International journal of automotive technology, tập 12, số 3, 2011

KIM 3 Incheon 403-714, Korea Daejeon 305-343, Korea Received 24 March 2009; Revised 8 September 2010 ABSTRACT− The effects of split injection, oxygen enriched air, and heavy exhaust gas

International Journal of Automotive Technology , Vol 12, No 3, pp 315 − 320 (2011)

315

EVALUATION OF IDLE STABILITY THROUGH IN-SITU TORQUE MEASUREMENT IN AUTOMATIC TRANSMISSION VEHICLES

Y SHIM 1)* , S K KAUH 1) and K.-P HA 2)

Hwaseong-si, Gyeonggi 445-706, Korea

(Received May 10 2010; Revised 1 November 2010)

ABSTRACT− Idle stability directly affects a vehicle’s NVH (Noise, Vibration and Harshness) and is closely related to driver satisfaction The present study proposes a method of measuring an engine’s idle roughness, which is useful in quantifying the idle stability Engine brake torque was measured directly using a torque sensor, which can be installed without modification

of the engine’s mounting structure In addition, angular acceleration was measured at the same position as the torque measurement, to compare dynamic characteristics of the angular acceleration with the torque variation Both torque and angular acceleration oscillate between positive and negative values In this study, torque data were divided into several regions, and each region starts from the point where the torque data changes its sign from negative to positive The root mean square values of both torque and angular acceleration were calculated for each region This calculation showed a very good correlation between the torques and the angular accelerations The idle stability was evaluated with the standard deviation of the measured torque, and the cycle-to-cycle variation is a more dominant factor in the idle stability than is the cylinder-to- cylinder variation Because it is easier to measure the angular acceleration than to measure the torque, the variations of angular accelerations are usually compared between engines However, the present study showed that the moment of inertia of an engine and the angular acceleration should be considered together when comparing the idle stability between engines.

KEY WORDS : Torque, Angular acceleration, Idle stability, Moment of inertia

NOMEMCLATURE

(kg·m2)

SD α : standard deviation of angular acceleration (rad/s2)

calculated torque (Nm)

IMEP : indicated mean effective pressure

RMS : root mean square

SDIMEP:standard deviation of indicated mean effective

pressure

1 INTRODUCTION

There has been a continuous effort to improve idle stability

because it is closely related to the driver satisfaction

Generally, the idle stability can be improved by increasingengine speed Fuel economy, however, should also bestrongly considered because the regulation of energy usagehas been reinforced Therefore, it is important to achievehigher idle stability with minimum fuel usage Typically,IMEP’s for 300 cycles were calculated for every cylinder,and the SDIMEP was calculated as the indication ofcombustion stability at idle Recently, various ways toevaluate idle stability have been studied as an alternative toSDIMEP

A cyclic averaging separation method was used toseparate the torque into a periodic component and arandom component The torque was calculated from thecombustion pressure, which was measured at everycylinder It was verified that the idle combustion stability isdirectly related to the random torque component caused bycycle-to-cycle variation and that the random torque moredirectly quantifies the excitation experienced by the vehiclethan does SDIMEP (Beikmann, 2001)

It takes significant time and effort to set up experimentalequipment for measuring the combustion pressure incylinders Additionally, the evaluation of idle combustionstability through combustion pressure cannot reflectdynamic crank motion Teng presented an alternative way

to evaluate combustion stability using flywheel angular

*Corresponding author. e-mail: mcjunkie75@gmail.com

acceleration It is based on the fact that the flywheel

angular acceleration has a direct relationship with the

engine torque The RMS values of the flywheel angular

acceleration were calculated within a cylinder’s

dominat-ing period Next, the cylinder-to-cylinder variation and

cycle-to-cycle variation were evaluated with the standard

deviation of those RMS values (Teng, 2003)

Zhen et al. utilized the crank position sensor signals to

calculate the crank angular acceleration, and they also

evaluated the combustion quality using the RMS values of

calculated angular acceleration An algorithm specially

developed for identifying a missing tooth or teeth was used

to convert the crank position sensor signal into crank

angular acceleration The correlation between RMS values

of angular acceleration and IMEP was presented (Zhen and

An, 2009)

In this study, the engine brake torque was measured

directly with a torque sensor that replaces the driveplate

that connects a crankshaft and a torque-converter of a

transmission In addition, the angular acceleration was

measured at the same position as the torque measurement,

to compare the dynamic characteristics of the angular

acceleration with the torque variation The moment of

inertia of the overall powertrain system, including the

engine and the transmission, and the mean torque were

determined experimentally The idle stability was evaluated

through the measured torque and angular acceleration, and

the results were compared

2 INSTRUMENTATION

2.1 Torque Measurement

Figure 1 shows the torque sensor installation position and

experimental setup The torque sensor is located between

an engine cylinder block and torque-converter, and the

engine brake torque can be measured directly without

modification of the engine’s mounting structure in a

vehicle The output signal from strain gauges that are

digital converter and a microprocessor The digitized signal

is transmitted wirelessly through a Bluetooth module A

telemetry system is essential because the torque sensor

rotates at a very high speed The Bluetooth module that

receives the transmitted data is fixed on the cylinder block,

and finally the data are transferred to a PC via a USBinterface (Lee, 2007)

2.2 Angular Acceleration MeasurementDesigning the circumferential part of the torque sensor asring gear shape, allowed the angular acceleration to bemeasured at the same position as the torque measurement.The magnetic pickup is fixed adjacent to the torque sensor,

as shown in Figure 1 The output signal from the magneticpickup is transferred to a PC after going through a timer

and the angular acceleration was calculated

In this study, the torque and the angular accelerationwere measured for two different SI engines The torquesensor was made individually for each engine and wascalibrated against a standard torquemeter Engine A is a 3.3

Figure 1 Schematic diagram of in-situ torque and angular

Variable valve timing

Variable valve timing

EVALUATION OF IDLE STABILITY THROUGH IN-SITU TORQUE MEASUREMENT 317

L, V6 type, and engine B is a 3.0 L, L6 type Table 1 shows

several major specifications of both engines

3 EXPERIMENTAL RESULT

3.1 Torque and Angular Acceleration

The torque of engine A in neutral idle is presented in

Figure 3 For 3 consecutive cycles, the torque variation per

each cycle is shown The cylinder-to-cylinder and

cycle-to-cycle torque variation can be seen easily, and the torque

oscillates between positive and negative values The

maximum value of each peak in the graph represents the

maximum positive torque reached during the combustion

stroke for each cylinder The positive torque means that the

force created from combustion drives a crankshaft On the

other hand, the negative torquemeans that the crankshaft is

driven by the inertial force before the combustion stroke in

the next cylinder

The angular acceleration at the same point in time is

presented in Figure 4 It shows a similar trend to the torque

graph, but the fluctuation of the angular acceleration is

larger than the fluctuation of the torque on the whole

The interaction of the torque from combustion in every

cylinder results in dynamic crankshaft motion Teng,however, considered that the region from a local minimumpoint to the next local minimum point is a cylinder’sdominating period The RMS values were calculated foreach region (Teng, 2003) Physically speaking, it isreasonable to divide the regions with reference to the localminimum points, but it is difficult to make an exactdivision This is because the fluctuation of measured data ismuch larger near the local minimum points In this study, acylinder’s dominating period is divided with reference toplus zero crossing points, at which the torque data changesfrom negative to positive Figure 3 and Figure 4 are theresults of the plus zero crossing of torque The validity ofplus zero crossing points as a reference will be discussed indetail later

Dividing the regions with reference to the torque’s pluszero closing points, the RMS values of torque and angularacceleration were calculated for each region Figure 5shows the correlation between the RMS torque and theRMS acceleration for engine A in neutral idle There is avery good correlation between the two Most notably, thecorrelation between torque and angular acceleration ishigher than that between IMEP and RMS acceleration

This means that the torque reflects dynamic crankshaftmotion better than IMEP It also suggests that the torquefluctuation can be understood relatively exactly through themeasurement of angular acceleration

3.2 Moment of Inertia and Mean TorqueThe crankshaft motion is affected by the torque from suchfactors as combustion, inertia, and friction Supposing thatthe effect of inertia and friction is consistent at a givenspeed, the relationship between engine brake torque andcrankshaft angular acceleration can be represented simply

Figure 4 Angular acceleration of engine A in neutral idle

during 3 consecutive cycles

Figure 5 Correlation between RMS torque and RMSacceleration for engine A in neutral idle

The torque created in the engine is generally reduced

due to the engine friction or lost from driving the engine’s

accessories, such as the valvetrain In this study, the engine

brake torque, on which the energy loss is already reflected,

was measured directly Therefore, there is no explicit

friction term in equation (1)

Supposing that both the moment of inertia and the mean

torque are constant in equation (1), the torque can be

calculated from equation (2) using the measured angular

acceleration

(2)The torque difference between measured torque and

calculated torque is defined by equation (3)

(3)The moment of inertia and the mean torque were

evaluated by the least square method to minimize the

square of the torque difference described in equation (3)

Then the torque was calculated with those moments of

inertia and the mean torque from equation (2)

For the following cases in both engines, the moment of

inertia and the mean torque are presented in Table 2

For engine A, the moment of inertia in neutral idle is

almost same as that in drive idle The mean torque in drive

idle is aproximately 3.5 times higher than in neutral idle,

which means that the engine was driven under higher load

than in neutral idle

The moment of inertia of engine B is smaller than that of

engine A, so it can be expected that the angular

accelera-tion variaaccelera-tion due to the torque variaaccelera-tion would be higher in

engine B The mean torque of engine B in neutral idle is

higher than that of engine A Though there is not much

difference in engine speed at neutral idle, engine B was

driven under approximately 1.7 times higher load

For engine A in drive idle, Figure 6 shows measured

torque, calculated torque, and torque difference Generally,

calculated torque is consistent with measured torque

Meanwhile, the fluctuation of calculated torque is larger

than the fluctuation of measured torque because the

fluctuation of measured angular acceleration is larger

The mean torque determined experimentally as described

above is almost the same as the arithmetic average of

measured torque This is because the engine brake torque

was measured directly, and it means that equation (1) isvalid

3.3 Evaluation of Idle StabilityIdle stability was evaluated through the measured torqueand angular acceleration Both cylinder-to-cylindervariation and cycle-to-cycle variation were quantified usingthe RMS values of torque and angular acceleration within acylinder’s dominating period The cylinder-to-cylindervariation can be represented by the standard deviation ofmean RMS values of all cylinders The average of thestandard deviations of the RMS values of all the individualcylinders gives the cycle-to-cycle variation

To calculate the RMS values of torque, a cylinder’sdominating period was divided by three differentreferences, which are local minimum, plus zero crossing,and mean torque crossing The cylinder-to-cylindervariation and the cycle-to-cycle variation were calculatedfor each case Figure 7 shows that the results are almost thesame, even though the references are different from each

T b c , = I o α m + T α

T b d , = T b m , – T b c ,

Table 2 Moment of inertia and mean torque

Engine ANeutral idle Drive idleEngine A Neutral idleEngine BMoment of

Figure 7 Cylinder-to-cylinder and cycle-to-cycle variation

of engine A in neutral The idle stability was evaluatedwith reference to local minimum, plus zero crossing, andmean torque crossing each

EVALUATION OF IDLE STABILITY THROUGH IN-SITU TORQUE MEASUREMENT 319

other Because the torque variation is large near local

minimum points, it is difficult to divide the region exactly,

but the plus zero crossing points can be determined more

clearly Considering the efficiency of numerical

manipula-tion, it is better to take the plus zero crossing points as

reference Therefore, the idle stability was evaluated with

reference to the plus zero crossing points

The evaluation results of idle stability for both engines

are presented in Figure 8 For all cases, the cycle-to-cycle

variation is larger than the cylinder-to-cylinder variation

For engine A, the difference in cycle-to-cycle variation

between neutral idle and drive idle is much bigger than that

in cylinder-to-cylinder variation

Comparing the idle stability at neutral idle between

engine A and engine B, both cylinder-to-cylinder variation

and cycle-to-cycle variation are much different It is highly

possible that the difference in cylinder-to-cylinder variation

is caused by the difference in engine specifications such as

engine type and valve drive mode The valve timing

control is working for both engines, and the difference in

control parameters may result in the difference in

cylinder-to- cylinder variation (Kim and Choi, 2006) The difference

in cycle-to-cycle variation can be considered through mean

torque difference As mentioned above, the mean torque of

engine B is higher than that of engine A This means that

the engine B was driven under higher load, and it

contribut-ed to a rcontribut-eduction in the cycle-to-cycle variation

The results described above show that the cycle-to-cycle

variation is a more dominant factor in the idle stability than

is the cylinder-to-cylinder variation, and the amount of the

cycle-to-cycle variation is especially sensitive to the engine

load

The idle stability was also evaluated with the angular

acceleration A cylinder’s dominating period to calculate

plus zero crossing points of angular acceleration The idle

stability of the engine A in drive idle and engine B in

neutral idle is presented in Figure 9 Contrary to the result

in Figure 8, the variation of engine B in neutral idle ishigher than that of engine A in drive idle

There exists the difference in moment of inertia between

engine B in Figure 9 was multiplied by the moment ofinertia of engine B in neutral idle, then divided by themoment of inertia of engine A in drive idle Figure 10shows the comparison result to which the difference inmoment of inertia is reflected A similar trend with torque

is shown The difference in moment of inertia should beconsidered when comparing the idle stability betweendifferent engines using angular acceleration only

4 CONCLUSION

The engine brake torque and the angular acceleration were

Figure 8 Comparison of idle stability for each case The

idle stability was evaluated with reference to the plus zero

crossing of torque

Figure 9 Idle stability of engine A in drive idle and engine

B in neutral idle, which are evaluated with angular tion The difference in moment of inertia between engines isnot considered

accelera-Figure 10 Idle stability of engine A in drive idle andengine B in neutral idle, which are evaluated with angularacceleration The difference in moment of inertia betweenengines is considered

measured for vehicles with automatic transmissions, and

the idle stability was evaluated The main result can be

summarized as follows:

(1) The engine brake torque was measured directly with the

torque sensor, and the sensor replaces the driveplate

that connects a crankshaft and a torque-converter of a

transmission To obtain the dynamic characteristics in

accordance with the torque variation, the angular

acceleration of the plate was measured

(2) The torque data were divided into several regions, and

each region is between the points where the torque data

changes its sign from negative to positive It was

verified that the angular acceleration has good

correlation with the torque

(3) The moment of inertia of the overall powertrain system,

including the engine and the transmission, and the

mean torque were determined experimentally using the

measured torque and the angular acceleration

(4) The cylinder-to-cylinder variation and the

cycle-to-cycle variation were calculated using the measured

torque to evaluate the idle stability The cycle-to-cycle

variation affects the idle stability more than does the

cylinder-to-cylinder variation

(5) Idle stability was also evaluated through the angular

acceleration, and the result was compared with that

using the torque The difference in moment of inertia

should be considered when comparing the idle stability

between different engines using angular acceleration

only

ACKNOWLEDGEMENT−The authors gratefully acknowledge the financial support by the second stage of the Brain Korea 21 Project in 2009.

REFERENCES

Beikmann, R S (2001) Roll-down considerations in idle

Hartwig, M., Via, J and Govindswamy, K (2005) tions of combustion parameters using engine speed

2005-01-2533.

Fundamentals Int Edn McGraw-Hill Singapore.Kim, D S and Cho, Y S (2006) Idle performance of an

SI engine with variations in engine control parameters

Int J Automotive Technology 7, 7, 763−768

Measurement of an Engine and Engine Accessories Using Bluetooth Ph D Dissertation School ofMechanical and Aerospace Engineering Seoul Nat’lUniversity Seoul Korea

Teng, C (2003) Evaluation of idle combustion stability

International Journal of Automotive Technology , Vol 12, No 3, pp 321 − 329 (2011)

321

SLIDING-MODE OBSERVER FOR UREA-SELECTIVE

CATALYTIC REDUCTION (SCR) MID-CATALYST AMMONIA

CONCENTRATION ESTIMATION

M.-F HSIEH and J WANG *

Department of Mechanical and Aerospace Engineering, The Ohio State University, Columbus, OH 43210, USA

(Received 14 May 2010; Revised 6 October 2010)

ABSTRACT− This paper presents an observer design for SCR mid-catalyst ammonia concentration estimation using tailpipe

NO x and ammonia sensors Urea-SCR has been popularly used by Diesel engine powered vehicles to reduce NO x emissions

in recent years It utilizes ammonia, converted from urea injected at upstream of the catalyst, as the reductant to catalytically convert NO x emissions to nitrogen To simultaneously achieve high SCR NO x conversion efficiency and low tailpipe ammonia slip, it is desirable to control the ammonia storage distribution along the SCR catalyst Such a control method, however, requires a mid-catalyst ammonia sensor The observer developed in this paper can replace such a mid-catalyst ammonia sensor and be used for SCR catalyst ammonia distribution control as well as serves for fault diagnosis purpose of the mid-catalyst ammonia sensor The stability of the observer was shown based on the sliding mode approach and analyzed by simulations Experimental validation of the observer was also conducted based on a medium-duty Diesel engine two-catalyst SCR system setup with emission sensors.

KEY WORDS : SCR, Ammonia concentration estimation, Sliding mode observer, Diesel engine

NOMENCLATURE

: catalyst ammonia coverage ratio

: number of mole of ammonia adsorbed on the catalyst

(mole)

: estimated state x

: estimation error of state x: (x- )

Bet SCRs:between SCR catalysts (mid-catalyst)

Bef SCRs:before SCR catalysts

Aft SCRs:after SCR catalysts

1 INTRODUCTION

Aftertreatment systems for Diesel engine emission

al., 2010; Lee et al., 2008; Hsieh et al., 2010; Wang, 2008;

Hsieh et al., 2009) Due to the stringent vehicle emissionregulations worldwide, selective catalytic reduction (SCR)systems have been widely used by Diesel engine vehicles

Among different SCR systems, urea-SCRs have beenproved of being able to reduce more than 90% of engine-

automotive industry (Song and Zhu, 2002) Urea-SCR(simply denoted as SCR in the rest of the paper) utilizes

nitrogen molecule and water by the catalytic reactions inthe SCR catalyst Because ammonia is considered ahazardous material that cannot be directly carried invehicles, 32.5% aqueous urea solution (AdBlue) has beenspecified as the standard precursor of ammonia for vehicleapplications (Shimizu and Satsuma, 2007) However,improper urea injection control (overdose) can lead totailpipe ammonia slip which is highly undesired On theother hand, less urea injection, even though tends to avoidtailpipe ammonia slip, can result in insufficient SCRreductant and thus induces the potential of higher tailpipe

NOx emissions Such contradictions necessitate

sophisticat-ed urea injection control particularly during engine transient

al., 2002; Shimizu et al., 2007; Hsieh et al., 2011a; Hsieh et

al., 2009; Chi et al., 2005; Schar et al., 2006; Upadhyay et

al., 2006; Herman et al., 2009; Wang et al., 2009).The study in (Willems et al., 2007) has pointed out that

θ NH3

MNH

3 Θ

feedback was necessary for SCR control to compensate

system uncertainties during real-world driving as well as

test cycles Recent control strategies utilize tailpipe

provide feedback signals for controlling the urea dosing

However, due to the nonlinearities and complexities of the

SCR dynamics, state distributions in the catalyst are

difficult to know from the tailpipe measurements,

especially for large size SCR catalysts where the state

variations are significant from upstream to downstream

With inadequate understanding about the in-SCR state

distributions, a high SCR efficiency, i.e low tailpipe NOx

and ammonia emissions and less urea injection, is difficult

to be realized in vehicle applications (Willems et al., 2007)

To achieve high SCR efficiency, the study of (Hsieh and

Wang, 2011a) suggested that one of the approaches is to

control the ammonia storage distribution in the axial

direction of the catalyst To achieve this control objective, a

backstepping based control algorithm has been designed in

strategy and controller were experimentally validated in

(Hsieh, 2010) For the ammonia storage distribution

control strategy, by retaining ammonia storage at a high

level at the upstream part of the SCR catalyst and by

limiting the amount of ammonia storage at the downstream

and the tailpipe ammonia emissions can be constrained to a

low level Experimental analyses based on the US06 test

cycle results in (Hsieh, 2010) also showed that, comparing

to the case without considering the storage distribution

along the axial direction of the catalyst, the cumulative

reduced by taking the ammonia axial direction dynamics

into considerations However, such control systems require

ammonia sensor at the middle of the SCR catalyst to

estimate the ammonia storage at the upstream part of the

SCR catalyst

The ammonia sensor at the middle of the SCR catalyst is

undesired from the production cost viewpoint To address

this concern, an observer is proposed in this study to replace

the physical sensor The observer utilizes the tailpipe NOx

and ammonia sensors to estimate the ammonia concentration

at the middle of the SCR catalyst Stability of the observer

was proved based on the sliding mode technique and

analyzed by simulations An experiment was also

conduct-ed basconduct-ed on a two-catalyst SCR system setup with an

ammonia sensor being placed between the two catalysts to

provide the actual mid-catalyst ammonia concentration

signal for validating the observer estimation

The rest of the paper is organized as follow A brief

introduction of basic SCR operational principles and the

SCR control-oriented model are introduced at first

Following that, the observer design is described with

theoretical proof of the error convergence Then, simulation

results and analyses are presented followed by an

experimental validation Finally, conclusive remarks aresummarized

2 UREA-SCR SYSTEM OPERATING PRINCIPLES AND CONTROL-ORIENTED MODEL

2.1 SCR System Operating Principles

steps: AdBlue to ammonia conversion, ammonia adsorption/

conversion generally consists of three reactions: AdBlueevaporation, urea decomposition, and isocyanic acidhydrolyzation (Piazzesi et al., 2006), as listed in Equation(1), Equation (2), and Equation (3), respectively

AdBlue evaporation:

(1)Urea decomposition:

NH 2 − CO − NH 2 → NH 3+HNCO (2)Isocyanic acid (HNCO) hydrolyzation:

NHCO+H 2 O → NH 3+CO 2 (3)For evaporation and decomposition reactions, studieshave pointed out that with sufficient exhaust-gastemperature (above 200 degree C), the reaction rates arevery fast and the AdBlue can usually be completelyconverted to ammonia and isocyanic acid before entering

Hydrolyzation, on the other hand, has limited reaction rateunder 400 degree C However, in the presence of a SCRcatalyst, this reaction becomes very fast, which can be two

geometry design, experimental studies have shown that thisreaction can be completed at the very upstream part of aSCR catalyst (Hsieh, 2010) Thus, it is rational to assume100% AdBlue to ammonia conversion before the SCRcatalyst (Hsieh, 2010)

The ammonia adsorption and desorption reactions in theSCR catalyst can be explained by the following equation

NH 3 ↔ NH 3* (4)

reductions are completed by the reactions with theadsorbed ammonia instead of the gaseous ammonia inexhaust gas, such that the ammonia adsorption anddesorption reactions are very critical in the SCR dynamics.The ammonia storage capacity can be seen as thesummation of the amount of ammonia adsorbed by the

NH 2 – CO NH – 2 ( liquid ) NH → 2 – CO NH – 2*+ xH 2 O

SLIDING-MODE OBSERVER FOR UREA-SELECTIVE CATALYTIC REDUCTION (SCR) 323

catalyst NH 3* and the free catalyst substrate θ free

summariz-ed by the following three reactions

The first reaction in Equation (5) is typically known as

the “standard SCR” due to the fact that this reaction rate is

second reaction in Equation (6) is called “fast SCR”,

because the reaction rate can be one order of magnitude

faster than the standard SCR reaction as studied in

(Grossale et al., 2008) The third reaction of Equation (7)

is commonly known as “slow SCR” because its reaction

rate is generally slower From these chemical reactions, it

can be understood that the adsorbed ammonia is utilized as

the reductant for the selective catalytic reduction Moreover,

it is important to note that reaction rates of these processes

increase with increasing amount of adsorbed ammonia and

2010) In other words, if the available ammonia amount is

fixed, to achieve better ammonia usage efficiency, i.e to

concentrations are high (Song et al., 2002)

Besides the aforementioned three main reaction processes,

SCR catalysts can also perform as an oxidation catalysts for

some specific gases The main oxidation reactions in the SCR

catalyst used in this study are listed below (Hsieh, 2010)

2.2 SCR Control-oriented Model

Based on the preceding reactions, by assuming the SCR

catalyst to be a continuously stirred tank reactor (CSTR), a

0-D control-oriented SCR model was developed as below

in the authors’ work (Hsieh and Wang, 2011b)

(10)

(11)(12)where K j, E j, S 1, S 2 are positive constants, and T is the

catalyst temperature (K)

The positive constants of the model (model parameters)were derived by minimizing the differences of the model

the SCR catalysts) and the sensor measurement based onseveral sets of experimental data Details of the parameterderivation and model validation can be found in the authors’work (Hsieh and Wang, 2011b)

2.3 Two-cell SCR Model for SCR ControlThe CSTR model assumes all the states inside the catalystare homogeneous This assumption is inappropriate whenthe catalyst volume is large In reality, SCR states, e.g NOx

etc, can change from upstream to downstream, and thevariations increase with enlarged catalyst volume A singleCSTR model in Equation (10) is not capable of capturingthe catalyst state changes along the axial direction To dealwith this problem, a multi-cell model is necessary In thisstudy, considering the system controllability and observability,

a two-cell model is used The model equations and aschematic presentation are shown in Equation (13) andFigure 1, respectively As can be seen in Figure 1 the SCRcatalyst is modeled by two separated smaller CSTR models(two-cell), i.e upstream SCR model and downstream SCRmodel By modeling the two smaller volumes separately,the CSTR assumption can better represent the real plantand also the state differences at upstream and downstreamparts of the catalyst can be pronounced

(13)

The corresponding physical states of the subscript i can

be found in Figure 1 The SCR model in Equation (13) hasbeen validated with experimental data, and the resultsshowed that the model can well capture the main SCRcatalyst dynamics at various engine operating conditions.Details of the SCR modeling work and experimentalvalidation can be found in authors’ work (Hsieh and Wang,2011b)

Such a model was also utilized by the studies in (Hsieh,

2010) to develop an ammonia storage distribution control

strategy which has been validated to be effective of

enhanc-ing the SCR system efficiency As schematic presentation of

the control strategy and utilized sensor locations are shown

in Figure 2 In this study, to achieve the sensor number

reduction, the model is then used for the observer design to

estimate the mid-catalyst ammonia concentration, i.e

ammonia concentration between the two catalysts

3 MID-CATALYST AMMONIA

CONCENTRATION OBSERVER DESIGN

3.1 Observer Design

The objective of the observer is to estimate the ammonia

sensors utilized are a tailpipe NOx sensor, a tailpipe ammonia

sensor, thermocouples, and other typical measurements from

can be ignored after the upstream SCR catalyst due to the

fast SCR reaction in Equation (6) (Hsieh, 2010), based on

the SCR model in Equation (13), the dynamics which are

considered in the observer design are summarized in the

following equations

(14)(15)(16)(17)

Equation (12) using temperature measurements, SCR

volumes V are available constants, exhaust flow rates F can

be estimated by engine speed, intake air flow rate, and fuel

(18)where

(19)(20)(21)

(22)

3.2 Observer Stability ProofThe proof of observer stability includes three parts In the

a finite period of time The third part is to prove that

factors for the ammonia storage distribution control, and its

2010) It will be proved in the third part and will bevalidated by simulations in the next section that by the

ammonia coverage ratio of the upstream part of the

Part 1: converges to

Selecting a Lyapunov function candidate as

(23)based on the observer in Equation (20), the time derivate ofthe Lyapunov function candidate becomes

(24)

is available within a finite period of time

2

-r4R 2,Θ 2 θ NH 3 , 2 F 2

V2 -CNH3,in ,

Figure 2 Schematic presentation of ammonia storage

distribution control studied in (Hsieh and Wang, 2011a;

Hsieh, 2010)

SLIDING-MODE OBSERVER FOR UREA-SELECTIVE CATALYTIC REDUCTION (SCR) 325

Part 2: converges to

Practically, the magnitude of

in Equation (24) is comparably large

seen in Figure 5 With this preliminary, we can consider

=0 as a mimic of a sliding surface and assume that

in a finite period of time (Utkin et al., 1999) Based on this

assumption, considering the ammonia concentration

Equation (16) we have:

zero in a finite period of time is guaranteed, which ensures

finite time (Drakunov, 1992; Drakunov and Utkin, 1995)

Therefore, from Equation (25) we have

(29)

time

Part 3: converges to

ammonia concentration estimation error

based on the observer in Equation (22)

becomes

(30)

Based on the equivalent control method,

at sliding mode in a finite time Therefore,

(31)

(32)Selecting a Lyapunov function candidate

(33)

By applying the observer in Equation (21) and therelation in Equation (32), the time derivative of theLyapunov function becomes

concentration before the SCR catalyst represents the

concentra-tion before the SCR catalysts was estimated based on theamount of urea injection and the assumption that theinjected urea can be completely converted to ammonia.Figure 4 shows the comparison of the estimated

Equation (20) can track the model value very well Figure 5shows the zoom-in of the comparison at the initial part Ascan be seen the initial value of the observer was set to 0.5

⎛ ⎞θ˜NH3, 2 = K′ NH3, 2 sign C ( NH3, 2 – C ˆˆ NH3, 2 )

C NH3, 2 Θ 2 r 4 F , 2 1

V 2 - r 4 R , 2 Θ

=

V ·θ

NH3 2, = θ ˜ NH3, 2 [ – θ ˜ NH3, 2 ( r 4 F , 2 C NH3, 2 V 2 + r 3 2 , C O2, 2 V 2 + r 4 R , 2 )

θ NH3, 2 r 1 2 , C NO , 2 C O2, 2 V 2 – –K θ , 2 sign θ(˜NH3,2)]

which is different from the model value of 0 The estimate

converged to the model value in a very short period of time

Figure 6 shows the comparison of the estimated tailpipe

As can be seen the estimated value follow the model value

very well Als Figure 7 shows the Zoom-in at the beginning

part of the simulation As it indicates the estimation

converg-ed to the desirconverg-ed value as expectconverg-ed

Figure 8 shows the comparison of observer estimation of

objective of the proposed observer As can be seen the

estimated based on the proposed estimation law Figure 9also show the Zoom-in of at the initial part of the simulation

As can be seen the estimated value converged to the actualvalue from a different initial value

Figure 10 shows the estimation of upstream catalystammonia coverage ratio and the corresponding modelvalue As can be seen that the observer in Equation (21)successfully utilized the mid-catalyst ammonia concentra-

start of the simulation As it indicates the estimation valuehad some perturbations at the initial and then converged to

Figure 4 Comparison of θ NH3, 1 (model) and θˆ NH3, 1 (observer)

Figure 5 Zoom-in of Figure 4

Figure 6 Comparison of C NH3, 1 (model) and Cˆ NH3, 1 (observer)

Figure 7 Zoom-in of Figure 6

Figure 8 Comparison of C NH3, 2 (model) and Cˆ NH3, 2 (observer)

Figure 9 Zoom-in of Figure 8

SLIDING-MODE OBSERVER FOR UREA-SELECTIVE CATALYTIC REDUCTION (SCR) 327

the model value The initial estimation perturbations were

Because the urea injection started around the 40th second,

function in the observer of Equation (21) cannot trigger

cannot happen in the situation when

ammonia slip after the SCR catalyst was presented and thus

well

4.2 Experimental Setup

The observer is validated based on a two-catalyst SCR

system, where the upstream catalyst serves as the upstream

part of the SCR catalyst in a regular single catalyst setup

and the downstream catalyst serves as the downstream part

of a regular single catalyst setup The SCR catalysts are

Fe-Zeolite catalysts with 4 L volume each The engine used in

the experimental tests is a medium-duty Diesel engine with

the peak power being 350 horsepower and peak torque being

880 Nm The engine is equipped with a high pressure-EGR,

a variable geometry turbocharger, and a high pressure

common rail injection system An ammonia sensor which

has been specially calibrated up to 500 ppm is placed

between the two catalysts to measure the mid-catalyst

ammonia concentration This measured value is treated as

installed at tailpipe (after the downstream SCR catalyst).These two emissions sensors together with other availablemeasurements from ECU are used as the inputs to the

SCR catalyst, which is used to monitor the engine exhaust

reading cannot be directly used An online extended Kalmanfilter to correct the NOx sensor cross-sensitive reading by anammonia sensor has been proposed and validated by theauthors in (Hsieh 2010; Hsieh and Wang, 2011c), and such acorrection is used in this study A schematic presentation andactual picture of the experimental setup are shown in Figure

Figure 11 Zoom-in of Figure 10

Figure 12 Schematic presentation of the experimental setup

Figure 13 Diesel engine and aftertreatment system testbench

4.3 Experimental Validation

Figure 14 shows the engine operation conditions and the SCR

catalyst temperatures during the test Figure 15 shows the

measured by the NOx sensors shows the NOx concentrations

before and after the SCR catalysts and the ammonia

concentration before the SCR catalysts As can be seen that

tailpipe NOx concentration decreased immediately after the

urea injection started

and the comparison of the mid-catalyst (between the two

SCR catalysts) ammonia concentrations estimated by the

seen the observer estimation converged to the sensor

second

The observer was not able to estimate the concentrationvery well in the first 1100 seconds It was because that the

zero in this region as can be seen in Figure 16 Such aphenomenon was caused by the same reason as explained

in the previous section, where the upstream ammoniacoverage ratio had difficulty to be estimated at initial due to

problem is more conspicuous in the experimental resultbecause the tailpipe ammonia concentration can stay atzero, instead of very low values as in the simulations, whenammonia storage in the catalyst is very low Because of thezero tailpipe ammonia slip, the sign function of the slidingmode observer in Equation (18) cannot trigger appropriate

tailpipe ammonia slip was measured However, in regular

capacity, the ammonia storage in the SCR catalyst isusually kept above a certain value (Hsieh, 2010) Withsufficient ammonia storage, ammonia slip, even though can

be small, is usually presented at the tailpipe With thetailpipe ammonia slip, the observer can successfullyestimate the mid-catalyst ammonia concentration as shown

in Figure 16 and the simulation validation in the previous

where ammonia injection was stopped and the tailpipeammonia slip was decreased immediately Even though theurea injection was stopped, sufficient ammonia was still

reduction in Figure 15 With the ammonia storage, there wasstill some ammonia slip at tailpipe and the observer was able

to estimate the mid-catalyst ammonia concentration based

on such low tailpipe ammonia measurement

5 CONCLUSIONS

A mid-catalyst ammonia concentration observer waspresented in this paper Such an observer can provideinformation on the ammonia concentration at the middle ofthe SCR catalyst The observer was designed based on thesliding mode technique and utilizes a set of tailpipe NOx and

analyses verified that the observer was able to estimateammonia concentration as desired, and the observer can bedirectly used to estimate the ammonia storage at theupstream part of the SCR catalyst Experimental validationwas also conducted using a two-catalyst SCR setup Theestimation results from the proposed observer showedconsistency comparing to the measurement from a speciallycalibrated ammonia sensor placed at the middle of the twocatalysts

C NH3, 1

C NH3, 2

Cˆ NH3, 1

Cˆ NH3, 2Figure 14 Engine/SCR operation parameters

catalysts measured by the NOx sensors

estimation

SLIDING-MODE OBSERVER FOR UREA-SELECTIVE CATALYTIC REDUCTION (SCR) 329

REFERENCES

Chi, J N and DaCosta, H F M (2005) Modeling and

SAE World Cong., SAE Paper No 2005-01-0966

Drakunov, S V (1992) Sliding-mode observers based on

and Control

Drakunov, S V and Utkin, V (1995) Sliding mode

observers tutorial Proc 34th Conf Decision and Control

Grossale, A., Nova, I., Tronconi, E., Chatterjee, D and

Weibel, M (2008) The chemistry of the NO/NO2-NH3

“fast” SCR reaction over Fe-ZSM5 investigated by

transient reaction analysis J Catalysts, 256, 312−322

Herman, A., Wu, M., Cabush, D and Shost, M (2009)

Model based control of SCR dosing and OBD strategies

Cong., SAE Paper No 2009-01-0911

Hsieh, M.-F., Wang, J and Canova, M (2010) Two-level

Measure-ment, and Control 132, 4, 13

Hsieh, M.-F., Canova, M and Wang, J (2009) Model

predictive control approach for AFR control during lean

NOx trap regenerations SAE Int J Fuels and Lubricants

2, 1, 149−157

Hsieh, M and Wang, J (2011a) A two-cell backstepping

based control strategy for diesel engine selective catalytic

Technology (in press), (DOI: 10.1109/TCST.2010.20984

77)

Hsieh, M and Wang, J (2009) Backstepping based

nonlinear ammonia surface coverage ratio control for

diesel engine selective catalytic reduction systems Proc.

ASME Dynamic Systems and Control Conf.

Hsieh, M.-F (2010) Control of Diesel Engine Urea Selective

Catalytic Reduction Systems Ph.D Dissertation Ohio State

University

Hsieh, M F and Wang, J (2011c) Design and experimental

concentration estimator in selective catalytic reduction

press), (DOI: 10.1016/j.conengprac.2010.12.002)

Hsieh, M F and Wang, J (2011b) Development and

experimental studies of a control-oriented SCR model for a

Practice (in press), (DOI: 10.1016/j.conengprac.2011.01

004)

Hsieh, M F and Wang, J (2010) Observer-based estimation

of selective catalytic reduction (SCR) catalyst ammonia

storage J Automobile Engineering, Proc Institution of

Mechanical Engineers, Part D, 224, 9, 1199−1211

Jung, S., Ishida, M., Yamamoto, S., Ueki, H and Sakaguchi,

D (2010) Enhancement of NOx-PM trade-off in a diesel

Technology 11, 5, 611−616

Kim, J Y., Ryu, S H and Ha, J S (2004) Numericalprediction on the characteristics of spray-induced mixingand thermal decomposition of urea solution in SCR

Division.Lee, S J., Jeong, S J., Kim, W S and Lee, C B (2008).Computational study on the effects of volume ratio ofDOC/DPF and catalyst loading on the PM and NOx

Measurement, and Control, 128, 737−741

Control in Electromechanical Systems Taylor & FrancisInc London

Wang, J (2008) Smooth in-cylinder lean-rich combustionswitching control for diesel engine exhaust-treatment

Electronic and Electrical Systems 1, 1, 340−348 Wang, D Y., Yao, S., Shost, M., Yoo, J., Cabush, D.,Racine, D., Cloudt, R and Willems, F (2009) Ammonia

Cong., SAE Int J Passenger Cars- Electronic and Electrical Systems 1, 1, 323−333

Way, P., Viswanathan, K., Preethi, P., Gilb, A., Zambon, N.and Blaisdell, J (2009), SCR performance optimizationthrough advancements in aftertreatment packaging

Proc SAE 2009 World Cong., SAE Paper No. 0633

2009-01-Willems, R., Cloudt, E V D., Eijnden, B D., Jager,Boomsma, W and Heuvel, I V D (2007) Is closed-loop SCR control required to meet future emission

2007-01-1574

Yim, S D., Kim, S J., Baik, J H., Nam, I., Mok, Y S., Lee,

J H., Cho, B K and Oh, S H (2004) Decomposition

Res., 43, 4856−4862

NANO-PARTICLE EMISSION CHARACTERISTICS OF EUROPEAN

AND WORLDWIDE HARMONIZED TEST CYCLES

FOR HEAVY-DUTY DIESEL ENGINES

C L MYUNG 1) , J KIM 1) , S KWON 2) , K CHOI 1) , A KO 1) and S PARK 1)*

(Received 31 May 2010; Revised 18 August 2010)

ABSTRACT− This study was conducted for the experimental comparison of particulate emission characteristics between the European and World-Harmonized test cycles for a heavy-duty diesel engine as part of the UN/ECE PMP ILCE of the Korea Particulate Measurement Program To verify the particulate mass and particle number concentrations from various operating modes, ETC/ESC and WHTC/WHSC, were evaluated Both will be enacted in Euro VI emission legislation The real-time particle emissions from a Mercedes OM501 heavy-duty golden engine with a catalyst based uncoated golden DPF were measured with CPC and DMS during daily test protocol Real-time particle formation of the transient cycles ETC and WHTC were strongly correlated with engine operating conditions and after-treatment device temperature The higher particle number concentration during the ESC #7 to #10 mode was ascribed to passive DPF regeneration and the thermal release of low volatile particles at high exhaust temperature conditions The detailed average particle number concentration equipped for golden DPF reached approximately 4.783E+11 #/kWh (weighted WHTC), 6.087E+10 #/kWh (WHSC), 4.596E+10 #/kWh (ETC), and 3.389E+12 #/kWh (ESC) Particle masses ranged from 0.0011 g/kWh (WHSC) to 0.0031 g/kWh (ESC) The particle number concentration and mass reduction of DPF reached about 99%, except for an ESC with a reduction of 95%.

KEY WORDS : European cycles, Worldwide harmonized test cycles, Nano-particles, Condensation particle counter, Differential mobility spectrometer, Diesel particulate filter

1 INTRODUCTION

To address the worldwide climate change problem, the EU

(European Union), Japan, California, and Korea have been

for automobiles Increasingly stringent exhaust emission

regulations have resulted in the development of more energy

efficient and environmentally friendly internal combustion

engine technologies (Kim and Lee, 2007; Zhao, 2010)

Diesel is the most efficient internal combustion engine

fuel providing more power and fuel efficiency than

gasoline, CNG (compressed natural gas), and LPG

(liquefi-ed petroleum gas) (Arcoumanis et al., 2008; Ristovski et al.,

2000, 2005) Although diesel engines perform well with

combustion inherently tends to produce significant amounts

of PM (particulate matter) caused by incomplete

combus-tion around individual fuel droplets in the combuscombus-tion zone

(Eastwood, 2008; Rakopolous and Giakoumis, 2009) PM

has been suspected of causing acute and chronic damage to

the human pulmonary and cardiovascular systems by the

2001, 2004; Kittelson, 1998; Kittelson and Simone, 1999).Currently, the diesel PM standard is not based on theparticle number (#/km), but rather on the particle mass (g/km) which can not sufficiently represent the toxicity ofdiesel-oriented nano-particles (Myung et al., 2009a; Sturgess

et al., 2008; Vouitsis et al., 2003)

Thus, the international PMP (Particle MeasurementProgramme) has been developing a new particle numbermeasurement technique through the ILCE (Inter-LaboratoryCorrelation Exercise) on LDVs (light-duty vehicles) forEuro V standards and on heavy-duty engines for Euro VIstandards, to complement or replace mass-based PM

et al., 2008, 2009; Roberto et al., 2007) Number-based PMmeasurement procedure prevents the possibility that theEuro VI PM mass limit is met using open filters that wouldenable a high number of ultra fine particles to pass

particular, two golden engines meeting the Euro IIIemission regulation were employed in the PMP to developand finalize a robust inter-laboratory guide for heavy-dutyengines testing under UN-GRPE phase 3 ILCE for the VE(validation exercise) with the IVECO Cursor 8 engine and

*Corresponding author. e-mail: spark@korea.ac.kr

332 C L MYUNG et al.

round-robin with Daimler engine, respectively (Anderson

and Clarke, 2008) For the round-robin tests, laboratories

from the EU, Japan, Korea, and Canada are participating In

test procedures for heavy-duty diesel engine, the WHTC

(worldwide heavy-duty transient cycle) and WHSC

(worldwide heavy-duty steady-state cycle) are scheduled to

replace the conventional ESC (european stationary cycle)

and ETC (european transient cycle) in the emission

legislation for Euro IV (Giechaskiel et al., 2008a; May et

The purpose of this study is to report the experimental

evaluation results on the particle numbers and mass

concentrations, particle size distribution, filtration efficiency

of DPF (diesel particulate filter), and repeatability of the

European and WHTC (world-harmonized test cycle) for a

heavy-duty diesel engine as part of the UN/ECE PMP ILCE

in the Korea PMP These results will be used to establish a

nano-particle numeric standard for heavy-duty diesel

engines, and provide repeatability and reproducibility

criteria for domestic round-robin testing

2 EXPERIMENTAL APPARATUS AND

METHOD

2.1 Specifications of Test engine and Procedure

Heavy-duty diesel engine and exhaust systems, including

exhaust pipes with DPF and EMS (engine management

systems), were provided by the PMP for the international

ILCE The engine is a turbocharged Euro III-compliant

Daimler OM501 equipped with catalyst-based uncoated

DPF Detailed engine specifications are summarized in

Table 1

Engine setting and speeds conditions for various test

modes were followed based on PMP procedures (Stein,

2008) Domestic RF-06 grade diesel fuel with sulfur

content below 10 ppm was employed A single batch of

fuel was supplied by SK Energy, and its specifications are

given in Table 2 Before testing the engine, 10W-40 grade

engine oil was flushed and filled to eliminate lubricant

effects on nano-particle emissions

The testing procedure for the heavy-duty engine was

strictly followed to provide the variability of repetitions inparticle measurements The test protocol for the ILCEconsisted of at least eight repetitions of cold WHTC, hotWHTC preceded by a 10 min soak, ETC, and ESC cycles,

as shown in Table 3 The continuity protocol between eachtransient cycle (defined as 5 min at idle and 5 min ofoperation at the ESC mode 7, plus 3 min at idle) wasapplied to ensure identical temperature profiles of the

engine-out emissions were measured to quantify the filtrationefficiency of the DPF

2.2 Primary Dilution and Particulate Analysis System

A full-flow constant volume sampler (CVS) exhaustdilution tunnel system (AVL CVS i60) meeting therequirements of Regulation 49 was used The flow rate of

standard reference conditions (i.e 20oC and 1 bar) The airused for the primary dilution of exhaust in the CVS tunnelwas first passed through a first HEPA (high efficiencyparticulate air) type charcoal-scrubbed filter, and then

Table 1 Golden engine specifications

Table 2 Specifications of reference fuel (RF-06 grade)

Table 3 Daily protocol of heavy-duty diesel engine

Hot WHTC

10 minute atWHSC mode 9WHSCCPETCCPESC

IFV : instrument Functional verification

CP : continuity protocol Precon : 15min ESC mode 10, 30min ESC mode 7

passed through a secondary HEPA filter

The mass of particulate material was measured using a

system composed of a particulate mass sampler, a sample

pre-classifier, and a filter holder assembly A sample probe

(sharp-edged and open ended) was fitted near the center

line in the dilution tunnel, 10 tunnel diameters downstream

of the gas inlet The dilution tunnel had an internal

point) was used The particulate mass system was heated

externally to 47±5oC using a heating controller The heated

elements included the filter holder and transfer tubing and

had a residence time greater than 0.2 s when calculated at

the required flow rate of 45 L/min Pallflex TX40

fluorocarbon coated glass filters were used to collect

particulates The mass collected was measured using as

analytical balance (Sintronix, model SE 2-F) with a

resolution of 0.1 in accordance with the test procedure

The number of particles emitted by the golden engine was

determined using a GPMS (golden particle measurement

system) The GPMS system has two main parts: 1) a particle

sampling system consisting of a sampling and particle

conditioning and measurement system consisting of a VPR

(volatile particle remover) and a PNC (particle number

counter) unit The VPR provides heated dilution, thermal

conditioning of the sample aerosol, and secondary dilution

for cooling and freezing of the sample evolution prior to

entry into the PNC The PNC unit is a particle counter (TSI

3010D) with the lower cut-off modified to 23 nm by the

manufacturer (Giechaskiel et al., 2008a; Lee et al., 2008;

Myung et al., 2009b; Anderson et al., 2010)

In addition to particle number concentration, the

DMS500 (Cambustion Co.) positioned at engine-out and

afterward DPF was used to analyze the particle size

distribution emitted from the heavy duty engine The

DMS500, which is based on the same operating principle

as the DMA (differential mobility analyzer) of the SMPS

(scanning mobility particle sizer) measures the number of

particles and their spectral weighting in the 5 nm to 2.5 µm

size range with a scan time of 200 ms (Cambustion, 2008).Details of the experimental system for heavy-duty round-robin testing at NIER (National Institute of EnvironmentalResearch) in Korea are shown in Figure 1

2.3 Calculation Procedure of Total Particle NumberTotal particle number (PN) emissions for each mode werecalculated by means of the following equation by particlenumber measurement procedure Regulation 49 (Giechaskiel

et al, 2008b)

(1)(2)

over the cycle given by the CPC, PRF is particle

(kg/h) is flow rate of diluted exhaust gas, t (h) is the duration

of the cycle, and Wact (kWh) is the actual cycle work

following equations Pi (W) is the power at mode i and WFi

is the weighting factor of mode i

(3)

(4)For the WHTC, the weighted results should be calculat-

ed using the adjustment factors of 14% cold WHTC plus86% hot WHTC as shown in Equation (5)

(5)Where:

Ncold : total number of particles emitted over the WHTCcold test cycles

hot test cycles

Wact, cold : cold actual cycle work in kWh

Wact, hot : hot actual cycle work in kWh

without periodically regenerating after-treatment kr=1)

3 RESULTS AND DISCUSSION

3.1 Real-time Particle Emission Characteristics withVarious Engine Operating Cycles

Figure 2 shows real-time particle emissions of variousengine operating cycles with an after-treatment device In thecase of the cold WHTC mode, 104 #/cm3 order of particles

=

e k r ( 0.14 N × cold ) 0.14 W × act cold ,

=

Figure 1 Experimental setup for heavy-duty round-robin

testing at the NIER

334 C L MYUNG et al.

were emitted during the 450 s from cold start, and after 750

s, the number of particles decreased by approximately 2

orders of magnitude due to the high filtration efficiency of

the DPF We note that the accumulated particles in the

previous preconditioning operation were released from the

DPF during the cold transient starting phase in which

temperature maintains still low condition Thus,

nano-particles were not deposited on the surface of the filter or

emitted through porosity inside the filter because the

diffusion velocity decreased during the low DPF temperature

condition zone (Eastwood, 2008; Rakopolous and

Giakoumis, 2009)

Compared with the cold WHTC mode, the particle

concentration was reduced by two order of magnitude and

ETC modes due to the high DPF temperature Thecumulative particle number emissions over 106 #/cm3 of the

WHTC and 104 #/cm3, respectively

Within the 13 measured points of the WHSC mode withDPF, the particle emission at the WHSC #10, for which theoperating conditions were 100% engine power and 87%rated engine speed, showed a clear trend of insufficientfiltration efficiency The particle spike can be seen in theWHSC #10 mode for both the real-time particle emissionand the cumulative particle concentration profile Theaccumulated particles showed a reduction of 2 orders ofmagnitude from 106 #/cm3 to 104 #/cm3 in the WHSC mode.From the ESC mode #3 with 50% engine load and 85%speed to mode #10 with 100% engine load and 85% speed,the particle concentration steadily increased even though theDPF was installed The real-time particle concentrationreached approximately 103 #/cm3 despite the after-treatmentdevice during the ESC #8~#10 modes because of passiveregeneration in the filter The total accumulated level of the

much higher than other test modes

Figure 3 shows the effects of the after-treatment device

on particle size distribution and number concentration forthe WHTC, WHSC, and ETC modes using the DMS500.Particle sizes from an internal combustion engine aregenerally classified into the nucleation and accumulationmodes as distinguished by particle diameter (Kittelson,1998) In this study, the nucleation and accumulation modehave particle diameters of less than approximately 50 nm,and from 50 nm to 1,000 nm, respectively In the case ofhot WHTC, WHSC, and ETC, a nuclei mode below Dp <

30 nm was observed after DPF, and the particle number

consecutively The particle size range of 30 nm < Dp < 400

#/cm3 to 104 #/cm3 after DPF Most particles with a DPFdecreased by two to three orders of magnitude except thecold WHTC mode that involved cold start operation

Figure 2 Real-time particle concentration and accumulated

particle number with European/WHTC modes

Figure 3 Particle size distribution and number concentration of WHTC, WHSC, and ETC modes

Figure 4 shows the time-resolved particle size

distribu-tion spectra of the DPF-equipped ESC mode using the

DMS500 Based on the particle spectra of the ESC mode,

the accumulation mode of 50 to 200 nm wasdistinctly

emitted during the #6 to #13 modes whose order of

nuclei mode, Dp < 50 nm in ESC # 8 to #10 modes occurred

because the higher exhaust temperature led to passive DPF

regeneration and thermal release of low volatility particles

3.2 Comparison of Particle Numbers and Mass Emissions

for Each Mode

Figure 5 shows the averaged particle number and mass of

various engine operating cycles with an after-treatment

device Particle number (particle mass) results over various

engine operation cycles showed that weighted WHTC was

4.783E+11 #/kWh (2.0 mg/kWh), WHSC 6.087E+10 #/kWh (1.1 mg/kWh), ETC 4.596E+10 #/kWh (1.5 mg/kWh), and ESC 3.389E+12 #/kWh (3.1 mg/kWh) Thefiltration efficiency of the particle number concentrationand mass reached approximately 99%, except for an ESCmass efficiency of 95% due to the passive regenerationduring the # 8 to #10 modes

The ratio of the standard deviation over the averagevalue of particle emissions (tested eight times), COV(coefficient of variance), is summarized in Table 4 When repeated emission tests are performed in the samelaboratory within a short period of time, the results should

be reasonably consistent The COV of the cold WHTCmode showed good repeatability at 26% for the particlemass and 28% for the particle number because of the sootloading procedure on the DPF through preconditioning TheCOV of particle number and mass ranged from 49% (hotWHTC) to 73% (ETC), and 22% (ESC) to 107% (ETC)

A COV of 22% of the particle mass in the ETC modemeans that its variation with DPF regeneration wasnegligible

Figure 6 shows an image of soot deposition andmorphology on fiber filters using an FESEM (field emissionscanning electron microscope, Hitachi S-4300) over varioustest modes The typical particle size distribution of thediesel engine exhibited a log-normal size around 100 nm, asshown in Figures 3 and 4, similar-sized particles thatappeared to be agglomerated chains or clusters (Eastwood,2008) were deposited on the fiber filter for various operat-ing cycles without DPF In addition, smaller and fewerparticles were observed with filtration, except in the coldWHTC mode, including the cold start phase with DPF-equipped cases

4 CONCLUSION

In this study, a comparison of nano-particle concentrationsand particle mass emissions between the European andWHTC for a golden heavy-duty diesel engine, as part of theUN/ECE PMP ILCE, was performed Based on the dailyprotocol suggested by the PMP, the filtration efficiency andrepeatability of golden DPF with particle size distributionsfor various test cycles were also investigated The majorfindings are summarized as follows

Figure 4 Time-resolved particle size distribution spectra of ESC mode after DPF

Figure 5 Average particle number and mass emissions for

various test modes

336 C L MYUNG et al.

the level steadily decreased to 101 #/cm3 The real-time

particle concentration in the ESC mode reached

device during the ESC #8 to #10 modes because of

passive regeneration in the filter

(2) Particle emissions over various engine certification

cycles showed that the particle number concentrations

(#/kWh) with DPF were 4.783E+11 (weighted WHTC),

6.087E+10 (WHSC), 4.596E+10 (ETC), and

3.389E+12 (ESC) Particle mass (g/kWh) ranged from

0.0011 (WHSC) to 0.0031 (ESC) The repeatability of

particle number and mass values reached approximately

22~107% and 25~73%, respectively The large COV of

the particle emissions was ascribed to the ETC mode,

which produces values of 107% and 73% The DPF

efficiency of particle number concentration and mass

reached around 99%, except for an ESC mass efficiency

of 95%

(3) The particle size distribution of WHTC, WHSC, ETC,and ESC with a DPF-equipped diesel engine was bi-modal, consisting of nucleation and accumulationmodes From the FESEM image, the agglomeratedchain or cluster particles around 100 nm were captureddespite DPF, but smaller particles were observed,except for in the cold WHTC mode

ACKNOWLEDGEMENT−This study was supported by the ECO-STAR project, KPMP, and Korea University Grant

Anderson, J., Giechaskiel, B., Munoz-Bueno, R., Sandbach,

E and Dillara, P (2007) Particle Measurement Programme(PMP) Light-duty Inter-laboratory Correlation Exercise(ILCE-LD) Final Report

Anderson, J., Mamakos, A., Giechaskiel, B., Carriero, M.and Martini, G (2010) Particle Measurement Programme(PMP) Heavy-duty Inter-laboratory Correlation Exercise(ILCE-HD) Final Report

Anderson, J and Clarke, D (2008) UN-GRPE PMP Phase

3 Inter-laboratory correlation exercise: Updated work and laboratory guide for Heavy Duty (HD) engine

Frame-Table 4 Detailed particle emissions and COV for various modes

Figure 6 Comparison of particle filters morphology by

FESEM on various test modes

Cambustion (2008) Fast Particulate Spectrometer User

Manual http://www.cambustion.co.uk

John Wiley & Sons Ltd UK

Giechaskiel, B., Alessandrini, S., Forni, F., Martinez-Lozano,

P., Lesueur, D., Carriero, M and Martini, G (2008a)

Particle Measurement Programme (PMP): Heavy-Duty

(HD) Inter-laboratory Exercise: Exploratory Work at JRC

(Oct ’07 – Feb ’08) EU Report 23426

Giechaskiel, B., Martinez-Lozano, P., Alessandrini, S., Forni,

F., Montigny, F., Fumagalli, I., Carriero, M and Martini, G

(2008b) Particle Measurement Programme (PMP) Heavy

duty (HD) Inter-laboratory Exercise: Validation Exercise

Tests at JRC (Feb 08) EU Report 23496

Kim, M Y and Lee, C S (2007) Effect of narrow fuel

spray angle and a dual injection configuration on the

improvement of exhaust emissions in a HCCI diesel

engine Fuel, 86, 2871−2880

Kittelson, D (1998) Engines and nanoparticles: A review

J Aerosol Sci., 29(5/6), 575−588

Kittelson, D and Simone, H (1999) Mechanism of

particulate matter formation in spark-ignition engines

Environ Sci Technol., 33, 3957−3967

Lee, J W., Jung, Y I., Jung, M W., Cha, K O., Kwon, S I.,

Kim, J C and Park, S (2008) Experimental investigation

and comparison of nano-particle emission characteristics

Automotive Technology 9, 4, 397−403

Lee, H., Kim, J., Myung, C L and Park, S (2009)

Experimental investigation of nanoparticle formation

characteristics from advanced gasoline and diesel fueled

light duty vehicles under different certification driving

1591−1601

May, J., Bosteels, D., Such, C., Nicol, A and Andersson, J

(2008) Heavy-duty engine particulate emissions:

Application of PMP methodology to measure particle

2008-01-1176

May, J., Bosteels, D., Nicol, A., Andersson, J and Such, C

(2007) The application of emissions control technologies

to a low-emissions engine to evaluate the capabilities of

future systems for European and world-harmonised

regulations 16th Aachen Colloquium

Myung, C L., Lee, H., Choi, K., Lee, Y J and Park, S

(2009a) Effects of gasoline, diesel, LPG, and low-carbonfuels and various certification modes on nanoparticleemission characteristics in light-duty vehicles Int J Automotive Technology 10, 5, 537−544

Myung, C L., Lee, H., Kwon, S., Lee, S., Jun, J., Lee, Y J.,Woo, Y., Lee, M., Bae, G N and Park, S (2009b) Inter-laboratory correlation exercise on a light-duty dieselpassenger vehicle to verify nano-particle emissioncharacteristics by Korea particle measurement program

J Mechanical Science and Technology, 23, 729−738

Engine Transient Operation – Principles of Operation and Simulation Analysis Springer UK

Ristovski, Z D., Morawska, L., Hitchins, J., Thomas, S.,Greenaway, C and Gilbert, D (2000) Particle emissionsfrom compressed natural gas engines J Aerosol Sci 31,

4, 403−413

Ristovski, Z D., Jayaratne, E R., Morawska, L., Ayoko, G

A and Lim, M (2005) Particle emissions and carbondioxide emissions from passenger vehicles operating on

Environment, 345, 93−98

Roberto, C., Volker, S., Rainer, V and Thorsten, B (2007).Measurement of nucleation and soot mode particleemission from a diesel passenger car in real world andlaboratory in situ dilution Atmo Environ., 41, 2125−

2135

Stein, J (2008) Installation and Operation Guide RoundRobin Engine for UN-GRPE PMP Phase 3 Inter-laboratoryCorrelation Exercise Daimler AG

Sturgess, J., Hill, B and Samuel, S (2008) Experimentalinvestigation into particulate size and number distributionfrom a spark ignition engine: In-cylinder combustion andtransient events J Automobile Engineering, 222, 1065−

1075

TRANS-WP29-GRPE-41 (2001) Development of aWorldwide Harmonized Heavy-duty Engine EmissionTest Cycle

Vouitsis, E., Ntziachristos, L and Samaras, Z (2003).Particulate matter mass measurements for low emitting

Energy and Combustion Science, 29, 635−672

Engine Technologies and Development - Volume 2: Diesel Engines Woodhead Publishing Ltd UK

International Journal of Automotive Technology , Vol 12, No 3, pp 339 − 350 (2011)

339

EFFECTS OF SPLIT INJECTION, OXYGEN ENRICHED AIR AND HEAVY EGR ON SOOT EMISSIONS IN A DIESEL ENGINE

L D K NGUYEN 1) , N W SUNG *1) , S S LEE 2) and H S KIM 3)

Incheon 403-714, Korea

Daejeon 305-343, Korea

(Received 24 March 2009; Revised 8 September 2010)

ABSTRACT− The effects of split injection, oxygen enriched air, and heavy exhaust gas recirculation (EGR) on soot emissions

in a direct injection diesel engine were studied using the KIVA-3V code When split injection is applied, the second injection

of fuel into a cylinder results in two separate stoichiometric zones, which helps soot oxidation As a result, soot emissions are decreased When oxygen enriched air is applied together with split injection, a higher concentration of oxygen causes higher temperatures in the cylinder The increase in temperature promotes the growth reaction of acetylene with soot However, it does not improve acetylene formation during the second injection of fuel As more acetylene is consumed in the growth reaction with soot, the concentration of acetylene in the cylinder is decreased, which leads to a decrease in soot formation and thus soot emissions A combination of split injection, a high concentration of oxygen, and a high EGR ratio shows the best results in terms of diesel emissions In this paper, the split injection scheme of 75.8.25, in which 75% of total fuel is injected

in the first pulse, followed by 8 o CA of dwell time, and 25% of fuel is injected in the second pulse, with an oxygen concentration of 23% in volume and an EGR ratio of 30% shows a 45% reduction in soot emissions, with the same NOx emissions as in single injection.

KEY WORDS : Split injection, Oxygen enriched air, Heavy EGR, Soot emissions, Diesel engine

NOMENCLATURE

k : turbulent kinetic energy [m2/s2]

in soot An aftertreatment device such as a diesel culate filter is needed to control soot emissions However,

parti-*Corresponding author. e-mail: nwsung@skku.edu

it has an additional cost A strategy to reduce NOx and soot

emissions simultaneously without the use of aftertreatment

devices is necessary, for example, a combination of split

injection, oxygen enriched air (OEA), and EGR

Split injection has been used as a method to reduce soot

the possibility of applying split injection with an electronic

unit injector Split injection became more practical later

with the development of a common rail injection system

By varying the amount of fuel in the first injection, Nehmer

and Reitz (1994) found that soot emissions were decreased

when more fuel was injected during the first injection In the

study of Han et al. (1996), soot emissions were reduced by

a factor of four, with a split injection scheme of 75.8.25,

whereby 75% of total fuel is injected in the first pulse,

followed by 8oCA of dwell time, and 25% of fuel is injected

in the second pulse, in comparison with a single injection

Other researchers (Hampson and Reitz, 1998; Bakenhus

and Reitz, 1999; Zhang and Nishida, 2003) using two-color

imaging optics to observe the soot emission process in

diesel engines also confirmed the benefit of split injection in

reducing soot emissions

The concept of using OEA in diesel engines was studied

a long time ago The use of OEA has advantages such as

reduction of soot emissions, CO, and unburned

hydro-carbons (Sekar et al., 1990; Virk et al., 1993; Lee et al.,

2007) However, the increase in NOx emissions and lack of

practical on-board oxygen enrichment devices prevented

any application of this concept In recent years, there has

been progress in the development of oxygen enrichment

devices such as a permeable oxygen membrane A compact

membrane module developed by Argonne National

Laboratory can increase the concentration of oxygen to 25%

in volume, and can be incorporated into vehicle design

(Stork and Poola, 1998) As this technology is developed,

oxygen enriched air becomes more attractive as a method to

reduce soot emissions

The effects of EGR on the reduction of NOx emissions

in diesel engines have been confirmed for years, and it is an

indispensable technology for modern diesel engines

Because of the dilution effect of EGR, the local flame

temperature is decreased, and thus NOx formation is

decreased The lower temperature, however, results in

increased soot emissions To avoid this drawback, many

researchers suggested that the EGR ratio should be

restricted to less than 20% (Uchida et al., 1993; Wagner et

al., 2000; Lee et al., 2007) However, if EGR is used

together with OEA, a higher EGR ratio can be applied to

reduce both NOx and soot emissions simultaneously

Taking into account the benefits of EGR for reducing NOx,

and split injection for reducing soot, Pierpont et al. (1995)

examined the possibility of the combined use of EGR and split

injection They reported that this combination is effective in

reducing both soot and NOx emissions, especially during a

high load condition when EGR causes a significant increase in

study under conditions similar to the experimental conditions

of Pierpont et al. (1995) to understand the soot mechanism.However, the two-step soot model of Hiroyasu et al. (1983)used in their study could not explain the details of sootproduction Later, Fusco et al. (1994) proposed an eight-stepphenomenological soot model Kazakov and Foster (1998)modified the Fusco model and successfully compared theirresults with experimental data The Foster model offers abetter description of soot formation and soot oxidation, andhas been widely applied in simulations of diesel emissions

In this study, the soot model is developed and applied to

a direct injection diesel engine to investigate the combinedeffects of split injection, OEA, and heavy EGR on sootemissions The purpose of this work is to understand themechanism of soot formation and oxidation, to obtaindetails of spatial distribution and time evolution of quan-tities, which are relevant to soot production, and to evaluatetheir influence on soot production during the combustionprocess

2 SOOT MODEL

The formation of soot is explained through a series ofprocesses proposed by Kazakov and Foster (1998) Aschematic diagram of the eight step soot model of Foster isshown in Figure 1 Under high temperature during combus-tion, precursor and acetylene are generated simultaneouslyfrom fuel pyrolysis For the calculation, the precursor, PR, isassumed to be C50, and the fuel is tetradecane, C14H30, due toits similar carbon/hydrogen ratio to diesel fuel

EFFECTS OF SPLIT INJECTION, OXYGEN ENRICHED AIR 341

Acetylene adds a carbon atom to the surface of soot

particles, making the soot particles larger Soot particles

also aggregate to form larger particles,

Because of the high temperature and the existence of

oxygen in the cylinder during the combustion process, the

precursor, acetylene, and soot particles are oxidized The

net soot is the result of soot formation and soot oxidation,

The relevant reaction rates are listed in Table 1 To

consider the effect of turbulent mixing in oxidation, the

global oxidation rates of precursor, soot, and acetylene are

calculated as the harmonic mean of the kinetic rate, ,

and the turbulent mixing rate, ,

(9)

where the mixing rate came from the eddy dissipationmodel of Magnussen and Hjertager (1977),

(10)Having all reaction rates r i (i = 1 ~ 8), the rates of change

of mass fraction of species are calculated simultaneously,

(11)(12)(13)(14)(15)(16)(17)

This soot model is incorporated into the KIVA-3V code(Amsden, 1999) The specifications of the engine and theconditions for calculation are listed in Tables 2 and 3 (Yi et

al., 2000)

Two injection schemes, single injection and split tion, are considered in this study The scheme of 75.8.25described above was chosen because this scheme was

dY O2

dt - MWO2

MW PR

- r 5

N A

–

-=

Table 1 Rates of reactions used in the soot model

=

r 4 4.2 10 4 12

RT - –

=

Table 2 Engine specifications

reported to be an optimal scheme for soot emissions in the

literature (Han et al., 1996) Figure 2 illustrates the pulses of

fuel injection and injection timing of the two schemes The

amount of injected fuel and the total duration of injection

(excluding the dwell) of split injection are kept the same as

that of single injection

At the beginning, an EGR ratio of 20% is used, but it is

increased to 30% in the case of heavy EGR When OEA is

applied, the oxygen concentration in the intake air is

increased up to 23% by volume from the normal

concentra-tion of 21% by volume

Figure 3 shows the computational meshes of the

combus-tion chamber It includes 20 cells in the radial direccombus-tion, 26 in

the axial direction, and 30 in the tangential direction It

represents one-sixth of the engine combustion chamber for

computational effici-ency because the injector has six holes

and the combus-tion chamber is axisymmetric The

computation is started from intake valve closing (IVC) and is

ended at exhaust valve opening (EVO) timing It takes

approximately three hours to calculate one case with a

Pentium 4 PC

3 RESULTS AND DISCUSSION

Figure 4 shows the variations in pressure during the tion process of a single injection without EGR, and of splitinjection with 10% EGR These results are compared with

NOx and soot emissions of a single injection There is agood agreement between experimental data and calculateddata for all injection schemes To illustrate the general results

of modeling, the single injection with 20% EGR ratio is usedfrom now on The temperature is an important factor that

Table 3 Computational test conditions

(ATDC: after top dead center; BTDC: before top dead center;

CA: crank angle)

Figure 2 Schemes of split injection and single injection

Figure 4 Comparison of calculated and measured cylinderpressure: (a) Single injection without EGR; (b) Splitinjection with 10% EGR

Figure 5 Comparison of calculated and measured NOx andsoot emissions in case of single injection without EGR

EFFECTS OF SPLIT INJECTION, OXYGEN ENRICHED AIR 343

controls the emissions of diesel engines

Figure 6 shows the variation of average gas temperature in

the cylinder The temperature rises quickly with a steep slope

after the start of combustion (SOC) The increased rate of

temperature is then slower, and the temperature reaches a

temperature in the cylinder gradually decreases Figure 7

shows the normalized burned fuel as a function of crank

angle At 40oATDC, 99% of the fuel is burned Based on the

variation of burned fuel, and the temperature in the cylinder,

the combustion process is divided into two stages

The early stage of combustion is from the SOC to

approximately 40oATDC It is characterized by the rela-tively

high temperature The later stage of combustion is after 40o

ATDC until EVO During this period, the temperature in the

cylinder is decreased, and most chemical reactions of the soot

production process are frozen, as shown later

Figure 8 shows the variations of precursor formation,

precursor oxidation, precursor conversion into soot

particles, and net precursor All reactions are activated

strongly during the early stage of combustion due to the

frozen Most precursors formed are converted into soot

particles, and the rest are oxidized The mass of net

precursor is reduced to zero at the end of the combustion

process The variations of acetylene formation, acetylene

oxidation, acetylene growth with soot, and net acetylene areshown in Figure 9 The trend is similar to that of theprecursor These reactions are activated strongly during theearly stage of combustion and frozen during the later stage

of combustion, except that the growth reaction of acetylenewith soot is continued until the end of combustion Thegrowth reaction requires a relatively lower temperature toactivate, just above 1200K, during the later stage ofcombustion The peak of net acetylene appears at the samemoment as the peak of net precursor In contrast to theprecursor, acetylene still remains in the cylinder at the laterstage of combustion This fact is important because theconcentration of acetylene affects the rate of soot formation.Figure 10 shows the variations of soot formation, sootoxidation, and net soot of single injection with 20% EGR.Soot formation is the sum of soot inception from precursorand the growth of initial soot particles with acetylene, butthe latter process has considerable influence over the totalmass of soot formation Because the growth reaction ofacetylene with soot continues during the later part ofcombustion, the mass of soot formation continuouslyincreases during this period Most of the soot formed isoxidized, and the net soot is the result of soot formation andsoot oxidation When split injection is applied, the change ininjection scheme affects the temperature

Figure 11 compares the variation in temperature of

Figure 6 Variation of average gas temperature in case of

single injection with 20% EGR

Figure 7 Normalized burned fuel

Figure 8 Variations of precursor formation, conversion intosoot, oxidation, and net precursor: (a) Precursor formation,conversion into soot, and oxidation; (b) Net precursor

75.8.25 split injection with that of single injection Both

cases have an EGR ratio of 20% In the case of the split

fuel injections, the increase of average temperature in the

cylinder is not as smooth as that of single injection From

the start of combustion to the end of first injection at

After 10oATDC, stopping of the first injection results in a

slower increasing rate of temperature

When the second fuel is injected at 18oATDC, temperature

increases with a faster rate Although the same total amount

of fuel is used in both cases, the peak temperature of splitinjection is not as high as that of single injection because of acooling effect of dwell period and retarded combustion ofthe second injected fuel during the expansion stroke.Because the second injection retards combustion, the peak intemperature is shifted to the right and the temperature of splitinjection is higher than that of single injection during thelater stage of combustion Acetylene is formed from fuelpyrolysis due to high temperature during combustion Theformation of acetylene depends strongly on temperature.Figure 12 shows the variation in acetylene formation, which

is affected by variation of the temperature In the early stage

of the combustion process, the mass of acetylene formed inthe split injection is lower than that of the single injectionbecause of its lower temperature When the first injection isstopped, temperature decreases; thus, the mass of acetylenedecreases as well, and the first peak appears In the secondinjection, the increase in temperature causes the second peak

of the curve The appearance of the second peak and therelatively higher temperature during the later stage ofcombustion leads to the higher formation of acetylene withsplit injection observed in this period Figure 13 shows thevariations of soot formation, oxidation, and net soot Sootformation depends on the concentration of acetylene in thecylinder During the early stage of combustion, the increas-

Figure 9 Variations of acetylene formation, oxidation,

growth with soot, and net acetylene: (a) Acetylene

forma-tion, oxidaforma-tion, and growth with soot; (b) Net acetylene

Figure 10 Variations of soot formation, oxidation and net

EFFECTS OF SPLIT INJECTION, OXYGEN ENRICHED AIR 345

ing rate of soot formation with split injection is significantly

lower than that of single injection The lower acetylene

concentration in split injection is the reason why soot

formation is lower

During the later stage of combustion, although the

absolute value of soot formation in split injection is still

lower than that of single injection, the increasing rate of

soot formation in the split injection is higher compared

with that of single injection due to the relatively higher

acetylene concentration The increase in soot formation in

split injection is shown clearly in Figure 14 In this figure,

the distributions of acetylene and the rate of soot formation

at a typical moment in the later stage of combustion, here

concentra-tion at the bottom of the cylinder results in the increase in

soot formation rate at this region However, the increase in

soot formation is surpassed by the increase in soot

oxidation, which is discussed below Finally, the net soot is

reduced, and soot emissions from split injection are 15%

lower than that of the single injection, as shown in Figure

13(b) Figure 15 compares the stoichiometric area and the

soot oxidation rate of the two cases on the cross section at

oxidation is activated near a stoichiometric zone with

enough oxygen and high temperature The stoi- chiometric

in the case of split injection due to the second injection.When a piston moves downward during the expansionstroke, this newly injected fuel flows downward to thebottom of the piston bowl, while the “old” fuel stays in the

Figure 13 Comparison of soot formation, oxidation, and

net soot between single injection and split injection: (a)

Soot formation and oxidation; (b) Net soot

Figure 14 Comparison of acetylene and soot formationrate of single injection and split injection at 90oATDC: (a)Acetylene; (b) Soot formation rate

Figure 15 Comparison of soot oxidation rate at differentcrank angles: (a) 20o ATDC; (b) 50oATDC; (c) 90oATDC

upper part of the cylinder Consequently, as shown at

which help the oxidation of soot in a larger area and at a

higher rate than with single injection This situation

remains throughout the later stage of combustion, as shown

cause of lower soot emissions in split injection

When OEA is applied, the concentration of oxygen in

the intake air is increased from 21% to 22% in volume

Figure 16 shows the burned fuel of two split injections with

21% and 22% oxygen volume Due to the increase in the

concentration of oxygen in the cylinder, the mixing of fuel

with oxygen to form a combustible mixture becomes

easier Thus, more fuel is burned, resulting in higher peak

temperatures, as shown in Figure 17 Figure 18 shows the

variations of acetylene In the beginning, the mass of

acetylene of the OEA case is higher as expected due to the

higher temperature, but, in the later stage of combustion , it

is lower than that of the 21% case The transition takes

to the period during the second injection of fuel The

reason for this transition is explained in Figure 19, which

shows the rate of acetylene formation and the growth of

acetylene with soot for the two cases The reaction of

acetylene formation requires a temperature over 1700K to

activate, whereas the growth reaction requires a lowertemperature, just over 1200 K During the dwell period, thecylinder temperature drops, which results in the substantialreduction of reaction rates of acetylene formation andgrowth When fuel is supplied by the second injection, thetemperature is increased again, but it is not as high as thetemperature during the first injection This increasedtemperature is not high enough to enhance acetyleneformation, but it promotes the growth reaction of acetylenestrongly, especially for the OEA case due to its higher

Figure 16 Comparison of burned fuel between normal case

Figure 17 Comparison of average gas temperature between

Figure 18 Comparison of net acetylene between a normal

Figure 19 Comparison of reaction rate of acetyleneformation and acetylene growth with soot between normal

formation rate; (b) Acetylene growth with soot rate

EFFECTS OF SPLIT INJECTION, OXYGEN ENRICHED AIR 347

temperature

Because more acetylene is consumed in the growth

reaction, the net mass of acetylene of the OEA case is

reduced very fast during the second injection period

Consequently, the mass of acetylene of the OEA case is

kept lower than that of the normal case until the end of

combustion The decrease in acetylene in the OEA case

results in its lower soot formation The variations of soot

formation, soot oxidation, and net soot of the two cases are

shown in Figure 20 During the early stage of combustion,

the higher acetylene formation of the OEA case results inhigher soot formation During the later stage of combustion,because the soot oxidation rate of the OEA case is higherthan that of the normal case, the net soot mass of the OEAcase is reduced significantly in this period

Figure 21 shows the trend of soot and NOx emissions in

a diesel engine with different oxygen concentrations, from21% to 23% in volume When oxygen concen-tration isincreased, soot emissions are decreased, and NOxemissions are increased For example, in the case of splitinjection, when the concentration of oxygen is increasedfrom 21% to 22% in volume, soot is decreased 37% andNOx is increased 102% compared with case of 21%oxygen If the concentration of oxygen is increased to 23%

in volume, an increase of 266% NOx is found To avoid alarge increase in NOx, the level of oxygen enrichmentshould not be greater than 22% in volume This limitationhas been suggested by other researchers in the literature(Virk et al.,1993; Rakopoulos et al., 2004)

It is of interest to examine the simultaneous reduc-tion ofNOx and soot with the combined use of split injection, ahigh concentration of oxygen, and a high EGR ratio.Although the literature suggests an EGR ratio lower thanFigure 20 Comparison of soot formation, oxidation, and net

(22% O2): (a) Soot formation and oxidation; (b) Net soot

Figure 21 Trend of NOx and soot emissions with different

oxygen concentrations and injection schemes

Figure 22 Comparison of temperature for heavy, highEGR, and OEA cases

Figure 23 Comparison of NOx emissions for heavy, highEGR, and OEA cases

20% and OEA lower than 22% in volume, the contrary

effects of the two technologies offer the possibility of a

synergistic effect when com-bined together A combination

of split injection with 30% EGR ratio and 23% OEA in

volume (called “heavy case”) is tested in this study These

results are compared with the case of 30% EGR ratio and

22% OEA (called “high EGR case”), and the case of 20%

EGR ratio and 22% OEA (called “OEA case”, which was

tested previ-ously) Figure 22 compares the variations in

temperature of these cases The variation in temperature for

a split injection with 23% OEA without EGR is shown for

reference The two curves of the heavy case and the OEA

case almost coincide, except that the temperature of the

heavy case is slightly lower at the beginning of the

combustion process It is hypothesized that the higher EGR

ratio in the heavy case results in a longer ignition delay and

a slower increase in temperature at the beginning This

hypothesis is confirmed when the curve of the heavy case

is compared with the curve of the high EGR case Both

cases have the same EGR ratio, and thus have the same

variation in temperature during the premixed phase After

level of the OEA case, and the temperature of heavy and

OEA case are higher than the temperature of high EGR

case Obviously, the higher oxygen concentration in the

heavy case results in more burned fuel and compensates for

the effect of high EGR on temperature reduction during

this period Figure 23 shows the variations in NOx of the

three cases It is known that NOx is formed mainly in the

early stage of combustion and that NOx formation is

affected strongly by temperature The lower NOx

forma-tion of the heavy case again confirms that its temperature at

the beginning of combustion is lower than that of the OEA

case

The lower temperature in the heavy case causes a lower

acetylene mass during the early part of combustion, as

shown in Figure 24 Compared with the OEA case, the

peak is slightly delayed, which results in a slight increase in

acetylene mass during the latter part of combustion Figure

25 shows the variation in net soot At the end of

combustion process, the higher acetylene mass in the heavycase results in more soot formation, and thus a slightincrease in soot emissions In comparison with OEA case,the soot emissions of the heavy case are 2% higher Figure

26 compares the NOx and soot emissions of differentcases The combination of a high EGR ratio and a highoxygen concentration offers more benefits than the use of ahigh oxygen concentration only For example, the case of30% EGR ratio with 23% OEA reduces NOx emissions by16% and with nearly the same soot emissions as the case of20% EGR ratio with 22% OEA Compared to singleinjection, the combination of split injection, 30% EGRratio, and 23% OEA reduces soot emissions by 45% whileNOx emissions is remained the same

Figure 24 Comparison of net acetylene between heavy,

high EGR, and OEA cases

Figure 25 Comparison of net soot between heavy, highEGR, and OEA case

Figure 26 Trend of NOx and soot emissions with differentoxygen concentrations, EGR and injection schemes

EFFECTS OF SPLIT INJECTION, OXYGEN ENRICHED AIR 349

product of fuel pyrolysis and plays an important role in

controlling soot emissions

When split injection is applied, a dwell time between two

injections causes a decrease in local temperature, which

leads to a decrease in precursor formation The soot

formation in split injection is decreased compared with

single injection When fuel is injected by the second pulse,

the downward movement of the piston during the expansion

stroke guides fuel to the bottom of the piston bowl, resulting

in an increase in the stoichio-metric area of the cylinder

With the increase in temper-ature due to a secondary

combustion of second injected fuel, the soot oxidation rate of

split injection is increased The net soot mass of the split

injection is lower than that of a single injection as a

consequence of the lower soot formation and higher soot

oxidation Compared to single injection, the soot emissions

are reduced by 15% under a split injection scheme of

75.8.25 The decrease in temperature during the early stage

of combustion also results in a 40% decrease in NOx

emissions

When OEA is used, the increase in temperature leads to

an increase in the acetylene consumed in the growth

reaction of acetylene with soot, which results in a lower

concentration of acetylene and lower soot formation

However, the increase in temperature causes an increase in

NOx emission The combination of split injection and 22%

oxygen concentration in volume shows a 47% decrease in

soot emissions with a penalty of an 18% increase in NOx

emissions in comparison with single injection

A combination of split injection, high EGR ratio, and

high oxygen concentration is the best solution to reduce

diesel emissions The high EGR ratio keeps a lower

temperature at the beginning of combustion process, which

results in low NOx emissions The high oxygen

concentra-tion increases temperature in the later stage of combusconcentra-tion,

which compensates for the drawback of EGR on soot

emissions during this period In this study, a split injection

reduced soot emissions by 45%, with NOx emissions

remaining the same as in single injection

ACKNOWLEDGEMENT− This study was supported by the

KOSEF Project R01-2006-000-10932-O from the Ministry of

Science and Education, Republic of Korea.

REFERENCES

to KIVA-3V Los Alamos National Laboratory Report,

LA-UR-99-915

Bakenhus, M and Reitz, R D (1999) Two-color

combus-tion visualizacombus-tion of single and split injeccombus-tions in a

single-cylinder heavy-duty D.I diesel engine using an

1999-01-1112

Chan, M., Das, S and Reitz, R D (1997) Modeling

multiple injection and EGR effects on diesel engine

Fusco, A., Knox-Kelecy, A L and Foster, D E (1994).Application of a phenomenological soot model to diesel

SAE Paper No. 980524

Han, Z., Uludogan, A., Hampson, G J and Reitz, R D.(1996) Mechanism of soot and NOx emission reduction

Hiroyasu, H., Kadota, T and Arai, M (1983) ment and use of a spray combustion modeling to predictdiesel engine efficiency and pollutant emissions Bulletin

Develop-of JSME, 26, 569−575

Kazakov, A and Foster, D E (1998) Modeling of sootformation during DI diesel combustion using a multi-

Lee, S S., Sung, N W., Kim, H S and Lee, J M (2007).Analytical study on soot formation in a diesel engine J Mech Science and Technology, 21, 113−122

Magnussen, B F and Hjertager, B H (1977) Onmathematical modeling of turbulent combustion with

16th Symp (Int.) Combustion, The Combustion Institute, Pittsburgh, 719−729

Nagle, J and Strickland-Constable, R F (1962)

Conf.,Pergamon, London

Nehmer, D A and Reitz, R D (1994) Measurement ofthe effect of injection rate and spilt injections on diesel

Pierpoint, D A., Montgomery, D T and Reitz, R D.(1995) Reducing particulate and NOx using multiple

950217

Pratsinis, S E (1987) Simultaneous nucleation,

Interface Sci., 124, 416−427

Rakopoulos, C D., Hountalas, D T., Zannis, T C andLevendis, Y A (2004) Operational and environmentalevaluation of diesel engines burning oxygen-enriched

Stork, K and Poola, R (1998) Membrane Based Air

Com-position Control for Light-duty Diesel Vehicles: A

Benefit and Cost Assessment. Argonne National Lab

Report ANL/ESD/TM-144

Uchida, N., Daiso, Y., Saito, T and Sugano, H (1993)

Combined effects of EGR and supercharging on diesel

Virk, K S., Korturk, U and Bartels, C R (1993) Effects

of oxygen-enriched air on diesel engine exhaust

931004

Wagner, R M., Green, J B., Storey, J M and Daw, C S.(2000) Extending exhaust gas recirculation limits in

Exposition, Salt Lake City,U.S.A

Yi, Y., Hessel, R., Zhu, G S and Reitz, R D (2000) Theinfluence of physical input parameter uncertainties onmultidimensional engine performance and emissions

SAE Paper No. 2000-01-1118

Zhang, Y and Nishida, K (2003) Vapor/Liquid behaviors

in split-injection DI diesel sprays in a 2-D model

International Journal of Automotive Technology , Vol 12, No 3, pp 351 − 358 (2011)

351

PISTON-TEMPERATURE MEASURING SYSTEM USING

ELECTROMAGNETIC INDUCTION WITH VERIFICATION

BY A TELEMETRY SYSTEM

H H LEE 1)* , H W BANG 1) , S K KAUH 1) and S I KIM 2)

Hwaseong-si, Gyeonggi 445-706, Korea

(Received 24 July 2009; Revised 1 August 2010)

ABSTRACT− The development of an inner-piston-chamber temperature measurement system is a necessary step in engine development or when solving other fundamental problems related to automotive engines There are various pre-existing measurement methods available, e.g., the linkage method, piston telemetry, templog, and the electromagnetic induction method In this study, we first redesigned the coil sensor used in the electromagnetic induction method using PEEK and then used Taguchi methods to reduce the number of experiments in the development process and finally utilized piston telemetry via Bluetooth to verify the precision and accuracy of the redesigned PEEK coil sensor and electromagnetic induction method The results displayed a reproducibility within 0.5 degrees and an accuracy within 2 degrees Celsius.

KEY WORDS : Piston temperature, Electromagnetic induction, Taguchi methods, Coil sensor, Piston telemetry, Temperature sensor, Thermistor, Ferrite, PEEK, Vehicle engine

1 INTRODUCTION

The pistons are the engine parts with the largest thermal

load, which decreases the tensile strength and hardness of

the piston material and simultaneously increases abrasion

of the piston ring by decreasing the oil-film thickness

between the piston ring and the liner Thermal load on the

piston is also known to cause aluminum cohesion on the

top-ring groove These problems cause undesirable side

effects, primarily a rise in the engine temperature For these

reasons, a temperature-measurement system should be

installed in the inner-piston chamber in the evolution of

engine design or when correcting engine flaws However,

the current limitations of piston-temperature measurement

systems make accurate measurements a difficult task The

difficulties are mainly due to piston revolution speeds over

6,000 r/m, accelerations over 2,000 ×g and temperatures in

excess of 300 degrees Celsius

Piston-temperature measurement is currently available

using a number of methods A traditional measurement tool,

the ‘linkage method’, utilizes a thermocouple to enable the

most accurate temperature measurement but a disadvantage

is that many engine parts need to be redesigned and a rather

long preparation time is required Templog, has a short and

simple preparation but the disadvantages are that it can only

measure a maximum temperature and its error range is quitelarge The telemetry method is used for more accuratetemperature measurement and does not require manychanges in engine structure design, but its disadvantage isthe difficultly in supplying power The electromagneticinduction method used in this study requires minimalchanges in hardware configuration, which enables a shorterpreparation time and requires less work Four to six real-time channel measurements per cylinder and an in-vehiclemeasurement are possible Compared to the linkage method,the number of measurable channels for the electromagneticinduction method is less than the linkage method; however,the measurement accuracy is almost on par with the linkagemethod and is much more durable, and the method does notrequire a long preparation time Because of these advantages,the electromagnetic induction method is widely utilized inthe automotive research field

Normally, ferrite is used for the core in coil sensors in theelectromagnetic induction method However, because thepermeability of ferrite varies greatly with the temperature ofthe surrounding environment, PEEK was chosen as the corematerial for this study To facilitate the design of anoptimized sensor taking into consideration piston vibrationand temperature, a simple experimental method, termed

‘design of experiments’, was utilized The selection ofexperiments performed was based on ‘tables of orthogonalarrays’, which is a fractional factorial design The final goal

*Corresponding author. e-mail: mighgus@naver.com

of this research was to verify the developed coil sensor with

piston telemetry utilizing a battery-powered board

2 TEMPERATURE-MEASUREMENT SYSTEM

2.1 Principle of the Electromagnetic Method

The electromagnetic method utilizes Faraday’s theory of

electromagnetic induction As the distance d between coils

L1 and L2 shown in Figure 1 decreases, the current I1 in the

primary coil L1 induces voltage in the secondary coil L2

This induced voltage in the secondary coil and its resistance

R2 causes a current I2 to flow, and this current in turn

induces a voltage in the primary coil As a result of this

induced voltage in the primary coil, the current in L1 varies

sinusoidally and causes a voltage amplitude change

between R1 and L1 This voltage amplitude change

between R1 and L1 varies with the resistance of R2 In the

sensor, R2 is equivalent to the resistance of the thermistor,

which varies with temperature Thus, temperature is

measured by estimating the resistance of R2 by measuring

the voltage amplitude change between R1 and L1

2.2 System Design of the Electromagnetic Method

The system consists of two major parts: the sensor and the

measuring equipment This study utilized a six-cylinder

engine and a 24-channel instrument in a program

develop-ed in 2006

2.2.1 Sensor component

Figure 2 is the schematic design of the sensor The primary

coil corresponds to L1 of Figure 1 and was installed on the

engine block The secondary coil corresponds to L2 of

Figure 1 and was installed on the lower part of the piston

When the piston reaches bottom dead center (BDC), the

primary coil is inserted into the secondary coil without

contact

Formerly, ferrite, with its high permeability, was used as

a core for the electromagnetic induction method However,

as permeability itself varies with temperature, measurement

results can differ greatly at the same temperature when

taken at different time points Figure 3 shows the change in

permeability versus temperature for ferrite Typical

engine-room temperature exceeds 100oC, and permeability decreases

steeply around this point

To resolve this temperature-related issue, PEEK wasutilized as the core material of the sensor coil for theelectromagnetic induction method PEEK is durable againstheat and abrasion, as verified by four independent engineexperiments Changing the core of the coil sensor fromferrite to PEEK resulted in a decrease in error due to thetemperature inside the engine room

The calibration tool maps temperature-voltage characteristics

to correct each sensor This tool consists of a thermostat forheating the thermistor and a thermocouple for measuringthe reference temperature In this paper, a commercialthermostat was used Figure 4 is a representation of thecalibration data measured with the calibration tool Thesedata are saved in the form of tables consisting of a voltagecolumn and temperature columns Each sensor has its own

Figure 1 Principle of the electromagnetic method

Figure 2 Schematic design of the sensor

Figure 3 Permeability of ferrite

Fiugre 4 Temperature vs voltage in the coil sensor

PISTON-TEMPERATURE MEASURING SYSTEM USING ELECTROMAGNETIC INDUCTION 353

calibration table for use in the measuring program

3 ROBUST DESIGN OF ELECTROMAGNETIC

COIL SENSOR USING TAGUCHI METHOD

A number of parameters associated with the coil sensor can

vary during optimization experiments In this study, eight

parameters were evaluated and the vibration of the engines

and the temperature inside the engine room were taken into

consideration as a function of background noise When

considering all the parameters in performing the

optimiza-tions, even if only three of the cases are assumed for each

in this research, the Taguchi method or “design of

experiments” was used

3.1 Outline of Robust Design Using Taguchi Method

In the Taguchi method, S/N Ratio (signal-to-noise ratio) is

considered crucial in providing robust characteristics i.e.,

the design is more robust against noise as the S/N ratio

increases Static and dynamic characteristics can be

considered when performing an optimization experiment

Among dynamic characteristics, temperature was used as a

signal parameter because voltage-amplitude variations are

proportional to temperature, as shown in Figure 4 For

dynamic characteristics, the results seen in Figure 5 show

that Experiment #15, with a higher S/N ratio of -17.9 dB,

was considered more robust than Experiment #5, with an S/

Experiment #15 Therefore, in the analysis of dynamiccharacteristics, temperature accuracy was not properlyreflected and the direction of the optimization wasdetermined to decrease the dispersion of the system Forthese reasons, the static characteristic “smaller-the-betterwas used instead of the dynamic characteristic

Using the static characteristic “smaller-the-better shown

in Figure 6, the ideal function was determined in a T-Vdiagram and the difference, shown as gaps in Figure 6,from each experiment was calculated as a minimum Thesmaller-the-better characteristic was applied here becausethe difference between the ideal function and each voltagecharacteristic of the experimental coils should be small toyield better performance In this analysis, Equation (1) wasused to calculate the S/N ratio, which was calculated foreach experiment to determine the best control parameter

(1)3.2 Experimental Equipment

An experimental equipment component diagram is shown inFigure 7 The variable resistor placed in the secondary coil

S N ⁄ i 10 1n - y ij2

j

∑

log –

=

Figure 5 T-V diagram (dynamic characteristic)

Figure 6 T-V diagram (smaller-the-better characteristic)

Figure 7 Experimental equipment component diagram.Table 1 Control parameters used; orthogonal-arrangementtable

B primary coil turns 210 turns 180 turns 240 turns

C ratio of primary coil turns to

was slowly varied and the resulting voltage displacement in

the primary coil was measured Through this process, an R-V

table was acquired and, using a T-R table for the thermistor, a

T-V diagram was then generated (Figure 4)

3.3 Control Parameters and Noise Parameters

In our system, it was important to investigate whether the

level of each parameter acts reciprocally or if the action is

rather small enough to ignore As shown in Table 1, the

oscillating sine-wave parameter was divided into two

levels and the other seven parameters were divided into

three levels, i.e., an L18 (21 × 37) orthogonal table was used

Here, G (the relative contact ratio) represents the position

of the primary coil relative to the coil contact with the

vertical center of the secondary coil

Among many noise parameters, engine vibration and the

temperature of the coil sensor were chosen as the main

noise parameters To simulate engine vibration, the

eccentricity of the coils was classified into two levels If the

position of the primary and secondary coils was concentric,

it was classified as N1, whereas if their relative position was

eccentric it was classified as N2 Similarly, the temperature

of the coil sensor was also classified into two levels

Maintaining primary and secondary coil temperatures at

25oC was classified as Q1 and at 120oC as Q2

3.4 Orthogonal-arrangement Table Analysis by the

‘Smaller-is-better Characteristic’

The fractional factorial experimental design was done

using the orthogonal-arrangement table As mentioned

above, the difference between each case and the ideal

function was calculated Specifically, referring to the

resistance characteristic table of the thermistor, the resistor

values for low-temperature, median-temperature and

high-temperature ranges were obtained and the S/N ratio was

calculated for each range Next, the Taguchi Method was

applied using the smaller-the-better characteristic and, as a

result, the minimal difference between the ideal function

and each experimental case was determined

The optimal combination of coil sensors for low

temperature, as shown in Figure 9, was A2B1C2D3E1F3G1H2,

and the same combination applied for median temperature

(Figure 10) The optimal combination of coil sensors for

high temperature was A1B1C2D3E1F3G1H1 (Figure 11)

Comparing the three cases confirmed that among the

control parameters shown below, C, F and G were the

dominant ones for the performance of the coil sensor.Considering only the oscillating sine wave (parameterA), there was a case were the S/N ratio was higher at 700kHz than at 600 kHz Nevertheless, as the frequency of theoscillated sine wave decreased it was observed that the peak

of the voltage displacement appeared when the variableresistor value was small (2007 Korean Society ofAutomotive Engineers Autumn Conference CollectedPapers Vol 1, 358–364) Because an NTC-type thermistorwas used for measuring the high-temperature ranges, there

is a greater advantage to set the frequency at 600 kHz.Considering the height of the primary coil (parameterD), the resulting S/N ratio was highest when the value was

9 mm However, this is a risk factor considering thestructure of an engine block because as the height of theprimary coil is increased the thickness of the jig holding thecoil must decrease, thereby weakening it Thus, the designheight should be kept to within 7 mm

Finally, considering only coil thickness (parameter H),using a coil with a diameter of 0.08 mm resulted in a higherS/N ratio in the median- and low-temperature ranges.Taking into consideration the manufacturability of the coilsensor, curls and cut-offs due to stiffness are majorobstacles for commercialization Thus, commercialization

of the coil sensor was considered when a coil diameter of0.1 mm was chosen for the design in this study

Summarizing the above results, the overall optimalcombination of control parameters for the coil sensor wasdetermined to be A1B1C2D1E1F3G1H1 This optimalcombination can be applied to all temperature ranges.Figure 8 First and second coil in the jig

Fiuger 9 Optimal combination for low temperature

Fiuger 10 Optimal combination for median temperature

PISTON-TEMPERATURE MEASURING SYSTEM USING ELECTROMAGNETIC INDUCTION 355

To verify the performance improvement of the optimal

combination, additional experiments were done, and the

results are shown in Table 2 For Case 1, an S/N ratio

increase of 11.4 dB was observed for the predicted value, as

depicted by the gain shown in Table 2, and an S/N ratio

increase of 9.0 dB for was observed for the verified value

(Table 2) This result shows an improved gain and that most

of the parameter effects on dispersion were reproduced For

Cases 2 and 3, the results were similar, verifying that the

results are credible

The plot in Figure 12 compares the performance of an

existing coil sensor and an optimized coil sensor from Case

3 It is clear that the sensing performance of the optimized

coil sensor was less dispersed against noise and closer tothe ideal function

4 RIG TEST OF ELECTROMAGNETIC INDUCTION METHOD USING PISTON TELEMETRY

4.1 Setup of the Rig System4.1.1 Design of piston telemetry

To verify the optimal specification for the coil sensor fromChapter 3, a telemetry system was used in this study Thetelemetry system consists of a data-acquisition module, apiston and a data-receiving module The data-acquisitionpart consists of a power supply, an ADC, a Bluetooth PCBand an aluminum case and is installed on the bottom of theconnecting rod The piston part consists of a piston and athermistor installed on the piston and is connected to thedata-acquisition part via an A-wire above the connectingrod The data receiver consists of a Bluetooth dongle and a

PC and processes temperature data in real time

To compensate for the inferior durability and longerpreparation time, which are the disadvantages of thetelemetry method, a thermistor was used as a temperaturesensor When using a thermocouple, it must be directlyconnected to the ADC input of the data-acquisition moduleinstalled at the bottom of the connecting rod due to cold-junction compensation, resulting in more setup time andlower durability Moreover, we revised ADC errors due totemperature increases inside the engine room to guarantee

Fiuger 11 Optimal combination for high temperature

Table 2 S/N ratios of optimal specification

Figure 12 Optimum design (T-V Diagram) for Case 3

Figure 13 Schematic design of the piston-telemetry system