Modeling of Diesel Exhaust Systems:

A methodology to better simulate soot reactivity

October 6th, 2011

Fabien Rioult, BASF

Dennis Anderson, BASF

DEER 2011 2

Outline

Modeling Capability

Application Example with Soot Regeneration Simulation

Improving the Fidelity of Systems Simulation

Customizing Soot Models

Conclusions and Path Forward

DEER 2011 3

Modeling Simulation Capability at BASF

Philosophy of Modeling at BASF : •Provide a theoretical foundation for catalyst submissions •Reduce amount of experimental work needed •Provide lookup tables to simplify customer calibration work •Provide capability to simulate exhaust system response on future engines

Optimizer

DEER 2011 4

CatSim Capabilities

CatSim incorporates individual 0-D, 1-D and/or 1-D + 1-D models for:

Monolithic flow through catalysts (TWC, DOC, SCR, AMOX, LNT)

Monolithic based filters (CSF, SCRoF)

Pipes and cones

Injectors with feedback control via Matlab or Excel

Cold EGR

Junctions

Each catalyst technology defined by washcoat properties and reaction kinetics

Additional features include:

Homogeneous reactions

Multiprocessor capable

DEER 2011 5

Filter Model: Processes Captured

CO, HC, NO, NH3,…

P, T, F

CO2, H2O, N2, NO2, …

Heat and Mass Transfer

Thermal losses to ambient

Axial conduction

Thermal conduction to can

Reactions

Soot Accumulation

Soot Regeneration

Backdiffusion of gases

DEER 2011 6

Example of Application with Soot Regeneration Simulation

DEER 2011 7

Filter Regeneration Simulation

CSF-out

0

100

200

300

400

500

600

700

0 200 400 600 800 1000 1200 1400Time (sec)

Tem

pera

ture

(°C

) ; C

once

ntra

tions

(p

pm)

NO measured NO2 measured THC measuredTemp CSFout measured NO simulated NO2 simulatedTHC simulated Temp CSFout simulated

DEER 2011 8

0

1

2

3

4

5

6

0 200 400 600 800 1000 1200 1400 1600Time (sec)

Soot

load

(g/l)

200

250

300

350

400

450

500

550

600

650

700

Tem

pera

ture

(C) -

in re

ar o

f CSF

soot load (1x injection)soot load (1.1x injection)soot load (1.2x injectioni)Temperature (1x injection)Temperature (1.1x injection)Temperature (1.2x injection)

Soot properties are different from one engine to another

Soot properties are different from one mode to another

The regeneration was estimated based on oxidation kinetics of soot produced in BASF’s lab. The soot generated by this engine might have different physical properties leading to different reactivity.

Filter Regeneration Simulation (2)

DEER 2011 9

Improving the Fidelity of Systems Simulation Customizing soot model

DEER 2011 10

Soot Particulates

Particulate composition and morphology depend on the combustion process: Different engines will generate different soot Even different mode can generate different soot Composition and morphology will impact reactivity with O2 and NO2

bronchus

lung nose

= Particle size distribution in diesel exhaust

100 nm 50 nm

Transmission Electron Microscopy (TEM) of diesel soot particulates

C50

1.7%2.3%

86.8%

3.1% 4.1%

2.0%

6.1%

SO4H2OSootNO3SOF-FuelSOF-Lube

Example of soot composition

[1] K. Al-Qurashi et A. Boehman, "Impact of exhaust gas recirculation (EGR) on the oxidative reactivity of diesel engine soot ", Combustion and Flame, 155, 675-695 (2008). [2] Juhun Song et al., "Impact of alternative fuels on soot properties and DPF regeneration”, Combust. Sci. and Tech., 179, 1991-2037 (2007). [3] R. Vander Wal et al., “HRTEM study of diesel soot collected from diesel particulate filters”, Carbon, 45, 70-77 (2007) [4] J. Rodriguez-Fernandez et al., “Characterization of the diesel soot oxidation process through optimized thermogravimetric method”, Energy and Fuels, 25, 2039-2048 (2011)

See Also:

DEER 2011 11

Soot Reactivity and Surface Area (BET)

0

100

200

300

400

500

600

700

800

Soot A Soot B Soot C Soot D Soot E Soot F

BET

SA

(m2/

g)

BET1 m2/gBET2 m2/g

Soot A presents the lowest surface area but the highest reactivity

TGA: mass loss during temperature ramp in air

DEER 2011 12

Soot Reactivity and Surface Oxygen (XPS)

Soot A presents more surface oxygen than other samples, which could be the cause of highest reactivity

TGA : mass loss during temperature ramp in air

Oxygen 1s Overlay Soot A Soot C Soot E

DEER 2011 13

Customizing Soot Model Process

Test Reactivity of Soot Samples

and Compare Kinetics

by TGA

Modify Reference Soot Model to Create New Soot Model

Determine Soot Oxidation

Kinetics on Reactor

Create Soot Model for

Reference Soot

Collect Reference Soot on Engine on Filter Cores

New Soot Sample

DEER 2011 14

Collection of Soot on Filter Cores Collect Reference

Soot on Engine On Filter Cores

DOC

Filter Cores

Monolith with plugged cells

DEER 2011 15

Test Protocol

Create Soot Model for

Reference Soot

Determine Soot Oxidation

Kinetics on Reactor

0

100

200

300

400

500

600

700

800

900

1000

0 500 1000 1500 2000Time (sec)

NO

2 (p

pm)

0

2

4

6

8

10

12

O2

(%)

NO2 ppmO2 %

0

100

200

300

400

500

600

700

800

900

1000

0 500 1000 1500 2000 2500Time (sec)

NO

2 (p

pm)

0

2

4

6

8

10

12

14

O2

(%)

NO2 ppmO2 %

Temperatures ( C): {200, 250, 300, 350, 400, 450}

Temperatures ( C): {500, 550, 600, 650}

Soot loading of the filter was determined by C balance (CO and CO2 produced)

DEER 2011 16

Reactor Results Used for Calibration of the Soot Model

Create Soot Model for

Reference Soot

Determine Soot Oxidation

Kinetics on Reactor

200 C 300 C

400 C 500 C

DEER 2011 17

Parity Plots Measured Vs. Calculated Concentrations

0

100

200

300

400

500

600

700

800

900

1000

0 200 400 600 800 1000Measured CO2 (ppm)

Cal

cula

ted

CO

2 (p

pm)

200C 250C 300C350C 400C 450C500C 550C 600C

0

100

200

300

400

500

600

700

800

900

1000

0 200 400 600 800 1000Measured CO (ppm)

Cal

cula

ted

CO

(ppm

)

200C 250C 300C350C 400C 450C500C 550C 600C

0

100

200

300

400

500

600

700

800

900

1000

0 200 400 600 800 1000Measured NO2 (ppm)

Cal

cula

ted

NO

2 (p

pm)

200C 250C 300C

350C 400C 450C

500C 550C 600C

0

100

200

300

400

500

600

700

800

900

1000

0 200 400 600 800 1000Measured NO (ppm)

Cal

cula

ted

NO

(ppm

)

200C 250C 300C

350C 400C 450C

500C 550C 600C

Create Soot Model for

Reference Soot

CO2

NO

CO

NO2

DEER 2011 18

Soot Model

C + O2 CO2

C + ½ O2 CO

C + 2 NO2 CO2 + 2 NO

C + NO2 CO + NO

C + NO2 + ½ O2 CO2 + NO

Create Soot Model for

Reference Soot

DEER 2011 19

Model Vs. Engine Data The soot model correctly simulates soot regeneration, NO2 consumption and CO formation Assumption: E/O soot rate was constant during the tests

DPF-out Emissions – C50 DPF-out Emissions – C100

C50 Temperature: 320 C Flow rate: 970 kg/h NOx: ~405 ppm

C100 Temperature: 434 C Flow rate: 940 kg/h NOx: ~ 370 ppm

model Exp. data

model Exp. data

DEER 2011 20

Thermo-Gravimetric Analysis (TGA) Weight Loss Due to O2 Oxidation

New Soot Reference Soot

Test Reactivity of Soot samples

and Compare Kinetics

by TGA

Ram

p-up in N2

Ram

p-up in N2

In Air In Air

DEER 2011 21

Thermo-Gravimetric Analysis (TGA) Weight Loss Due to NO2 Oxidation

New Soot Reference Soot

Test Reactivity of Soot samples

and Compare Kinetics

by TGA

Ram

p-up in N2

Ram

p-up in N2

In 2% NO2 / 6% O2 In 2% NO2 / 6% O2

DEER 2011 22

Thermo-Gravimetric Analysis Kinetic Constants Comparison

•Kinetics information obtained by TGA is used to modify the reference soot model (the relative difference in pre-exponential coefficients is the information used to modify the reference soot model and create a model for the “new soot”). •This allows a better estimation of regeneration durations depending on the conditions (temperature, flow, NO2…) and interaction with deNOx system (ex: NO2 at SCR-in).

Modify Reference Soot Model to

Create Customer’s Soot Model

y = -65.11x + 8.41R2 = 1.00

y = -59.84x + 8.10R2 = 0.99

y = -138.78x + 17.08R2 = 1.00

y = -183.92x + 22.07R2 = 1.00

-8

-7

-6

-5

-4

-3

-2

-1

00.13 0.15 0.17 0.19 0.21 0.23 0.25

1000/RT

Ln(k

)

New Soot - O2 New Soot - NO2Reference Soot - O2 Reference Soot - NO2Rerun Reference Soot - O2 Rerun Reference Soot - NO2

Simulated Passive Regeneration E/O 350°C, 500 ppm NO, 13 %O2

DOC +CSF

0

1

2

3

4

0 2 4 6 8 10Time (h)

Soot

Loa

ding

(g/l)

050100150200250300350

Filte

r-out

NO

2 (p

pm)

Soot Loading - Reference Soot Soot Loading - New SootFilter-out NO2 - Reference soot Filter out NO2 - New Soot

DEER 2011 23

Conclusions / Path Forward

Conclusions

A soot model was developed for soot oxidation on reactor and used as a reference.

– This model allowed a good prediction of engine data

A methodology to simulate soot from different origins was proposed.

– use a different soot sample from another origin (different engine) and compare its reactivity to the reference soot sample by solely using thermo-gravimetric analysis. The relative difference of reactivity was used to create a model of this “new soot” sample.

Path Forward

Final validation: compare predictions from this soot model generated with this methodology to engine data.

DEER 2011 24

Thank You For your attention