PM Sensor Development and Simulation for Diesel

Renewable energies | Eco-friendly production | Innovative transport | Eco-efficient processes | Sustainable resources

PM Sensor Development and Simulation for Diesel Particulate Filter

On Board Diagnostic CLEERS Workshop 2013Dearborn, April 10-12th, 2013

J. Lavy, C.N. Millet, Y. Creff (IFP Energies nouvelles)

F. Duault, S. Raquin (EFI Automotive)

Ph. Breuil, J.P. Viricelle (Ecole des Mines de St Etienne)

CLEERS / 2013 / J Lavy2

©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outline

Context and objectives PM sensor basic approaches PM sensor design optimization PM sensor response characterization and analysis Vehicle evaluation on chassis dyno PM sensor response modeling Conclusion and outlook


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outline



©20

13 -

IFP

Ene

rgie

s no

uvel

les

Context and objectives

DPF now generalized to cope DPF now generalized to cope DPF now generalized to cope DPF now generalized to cope with more and more stringent with more and more stringent with more and more stringent with more and more stringent emission standards, especially emission standards, especially emission standards, especially emission standards, especially for onfor onfor onfor on----road vehicle applicationsroad vehicle applicationsroad vehicle applicationsroad vehicle applications

PM10 annual mean daily value (µg/m 3) - 2012

WHO guidelines : 20 µg/m3EU regulation limits : 40 µg/m 3

Air quality standards in many Air quality standards in many Air quality standards in many Air quality standards in many countries aim at reducing countries aim at reducing countries aim at reducing countries aim at reducing population exposure to air population exposure to air population exposure to air population exposure to air pollutantspollutantspollutantspollutants PMPMPMPM10101010, PM, PM, PM, PM25252525, SO, SO, SO, SO2222, NO, NO, NO, NO2222, CO, O, CO, O, CO, O, CO, O3333


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Context and objectivesEurope LDV

PM mass certification and EOBD threshold limits (mg/km)

020406080

100120140160180200

Euro

3Eur

o 4

Euro

5 a

(200

9)

Euro

5 b

(201

1)

Euro

6.1

(201

4)

Euro

6.2

(201

7)

EOBD

Certification

US LDV

PM OTL/Certification limitratio

0

1

2

3

4

5

6

2004

-200

9

2010

-201

2

2013

+

Similar trend for HDVNeed for accurate DPF filtration efficiency diagnostic


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Context and objectives Today, detection of DPF failure is based on differential Today, detection of DPF failure is based on differential Today, detection of DPF failure is based on differential Today, detection of DPF failure is based on differential

pressure sensor technologypressure sensor technologypressure sensor technologypressure sensor technology Limited sensitivity: only severe DPF failures can be detectedLimited sensitivity: only severe DPF failures can be detectedLimited sensitivity: only severe DPF failures can be detectedLimited sensitivity: only severe DPF failures can be detected Technology not able to meet future stringent EOBD threshold limiTechnology not able to meet future stringent EOBD threshold limiTechnology not able to meet future stringent EOBD threshold limiTechnology not able to meet future stringent EOBD threshold limitstststs

Need for a more sensitive technology such as an onononon----board board board board PM sensorPM sensorPM sensorPM sensor downstream of the DPFdownstream of the DPFdownstream of the DPFdownstream of the DPF

Various PM sensor technologies under developmentVarious PM sensor technologies under developmentVarious PM sensor technologies under developmentVarious PM sensor technologies under development Continuous and cumulative measurement concepts Continuous and cumulative measurement concepts Continuous and cumulative measurement concepts Continuous and cumulative measurement concepts


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outline



©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM sensor basic approaches:Spark discharge

Source : Journal of Engineering for Gas Turbine and Power / March 2009, Vol 131 / Allen et al

Source : SAE 2006-05-0285 / Gheorghiu et al Source : SAE 2012 OBD Symposium /

Stuttgart / Gheorghiu

PM concentration related to changes in the voltage waveform of a repetitive, low energy spark discharge

Cross-sensitivities to check High voltage to manage


©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM sensor basic approaches:Electrochemical polarization Produced by a difference in O2 partial pressures

between 2 electrodes, which is due to PM deposit and oxidation on the anode

Source : SAE 2011-01-2059 / Yoshihara et al. Very localized phenomenon


©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM sensor basic approaches:The electrostatic concept: image charge Measurement of the current emitted by the inherent

PM electrical charge

Source : ETH Conference 2004 / Kittelson et al / University of Minnesota

Poor response signal due to the globally neutral particulate charge within the exhaust pipe


©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM sensor basic approaches:The electrostatic concept: corona discharge

Source : DEER 2009 / Hall et al

The corona discharge increases PM electrical charge

High voltage management necessary

Source : SAE 2011-01-0626 / Zamaras et al


©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM sensor basic approaches :Electrostatic capacity Detection of PM accumulation related to changes in

electrostatic capacity

Source : SAE 2011-01-0302 / Kono et al.

High voltage to manage


©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM sensor basic approaches:The resistive concept Resistance decreases as conductive carbonaceous

particulates accumulate between the electrodes

Source : IFPEN patent FR2760531

Time

Resistance (Ohm)

Concept widely studied due to simplicity and low cost


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outline



©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM sensor design optimizationFrom the first prototype in 2000

1 cm

Source : SAE 2000-01-0472 / Bouchez et al

Glow plug for soot burning

Two electrodes for resistance measurement

Prototype developed for concept validation


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Sensitive element

Interdigitalelectrodes for

resistance measurement

Heater for

soot burning

(regeneration)

PM sensor design optimizationto a pre-industrial version today

Packaging & control unit


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Objectives:Objectives:Objectives:Objectives: homogeneous PM depositionhomogeneous PM depositionhomogeneous PM depositionhomogeneous PM deposition avoid deposition of large particlesavoid deposition of large particlesavoid deposition of large particlesavoid deposition of large particles limited heat exchanges for efficient and fast soot burninglimited heat exchanges for efficient and fast soot burninglimited heat exchanges for efficient and fast soot burninglimited heat exchanges for efficient and fast soot burning

PM sensor design optimization3D CFD of the flow around sensor shield


©20

13 -

IFP

Ene

rgie

s no

uvel

les

K-Epsilon RNG turbulence model for fluid Lagrangian representation for particles

Based on spray liquid injection models (no evaporation… Turbulent dispersion also taken into account for particles Simultaneous injection of different particle sizes

255 000 cells and 268 000 nodes 64 processors 25 to100 h (depending on flow conditions…



©20

13 -

IFP

Ene

rgie

s no

uvel

les



©20

13 -

IFP

Ene

rgie

s no

uvel

les


PM deposited on ΣsDeposition ratio =

PM flowing through Σin

Σs

Σin


©20

13 -

IFP

Ene

rgie

s no

uvel

les

0,00

2,00

4,00

6,00

8,00

10,00

1 10 100 1000 10000

Particles size Particles size Particles size Particles size ((((nmnmnmnm…………

Dep

ositi

on ra

tio (%

…D

epos

ition

ratio

(%…

Dep

ositi

on ra

tio (%

…D

epos

ition

ratio

(%… With turbulent dispersion

W/o turbulent dispersion

PM sensor design optimization3D CFD of the flow around sensor shield Turbulent dispersion and particles inertia effects

Particles above 5 µm deposited by inertial impaction

Particles below 5 µm deposited by turbulent dispersion

Typical Diesel Particulate sizedistribution


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Reference design : large particles impact a lot

Config #A : no impact of large particles low impact of smaller

particles

Config #B : no impact of large particles impact of smaller particles

remains significantBest experimental results with Config#B

0,00

2,00

4,00

6,00

8,00

10,00

1 10 100 1000 10000

Particles size Particles size Particles size Particles size ((((nmnmnmnm…………

Dep

ositi

on ra

tio (%

…D

epos

ition

ratio

(%…

Dep

ositi

on ra

tio (%

…D

epos

ition

ratio

(%… Reference

Config #AConfig #B

PM sensor design optimization3D CFD of the flow around sensor shield Various configurations tested


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outline



©20

13 -

IFP

Ene

rgie

s no

uvel

les

Heater

Particulatespreading out

system

PM sensor

Particulatetank

PM sensor response characterization and analysisDevelopment of a specific PM test bench rig

ΦΦΦΦ: 39 mm Flow rate: 0 350 m3/h

Flow velocity: 0 80 m/sTemperature: 20°°°°C 450 °C

PM concentration: 0 250 mg/m3

Particulate mean diameter (nm)

120165

240

10

100

1000

Engineout

Enginesoot

Carbonblack

Particulate test rig

Eng

ine


©20

13 -

IFP

Ene

rgie

s no

uvel

les

BIP - capteur L18P3D5 - 150 °C - 34.5 m/s - suies m oteur - effet concentration PM

100

1000

10000

100000

0 100 200 300 400 500 600

temps (s)

rési

stan

ce fi

ltrée

cap

teur

(kO

hm)

PM 11 mg/m3 PM 4 mg/m3

time (s)

PM

sen

sor

resi

stan

ce(k

Ω) 150°C / 35 m/s

PM : 11 mg/m3 PM : 4 mg/m3

110 s 430 s

PM sensor response characterization and analysis

Loading time (tload)


©20

13 -

IFP

Ene

rgie

s no

uvel

les

26


0

100

200

300

400

500

600

700

0,0 1,0 2,0 3,0 4,0 5,0 6,0 7,0 8,0 9,0 10,0

PM flow rate (mg/s)

Tim

e to

rea

ch 1

Moh

m (

s)

Particulate test rig feeded by engine soot

Engine test rig

Particulate test rig fed by engine soot

PM flow rate (mg/s)

Load

ing

time

t load

(s)

Comparison between particulate and engine test rigs results

Similar results

Particulate test rig representative of engine conditions and phenomena


©20

13 -

IFP

Ene

rgie

s no

uvel

les

27

0

20

40

60

80

100

120

0 10 20 30 40

Electrode polarization (V)

Load

ing

time

(s)

Polarization favors formation of particle bridges

Better deposition rate ? (to be confirmed)


Fixed flow rate and soot concentration

electrode

electrode

PM sensor sensitivity enhanced by electrode polarization


©20

13 -

IFP

Ene

rgie

s no

uvel

les

BIP - capteur L18P3D5 - 150 °C - 34.5 m/s - suies m oteur - effet concentration PM

100

1000

10000

100000

0 100 200 300 400 500 600

temps (s)

rési

stan

ce fi

ltrée

cap

teur

(kO

hm)

PM 11 mg/m3 PM 4 mg/m3

Time (s)

sens

orre

sist

ance

(kΩ

)

150°C / 35 m/s

PM : 11 mg/m3 PM : 4 mg/m3

tload


Pipe section

Qsoot( g/mm2)

Qsoot should be a constant value in perfect conditions


©20

13 -

IFP

Ene

rgie

s no

uvel

les

29

flow velocity (m/s)

Particulatetest rig

flow velocity (m/s)

EngineBench

PM sensor response characterization and analysis Test repartition vs PM, flow velocity and temperature


©20

13 -

IFP

Ene

rgie

s no

uvel

les

30


Similar tendency whatever the test benchDominant effect of flow velocity and soot concentration

on sensor response

0

20

40

60

80

100

120

0 10 20 30 40 50 60

PM (mg/m3 )Q

soot

(µg/

mm

²)

8-18 m/s

19-33 m/s

37-46 m/s

0

20

40

60

80

100

120

0 10 20 30 40 50 60

PM (mg/m3 )

Qso

ot (µ

g/m

m²)

8-18 m/s

19-33 m/s

37-46 m/s

0 10 20 30 40 50 60

vitesse 9 m/svitesse 34 m/svitesse 60 m/s

velocityvelocity

velocity


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outline



©20

13 -

IFP

Ene

rgie

s no

uvel

les

Vehicle evaluation on chassis dyno

Citroën C4 - €4 DV6 engine

Drilled DPF to simulate different failure levels

EXXOtestCAN logger

IXXAT CANbridge

Instantaneous PM concentration measurement by AVL 483 (MSS)


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Slightly damaged DPFPM 4.6 mg/km

Vehicle evaluation on chassis dynoNo DPF

PM 15.2 mg/km

PM sensor resistance (kΩ)

Vehicle speed (km/h)

PM (mg/m3)

4 successive hot NEDC driving cycles


©20

13 -

IFP

Ene

rgie

s no

uvel

les

0

0,5

1

1,5

2

2,5

3

3,5

4

4,5

5

0 5 10 15 20

PM mass NEDC (mg/km)

Pm

sen

sor

load

ings

/ N

ED

C

Vehicle evaluation on chassis dyno

PM sensor loading frequency proportional to PM mass

PM sensor able to detect low PM level (~ 4 loadings during NEDC @ 12 mg/km)

PM sensors loading frequency function of the PM mass

12


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outline



©20

13 -

IFP

Ene

rgie

s no

uvel

les

WhyWhyWhyWhy To analyze the effects of flow and PM sensor design To analyze the effects of flow and PM sensor design To analyze the effects of flow and PM sensor design To analyze the effects of flow and PM sensor design

parametersparametersparametersparameters To be used in modelTo be used in modelTo be used in modelTo be used in model----based DPF onbased DPF onbased DPF onbased DPF on----board diagnostic board diagnostic board diagnostic board diagnostic

algorithms (both modelalgorithms (both modelalgorithms (both modelalgorithms (both model----based and non modelbased and non modelbased and non modelbased and non model----based based based based diagnostics algorithms developed at IFPEN…diagnostics algorithms developed at IFPEN…diagnostics algorithms developed at IFPEN…diagnostics algorithms developed at IFPEN…

How: by coupling two subHow: by coupling two subHow: by coupling two subHow: by coupling two sub----modelsmodelsmodelsmodels PM deposition on the sensing zonePM deposition on the sensing zonePM deposition on the sensing zonePM deposition on the sensing zone Resistive response according to PM quantity deposited over Resistive response according to PM quantity deposited over Resistive response according to PM quantity deposited over Resistive response according to PM quantity deposited over

the sensing zonethe sensing zonethe sensing zonethe sensing zone

PM sensor response modeling


©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM depositionsub-model

Soot depositedquantity

qsoot (t)

Resistive responsesub-model

qsoot

Res

ista

nce

PM sensor response modelingExhaust pipe ØFlow velocity V(t)PM concentration C(t)Temperature T °°°°(t)

Qsoot (t)

time

Res

ista

nce

Rs(t)


©20

13 -

IFP

Ene

rgie

s no

uvel

les

PM deposition ratio (Dr… is a function of flow velocity and PM concentration

0

2

4

6

8

10

0 10 20 30 40 50 60PM (mg/m3)

PM

dep

ositi

on r

atio

(‰

)

Model - 9m/sModel - 34 m/sModel - 60 m/sExperimental - 9 m/s

Experimental - 34 m/sExperimental - 60 m/s

)....1.( 3210 CVACAVAADr +++=

PM sensor response modelingPM deposition ratio sub-model

A1, A2, A3 calibrated from stabilized test results

A0 calibrated from either stabilized or transient test results


©20

13 -

IFP

Ene

rgie

s no

uvel

les

∑+= is RRR /1/1/1 0

PM sensor response modelingResistive response sub-model

Particles in a bridge are series resistors and bridges are resistors connected in parallel

l 0R

Ri

Ri²..

..4

Dm

i Nb

lR

πρ=

PM: mean diameter: resistivity (no specific data available, calibrated

parameter from either stady state or transient test results)ρ

mD


©20

13 -

IFP

Ene

rgie

s no

uvel

les


A0 and calibrated from this reference test PM sensor resistance – experimental and modeling (kΩ)

Vehicle speed (km/h)

PM (mg/m3)

No DPF – Hot successive NEDC - PM 15.2 mg/km

Good accordance with experimental data


©20

13 -

IFP

Ene

rgie

s no

uvel

les


0

1

2

3

4

5

6

0 5 10 15 20

PM mass NEDC (mg/km)

Pm

sen

sor

load

ings

/ N

ED

C

Exp.Model

Application of the model for various DPF failure levels

Accurate prediction of the PM sensor loading frequency whatever the DPF failure level

Model evaluation for others driving cycles to be done


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Conclusion

3D3D3D3D----CFD simulationCFD simulationCFD simulationCFD simulation to better understand the particle deposition processesto better understand the particle deposition processesto better understand the particle deposition processesto better understand the particle deposition processes to optimize the sensor collecting tip designto optimize the sensor collecting tip designto optimize the sensor collecting tip designto optimize the sensor collecting tip design

Development of a specific particulate test rigDevelopment of a specific particulate test rigDevelopment of a specific particulate test rigDevelopment of a specific particulate test rig easy and independent control of flow velocity, temperature, PM easy and independent control of flow velocity, temperature, PM easy and independent control of flow velocity, temperature, PM easy and independent control of flow velocity, temperature, PM

concentration and nature (synthetic or engine soot…concentration and nature (synthetic or engine soot…concentration and nature (synthetic or engine soot…concentration and nature (synthetic or engine soot… sensor response analysis and results in accordance with engine tsensor response analysis and results in accordance with engine tsensor response analysis and results in accordance with engine tsensor response analysis and results in accordance with engine testsestsestsests

Engine test benches and vehiclesEngine test benches and vehiclesEngine test benches and vehiclesEngine test benches and vehicles sensitivity validation in steadysensitivity validation in steadysensitivity validation in steadysensitivity validation in steady----state and transient conditionsstate and transient conditionsstate and transient conditionsstate and transient conditions

Tools used to develop a new resistive PM sensor


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Conclusion

High sensitivity to low PM levels, complying with the 12 High sensitivity to low PM levels, complying with the 12 High sensitivity to low PM levels, complying with the 12 High sensitivity to low PM levels, complying with the 12 mg/km European OBD threshold limit (Euro 62 in 2017…mg/km European OBD threshold limit (Euro 62 in 2017…mg/km European OBD threshold limit (Euro 62 in 2017…mg/km European OBD threshold limit (Euro 62 in 2017…

Nearly continuous DPF monitoring despite a basic Nearly continuous DPF monitoring despite a basic Nearly continuous DPF monitoring despite a basic Nearly continuous DPF monitoring despite a basic "cumulative process""cumulative process""cumulative process""cumulative process"

Model of PM deposition rate and sensor resistance Model of PM deposition rate and sensor resistance Model of PM deposition rate and sensor resistance Model of PM deposition rate and sensor resistance response developed and validated in both steady state response developed and validated in both steady state response developed and validated in both steady state response developed and validated in both steady state and transient (NEDC cycle… conditionsand transient (NEDC cycle… conditionsand transient (NEDC cycle… conditionsand transient (NEDC cycle… conditions

This on-board PM sensor demonstrated its strong ability to detect DPF malfunction or failure as required by the future OBD standards


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Outlook Validation of DPF failure diagnostic algorithm in real life Validation of DPF failure diagnostic algorithm in real life Validation of DPF failure diagnostic algorithm in real life Validation of DPF failure diagnostic algorithm in real life

conditionsconditionsconditionsconditions SAE paper 2013SAE paper 2013SAE paper 2013SAE paper 2013----01010101----1334 to be presented next week at SAE 1334 to be presented next week at SAE 1334 to be presented next week at SAE 1334 to be presented next week at SAE

World CongressWorld CongressWorld CongressWorld Congress Durability tests under wayDurability tests under wayDurability tests under wayDurability tests under way

aging: 600 h, 2400 regenerations achieved so far on an engineaging: 600 h, 2400 regenerations achieved so far on an engineaging: 600 h, 2400 regenerations achieved so far on an engineaging: 600 h, 2400 regenerations achieved so far on an engine poisoning: from fuel and lubricant additivespoisoning: from fuel and lubricant additivespoisoning: from fuel and lubricant additivespoisoning: from fuel and lubricant additives

Evaluation of PM sensor response to particulate Evaluation of PM sensor response to particulate Evaluation of PM sensor response to particulate Evaluation of PM sensor response to particulate numbernumbernumbernumber Diesel engineDiesel engineDiesel engineDiesel engine GDI engine (GPF developed to comply with future PN GDI engine (GPF developed to comply with future PN GDI engine (GPF developed to comply with future PN GDI engine (GPF developed to comply with future PN

legislation…legislation…legislation…legislation…


©20

13 -

IFP

Ene

rgie

s no

uvel

les

Acknowledgements To coTo coTo coTo co----authorsauthorsauthorsauthors To French authorities for financial support of To French authorities for financial support of To French authorities for financial support of To French authorities for financial support of

"CICLAMEN 1&2" projects"CICLAMEN 1&2" projects"CICLAMEN 1&2" projects"CICLAMEN 1&2" projects

To "CICLAMEN 1&2" project partnersTo "CICLAMEN 1&2" project partnersTo "CICLAMEN 1&2" project partnersTo "CICLAMEN 1&2" project partners

Renewable energies | Eco-friendly production | Innovative transport | Eco-efficient processes | Sustainable resources

www.ifpenergiesnouvelles.com

Thank you for your attentionContact : [email protected]

Documents

PM Sensor Development and Simulation for Diesel