximum temperature rise on the oxygen concentration for a soot loading of 10 g/L; v = 12 cm / (cm -s) 58 Figure 3.7 Moving temperature profiles originating from end of DPF. Soot loading of 10 g/L and O2=10 vol.%, v = 12 cm3/ (cm2-s). (Same experiment as that in Figure. 3.5b) 59 Figure 3.8 Coalescence of temperature fronts originating from two ignition points. Soot loading of 10 g/L and 02=15 vol.%, v = 12 cm / (cm -s). (Same experiment as that in Figure. 3.5d) 60 Figure 3.9 Effect of oxygen concentration on the relation between maximum front temperature and its propagation velocity; v = 12 cm3/ (cm2-s); Soot loading of 10 g/L 61 Figure 3.10 IR images of soot combustion for soot loading of 20 g/L and oxygen concentration of (a) 7.5 vol.% ; (b) 10 vol.% ; (c) 12.5 vol.% ; (d) 15 vol.%. 63 Xll

Figure 3.11 Dependence of the maximum temperature rise on the oxygen concentration for soot loading of 20 g/L. The sequence of IR images corresponding to cases b, c, and d are shown in Figure.3.10. At a uniform combustion occurs. 64 Figure 3.12 Effect of oxygen concentration on the maximum temperature of upstream and downstream moving fronts; v = 12 cm / (cm -s); Loading of 20 g/L. ...65 Figure 3.13 Thermal images of soot regeneration for soot loading of 10 g/L at several gas flow rates through the filter (a) 5 cm3/(cm2-s); (b) 7 cm3/(cm2-s); (c) 12 cm3/(cm2-s) 68 Figure 3.14 Moving temperature profiles originating from front of DPF; Soot loading of 10 g/L and v = 5 cm3/ (cm2-s). (same experiment as that in Figure 3.13 a) .69 Figure 3.15 Moving temperature profiles originating from middle of DPF. Soot loading of 10 g/L and v = 7 cm3/ (cm2-s). (same experiment as that in Figure 3.13b) 70 Figure 3.16 Impact of filter wall flow rate on maximum front temperature at several 02 concentrations and soot loading of 10 and 20 g/L 72 Figure 4.1 Moving temperature zone during DPF regeneration. Feed temperature of 620°C, 02 = 10 vol.% and v = 12 cm3/ (cm2-s). PM Loading (a) 10 g/L; (b) 20 g/L 84 Figure 4.2 Thermal images of motion of hot region during DPF regeneration. Feed temperature was decreased from 620°C to 520°C. O2 = 10 vol. % and v = 12 cm3/(cm2-s) 84 Figure 4.3 Moving ignited region originating from downstream of DPF 85 xiii

Figure 4.4 Upstream moving temperature profiles with the highest peak for 10 g/L; Feed temperature was decreased from 620°C to 520°C when temperature front was at (1) downstream (2) middle (3) upstream section of DPF 87 Figure 4.5 Temporal temperature measured by thermocouples in the reactor; Thermocouple 2-5 are attached to DPF, 1 ahead of DPF 88 Figure 4.6 Thermal images of motion of hot region during DPF regeneration. Feed temperature was decreased from 620°C to 520°C. O2 = 10 vol. % and v = 12 cm3/ (cm2s); PM Loading of 20 g/L 89 Figure 4.7 Moving temperature profiles with the highest peak for 20 g/L; Feed temperature was decreased from 620°C to 520°C when temperature front was at (1) downstream (2) middle (3) upstream section of DPF 90 Figure 4.8 Moving temperature front profiles following a rapid decrease of filtration velocity from 12 to 8.3 cm3/ (cm2-s); T, = 620°C and 02 = 10 vol.%; (a) PM loading of 10 g/L; (b) PM loading of 20 g/L 92 Figure 4.9 Impact of filtration velocity on temporal temperature profiles following a rapid shift in T, = 620 °C to 520 °C and 02 from 10 to 15 vol. % (a) v = 5 cm3/ (cm2-s); (b) v = 12 cm3/ (cm2-s); PM loading of 10 g/L 98 Figure 4.10 Upstream moving temperature profiles with the highest peak for 10 g/L; Feed temperature and filtration velocity were decreased when temperature front was at (1) downstream (2) middle (3) upstream section of DPF 99 Figure 4.11 Moving temperature front profiles following a rapid shift in Ti = 620°C to 520°C and 02 from 10 to 15 vol. %; v = 12 cm3/ (cm2-s); PM loading of 20 g/L 100 XIV

Figure 4.12 Moving temperature front profiles following a rapid shift in Ti = 620°C to 520°C and filtration velocity from 12 to 8.3 cm3/ (cm2-s); 02 = 10 vol. %. 101 Figure 4.13 Moving temperature profiles with the highest peak; Feed filtration velocity was decreased from 12 to 8.3 cm3/ (cm2-s), T, = 620 °C to 520°C when temperature front was at (1) upstream (2) middle (3) downstream section of DPF); 02 = 10 vol.% 102 Figure 4.14 Downstream Moving temperature wave profiles following a sudden shift in O2 = 10 to 15 v. % and filtration velocity from 12 to 8.3 cm3/ (cm2-s); T, = 620°C 104 Figure 4.15 Response of an upstream moving front to a step change at 42 seconds after feed introduction to the DPF 107 Figure 4.16 Moving upstream temperature profiles with the highest temperature rise following a step change in the feed at the locations marked on the horizontal line. Initial feed conditions as in Figure 4.15 109 Figure 4.17 Response to a step-change before a downstream ignition occurred 20 seconds after feed introduced to DPF. Initial conditions as in Figure 4.15 110 Figure 4.18 Response of a downstream moving front to a step-change at 36 seconds after introduction of feed to the reactor. Initial DPF and feed temperature of 650°C I l l Figure 4.19 Response of a downstream moving front to a step-change at 42 seconds after introduction of feed to DPF. Initial soot loading of 20 g/L 113 XV

Figure 4.20 Moving upstream temperature profiles with the highest temperature rise following a step-change in the feed at the locations marked on the horizontal line. Initial soot loading of 20 g/L 114 Figure 4.21 Response to a step-change before ignition occurred 20 seconds after introduction of feed to DPF. Initial soot loading of 20 g/L 115 Figure 5.1 A schematic of (a) the experimental system; (b) of the reactor 124 Figure 5.2 Thermal images of a moving of hot region during the soot regeneration; T, = 620°C, 02 = 10 vol. % and PM = 10 g/L; (a) v = 5 cm3/ (cm2-s); (b) v = 12 cm3/(cm2-s); 127 Figure 5.3 (a) Dependence of the surface peak temperature on the thickness of the soot filter; (b) Dependence of the average shrinking velocity on the soot layer thickness 128 Figure 5.4 Dependence of the average soot layer shrinkage velocity on the peak front temperature at three oxygen concentrations; Soot loading of 10 g/L 129 Figure 5.5 Impact of the oxygen concentration on the temperature distribution; PM = 10 g/L 131 Figure 5.6 Impact of filtration velocity on the temporal temperature profiles; PM = 10 g/L. 132 Figure 5.7 Impact of oxygen concentration on the temperature distribution; PM = 20 g/L. 133 Figure 5.8 Impact of filtration velocity on the temperature distribution inside the soot layer and the ceramic filter; PM = 20 g/L 134 Figure 6.1 A schematic diagram of a pair of inlet and outlet channels 140 xvi

Figure 6.2 (a) Temporal axial temperature profiles during PM regeneration under stationary feed condition (b) Corresponding temporal deposited PM profiles (c) Evolving dimensionless filtration velocity profiles (d) Profile of O2 concentration 148 Figure 6.3 Impact of filtration velocity on temporal temperature profiles. In both cases temperature perturbation is initiated when temperature front reached x/L = 0.5. 150 Figure 6.4 Dependence of the amplitude of the wrong-way temperature rise on (a) The amplitude of the sudden temperature decrease (b) The position of the temperature front when the feed temperature i s decreased 152 Figure 6.5 (a) Peak local temperature for a DPF length of 0.09 m and 0.18 m (b) Impact of DPF length on the maximum temperature under stationary feed conditions (dashed line) and following a sudden temperature decrease (solid line) 153 Figure 6.6 (a) Wrong-way behavior of PM combustion. Feed temperature is decreased at when front reaches x/L = 0.5; (b) Corresponding transient PM deposit thickness profile 155 Figure 6.7 Impact of filtration velocity on wrong-way behavior during PM combustion. PM loading of 8 g/L. Feed temperature is decreased when temperature front reaches x/L = 0.5 156 Figure 6.8 Impact of the heat transfer coefficient on the wrong-way behavior of PM combustion. Initial DPF temperature is (a).700 K (b). 900 K. Solid line h = 160 W/(m2-K); Dashed line h = 320 W/(m2-K) 158 xvii

Figure 6.9 Impact of the filtration velocity on the pressure drop for soot loadings of 6 g/L, 8 g/L and 10 g/L 160 Figure 6.10 Impact of the feed velocity on the peak temperature at stationary feed conditions (a) Tin, = 700 K; (b) Tini = 900 K 162 Figure 6.11 Transient temperature profiles following a reduction in filtration velocity (a) velocity was decrease from 3.34 to 1.67 cm/s; (b) velocity was decrease from 3.34 to 0.83 cm/s; (c) velocity was decrease from 4.67 to 3.34 cm/s; Tj =700 K; Tmi = 900 K 165 Figure 6.12 Transient temperature profiles following a reduction in filtration velocity (a) velocity was decrease from 3.34 to 1.67 cm/s; (b) velocity was decrease from 3.34 to 0.83 cm/s; (c) velocity was decrease from 4.67 to 3.34 cm/s; Tj = Tlni = 900 K 167 Figure 6.13(a) Temporal axial temperature profiles during homogeneous PM regeneration with relative low feed temperature of 823 K; (b) Corresponding temporal deposited PM profiles 170 Figure 6.14(a) Temporal axial temperature profiles during PM regeneration under stationary feed condition (b) Corresponding temporal deposited PM profiles 172 Figure 6.15 Impact of sudden (a) increase of feed oxygen concentration (b) decrease of filtration velocity (c) feed temperature on temporal temperature profiles. In all cases feed variable perturbation is initiated when temperature front reaches x/L = 0.5 174 xvi u

Figure 6.16 Moving temperature front profiles following a rapid shift in (a) O2 and filtration velocity (b) feed temperature and O2 (c) feed temperature and filtration velocity; PM loading of 10 g/L 176 Figure 6.17 Impact of step change feed conditions on the wrong-way behavior of PM combustion. Initial DPF temperature is (a). 700 K; (b). 923 K. Temperature is decreased when front reached x/L = 0.5 179 Figure 6.18 Moving temperature front profiles following changes of feed conditions over 30 s periods 180 Figure 6.19 (a) Dependence of the temperature rise on the position of the temperature front when the feed conditions are changed, (b) Moving temperature profile following a step-change of the feed 10 seconds after feed introduced to DPF. 182 Figure 6.20 (a) Moving temperature front profiles following a rapid shift in feed conditions; PM loading of 8 g/L; (b) Impact of soot loadings on the maximum temperature during stationary and transient states 183 XIX

List of Tables Table 1-1 US Tier 2 Emission Standards, Federal Test Procedure (FTP) 75, g/mi 3 Table 1-2 Material Candidates for Diesel Particulate Filters 6 Table 2-1 Merlin-Mid InSb MWIR Camera System Specifications 37 Table 6-1 Geometry and physical properties of the DPF and soot deposit used in the simulations 140 XX

Nomenclature Cp = specific heat capacity, J/ (kg-K) D = effective mass diffusivity, m /s d = hydraulic diameter of clean channel, m dj = hydraulic diameter of channel i, m dp = mean pore size, m E = activation energy, J/mol h = convective heat transfer coefficient, W/(m -K) AHco = specific heat of CO formation, J/mol AHco2 = specific heat of CO2 formation, J/mol k, = mass transfer coefficient of O2 in the channel i, m/s k0 = reaction rate frequency factor kox = reaction rate coefficient kp = permeability of soot deposit, m ks = permeability of ceramic filter, m2 L = length of ceramic filter, m M = molecular weight, kg/mol xxi

Pe = Peclet number P = exhaust gas pressure, Pa R = gas constant, m3-Pa/ (mol-K) Rsoot= reaction rate, mol/ (m -s) sp = specific area of soot deposit layer, m"1 Sh = Sherwood number T = Temperature, K t = time, s v = velocity, m w = thickness of soot deposit, m ws = thickness of ceramic filter wall, m x = distance, m y = oxygen concentration of the exhaust gas, mole fraction z = axial distance, m Greek letters a = index for the completeness of thermal oxidation xxii

cti = constant in channel pressure drop correlation AH = reaction heat, J/mol AP = backpressure, Pa Xp = soot deposit thermal conductivity, W/(m-K) Xs = ceramic filter thermal conductivity, W/(m-K) 8 = soot deposit porosity |i = exhaust gas viscosity p = density, kg/m3 Subscripts g = exhaust gas i = channel index, 1 = inlet channel 2 = outlet channel p = particulate layer s = substrate layer w = wall-outlet channel interface XXI11

Chapter 1 Introduction 1.1 Introduction Currently a diesel engine is the most efficient internal combustion engine and its use is widely expanding. The high efficiency and long durability of diesel engines have motivated their use in passenger cars, heavy-duty trucks and off-road vehicles throughout the world. As the diesel engine operates under very lean conditions even at full load, at which point the quantity of fuel injected per cycle is still about 50% lean of stoichiometric, less gaseous carbon monoxide and hydrocarbon emission are produced. ' One of the main disadvantages of diesel engines is the emission of particulate matter (PM), which consists of unburned carbon compounds. This emission is caused by local low temperatures where the fuel is not fully atomized. These local low temperatures occur at the cylinder walls and at the outside of large droplets of fuel. Compared to the overall fuel/air ratio which is lean, the fuel/air ratio at these areas where it is relatively cold is rich and less air is available at these areas to burn the fuel and some of the fuel turns into a carbon deposit.3"5 Once the particulate matter is emitted to the atmosphere it can penetrate deeply into the lung and is a potential occupational carcinogen of the human respiratory system, thus posing a serious threat to the public health.6'7 Because of this, several diesel vehicle emission standards have been developed in US and Europe to limit these emissions and those standards become more stringent in time. As early as 1991 the US Environmental Protection Agency (EPA) published the Tier 1 emission standards for the passenger cars and light-duty vehicles. This regulation was fully implemented in 1997. A more stringent numerical emission limits standards, 1

Tier 2, was introduced in 1999 and phased-in between 2004 and 2009. Under the Tier 2 regulation, the same emission limits applicable to light-duty emission standards have been applied to cover some of the heavier vehicle categories regardless of the fuel they use, which means vehicles fueled by gasoline, diesel, or alternative fuels must meet the same standards. Since light-duty emission standards are expressed in grams of pollutants per mile, vehicles with large engines such as light trucks or SUVs have to use more advanced emission control technologies than vehicles with smaller engines in order to meet the standards. Table 1.1 shows the Tier 2 standards for passenger and light duty vehicles. Tier 2 emission standards incorporate 8 permanent and 3 temporary certification levels of different stringency, called certification bins, and an average fleet standard for NOx emissions. For particular vehicles vehicle manufacturers have a choice to certify to any of the available bins. In the Tier 2 emission standards the average NOx emissions of the entire light-duty vehicle fleet sold by each manufacturer have to meet the average NOx standard of 0.07 g/mi. The temporary certification bins (bin 9, 10, and an MDPV bin 11) with more relaxed emission limits are only available in the phase-in period that has already expired since the 2008 model year. For Tier 1 a full useful life particulate matter emission standard was 0.10 g/mile. The current standards require 90% reduction of PM emission from the diesel engine exhaust. Since the excess sulfur in the fuel may poison the catalyst and lead to the failure of certain after treatment technologies, the EPA provided regulations concerning the sulfur level in the fuel. The current fuel regulation limits the sulfur content in diesel fuel in the U.S is 15 ppm. 2

Table 1-1 US Tier 2 Emission Standards, Federal Test Procedure (FTP) 75, g/mi Bin# Intermediate life (5 years / 50,000 mi) NMOG* CO NOx PM HCHO Full useful life NMOG* CO NOxf PM HCHO Temporary Bins 11 MDPVC i /-va,b,d,f qa,b,e,f 0.125 (0.160) 0.075 (0.140) 3.4 (4.4) 3.4 0.4 0.2 - - 0.015 (0.018) 0.015 0.280 0.156 (0.230) 0.090 (0.180) 7.3 4.2 (6.4) 4.2 0.9 0.6 0.3 0.12 0.08 0.06 0.032 0.018 (0.027) 0.018 Permanent Bins 8b 7 6 5 4 3 2 1 0.100 (0.125) 0.075 0.075 0.075 - - - - 3.4 3.4 3.4 3.4 - - - - 0.14 0.11 0.08 0.05 - - - - - - - - - - - - 0.015 0.015 0.015 0.015 - - - - 0.125 (0.156) 0.090 0.090 0.090 0.070 0.055 0.010 0.000 4.2 4.2 4.2 4.2 2.1 2.1 2.1 0.0 0.20 0.15 0.10 0.07 0.04 0.03 0.02 0.00 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.00 0.018 0.018 0.018 0.018 0.011 0.011 0.004 0.000 * for diesel fueled vehicle, NMOG (non-methane organic gases) means NMHC (non- methane hydrocarbons) f average manufacturer fleet NOx standard is 0.07 g/mi for Tier 2 vehicles a - Bin deleted at end of 2006 model year (2008 for HLDTs) b - The higher temporary NMOG, CO and HCHO values apply only to HLDTs (Heavy Light-Duty Truck ) and MDPVs (Medium-Duty Passenger Vehicle) and expire after 2008 c - An additional temporary bin restricted to MDPVs, expires after model year 2008 d - Optional temporary NMOG standard of 0.195 g/mi (50,000) and 0.280 g/mi (full useful life) applies for qualifying LDT4s and MDPVs only e - Optional temporary NMOG standard of 0.100 g/mi (50,000) and 0.130 g/mi (full useful life) applies for qualifying LDT2s only f - 50,000 mile standard optional for diesels certified to bins 9 or 10 3

1.2 Diesel Particulate Filter The filter for particulate matter control has been introduced on a large scale worldwide during early 2000s. Currently the Diesel Particulate Filter (DPF) is the most efficient device for PM removal from the engine effluents.8"11 It consists of thousands of square parallel channels, with the opposite ends of adjacent channels being plugged. Figure 1.1 is a schematic of DPF channels. The exhaust gas passes through the filter porous walls into the adjacent channels, while more than 95% of the PM accumulates on the filter.12 Such kind of honeycomb wall flow filter was made from extrusion of ceramic honeycomb catalyst carriers with ceramic plugging of the opposite end of adjacent channels. Usually the DPF with 200 or 300 cells per square inch (cpsi) has a wall 1 ^ thickness of about 350 to 400 um. The average filter pore size and pore volume vary depending on the regeneration strategy. Typical values of porosity for filters using a fuel- borne regeneration additive are between 45 and 50%) while the mean pore size is about 10 um. Filter substrates used for washcoat/catalyst coating have a porosity of 65% and 20 um mean pore size. The overall size of the filter depends on the exhaust volume of the engine as it is essential to keep the backpressure low. Typical filter dimensions for filter with cell density of 200 cpsi for passenger cars are 144 mm (5.66 inch) diameter and 152 mm (6 inch) length, almost 2.5 L in total volume, with a total filtration area of 1.9 m2. This leads to an average filtration velocity of some cm/s.14 The total weight for a 200 cpsi DPF is about 1.8 kg made from Silicon carbide or 1.2 kg made from Cordierite. A picture of a commercial available DPF made from Silicon carbide is shown in Figure 1.2. 4

Figure 1 1 The schematic of DPF channels Figure 1 2 The commercial available DPF 5

The ideal DPF material should have high temperature stability, resistance against thermal strain, corrosion stability and sufficient mechanical strength to fulfill all the requirements on filtration, regeneration and application. Another important consideration is the availability of such a material for mass production at low price and low weight. Table 1.2 compares some material candidates for diesel particulate filter.15 Table 1 -2 Material Candidates for Diesel Particulate Filters Material Density(g/cm3) Thermal conductivity (RT) (W/mK) CTE20-1000°C(10"61/K) Young's modulus (GPa) Thermal limit for application (air) (°C) Corrosion resistance * Price * Cordierite 2.1 1-3 0.9-2.5 130 1350 - ++ SiC 3.1- 3.2 90 4.7- 5.2 410 1500 + - Silicon 2.33 120 4.4 110 1350 0 — Mullite 2.9 4-5 4.4 150 1600 0 + Al- Titanate 3.3 1.5-3 -0.5-3 20 1500 0 + FeCr Ni 8.1 14 17 200 1250 - — *These properties cannot be compared directly. A rough estimation gives: ++, very good; +, good; o, ambient; --, negative; --, very negative. The silicon in this table is not used alone but in combination with silicon carbide. Cordierite and silicon carbide are the two most commonly used filter materials. Cordierite is a synthetic ceramic, made from kaolin and talc. It was originally used as a flow through catalyst substrate and later adapted for the filter application. Besides the low price advantage the cordierite has a very low coefficient of thermal expansion (CTE). It can be optimized to be up to 0.4x10"6 1/K parallel to the extrusion direction, which diminishes thermal stresses during regeneration. The main problem associated with

Full document contains 235 pages
Abstract: A major technological challenge in the operation of diesel particulate filters (DPFs) is prevention of occasional melting of the ceramic filters during regeneration. The dynamic features of the combustion of particulate matter (PM) on a single layer diesel particulate filter were studied using IR imaging. The experiments showed that at stationary feed conditions the soot combustion may proceed in three different modes: either by a single moving hot zone or several hot zones generated at different ignition points or uniform combustion all over the surface. The maximum temperature of the moving fronts was much higher than those attained during uniform combustion. The highest temperature attained under stationary (constant feed) combustion is too low to cause the filter melting (melting temperature ∼1250°C). We conjecture that high temperature excursions are a counter-intuitive response to a rapid deceleration which decreases the exhaust gas temperature and flow-rate and increases the oxygen concentration. The experiments showed that a step-change of the feed temperature led to a transient temperature exceeding the highest attained under stationary operation with the initial feed temperature. A simultaneous step-change of the feed temperature, flow rate and oxygen concentration in the feed led to a transient temperature that exceeded the highest attained for stationary operation under either the initial or final operation conditions. It was also higher than those generated either by a step change of any single feed input or by a step-change of any pair of feed inputs. The temperature rise depended in a complex way on several factors, such as the direction of movement of the propagating front, the location of the front when the feed was step-changed, and whether the step-change was done either before or after formation of a moving temperature front. The simulations provided insight about the dependence of the amplitude of the temperature rise on the step change of the operating variables. The understanding generated by the simulations should help develop operation and control protocols that circumvent or at least decrease the probability of the occurrence of the destructive melting of the DPF.