Laser velocimeter measurements in a turbocharger turbineE F Benisek and A G Struble

1C405/009

Flows in vaned and vaneless stators of radial-inflow turbocharger turbinesN C Baines and M Lavy

7C405/005

A method to predict performance of vaneless radiai turbines under steady andunsteady flow conditionsH Chen and D E Winterbone

C405/00813

23Investigation of the flow characteristics of radiai inflow turbocharger volutesA Whitfield and A B Mohd Noor

C405/023

Alternative turbocharger systems far the automotive diesel engineB E Walsham

39C405/036

Double resonance system-a new way to improve the low-speed operation of. . .C405/013

51

C405/05459

Effects of controlled turbine bypass OD a commerciaI diesel turbochargedengIneM W G Brown and A B Homer

C405/03771

Optimal control of an integrated engine transmission with controllablepressure chargingJ Hall and F J Wallace

C405/02283

The transient response of a turbocharger vehicle diesel engine with injectiontiming controlC Arcoumanis, Z Bazari and S H Chan

C405/03899

On-road perfornlance of two alternative transmission systems for heavyvehiclesF J Wallace and Z Dequan

Pulsating flow behaviour of a twin-entry vaneless radial-flow turbineJ H Yeo and NC Baines

C405/004

Unsteady flow performance of turbocharger radiaI turbinesM Capobianco and A Gambarotta

C405/017

Development of a model to predict the vibration response of a largeturbocharger rotor under pulsating flow conditionsG M Chapman and J B Turnbu//

C405/019

A compact two-stage turbocharger moduleD WHTennant

C405/042

C405/015 Measurement of the performance of a radiaI inflow turbine in conditionalsteady and unsteady flowD E Winterbone, B Nikpour and G I Alexander

Development work on turbochargers for two- and four-stroke diesel engineswith and without turbo-compound systemsM Appel and J Bucher

C405/053163

The interaction of diesel engine turbocharging and tuned inlet manifoldsystems under steady speed and transient operationK Banisoleiman, L A Smith and B A French

C405/034

Application of variable valve timing to a highly turbocharged diesel engineS J Charlton, A J Keane, H J Leonard and C R Stone

189C405/044

supercnargea engmesG Cser

Investigation of the flow phenomena in the inlet of an unshrouded centrifugaIcompressor impeller operating at part loadR Girsberger and K H Rohne

C405/041

C405/040 A simple method for predicting centrifugaI compressor performancecharacteristicsE Swain

207

C405/024219

Turbocharger compressor developments far broad range and high pressure

229Turbocharging for submarines-a special caseE T F Kirkman and RA Hopper

C405/045

C405/012241

The role of digital dynamic simulation in the design of a turbocharger teststandD A Penz, C D Mullinax, M Y Young and A G Stroble

255A severe turbocharger bench engine test for lubricant evaluationG B Toft and R Mainwaring

C405/011

267Towards a more reliable turbocharger for rail tractionK Dunn and U Gribi

C405/021

277The application of ball bearings to automotive turbochargersJETB/ake

C405/049

283C405/057 Experiences with supercharging the Porsche 944 engineH Richter and N Hemmerlein

289C405/032 Comprex with gas pocket controlA Mayer, lEI Nashar and J Perewusnyk

295C405/060 Ceramic rotors for passenger car turbochargersB Engels and R Lingenauber

C405/025 Study of cerarnic rotor and turbine housing materia! of a turbocharger up tothe turbine inlet gas temperature lO50.CY Miyagi, K Miyashita, H Sugihara and T Tomita

317C405/055 Turbocharging techniques for sports car enginesC Briistle, J Wagner, K Tran Van and K Burk

329The benefits of variable geometry turbocharging tram the military aspectJ R Starkey and P C Franklin

C405/043

339Turbochargers and the military vehicleD P Hartwell

C405/052

ratto applicationsD Flaxington and B Mahbod

Unsteady flow performance of turbochargerradiai turbines

M CAPOBIANCO and A GAMBAROTT ADepartment of Energetic Engineering, University of Genoa, Italy

SYNOPSIS The unsteady flow performance of turbocharger radial turbines are affected by thecharacteristics of the pulsating flow. The paper presents the results of an extensiveinvestigation developed at the University of Genoa. A new arrangement of the turbocharger testrig and an original injection device to measure turbine characteristics almost to zero mass flowconditions are described.Steady and pulsating flow performance of a wastegated turbocharger turbine were measured. Theinfluence of the main parameters of the pulsating flow on average turbine mass flow and torquewere analysed. Additionally, a wide range of quasi-steady flow methods were developed, anddifferent procedures far calculating turbine overall performance were proposed and tested.

INTRODUCTION flow and power starting from measured steady-flow curves and pressure diagrams wereconsidered. The results produced by theseprocedures and their deviations from meanmeasured values are discussed in this papero

2 EXPERIMENTAL SYSTEM

The DINE turbocharger test facility was fullydescribed in previous papers (6, 7). Thisapparatus operates at low air temperature at theturbine inlet (up to 400 K). Both turbine andcompressor, which is used as a dynamometer, canbe supplied with compressed air at controlledpressure levels by an appropriate piping system.

The present widespread utilization of radiaiflow compressors and turbines far turbochargingautomotive engines needs a better knowledge oftheir behaviour under steady and pulsating flowconditions, in arder to optimize matchingcaIcuIations and increase engine efficiency.Referring to the radiaI turbine, the resuIts ofsome extensive theoretical and experimentalstudies have been presented in the Iiterature(1, 2,3,4,5). Test rig investigations seemparticuIarly suitable to highIight how thecharacteristics of unsteady fIow (frequency,pulse amplitude and mean vaIue) affect turbineperformance.

A specific research programme on thesetopics is in progress at the Department ofEnergetic Engineering of the University of Genoa(DINE). Following the first results, reported ina recent paper (6), the development of theactivity is here presented.

Rotating valve pulse generators installedupstream of the turbine produce unsteadiness inthe case of pulsating flow tests. The flow areadiagram produced by the pulse generator can bemodified by replacing the rotar and the statorports of the device, and the pulse frequency canbe adjusted within the typical range ofautomotive engines exhaust systems (6).

The previous stage of the investigationsuggested some improvements of the DINEturbocharger test facility.

A new supply line was then set up, by whichtests can now be made on single and two-entryturbines with independently controlled pulseparameters far each entry.

Furthermore, an air injection system actingon the compressor side of the turbochargerallowed to measure the turbine characteristicsclose to zero mass flow conditions withoutremoving the compressor wheel.

The arrangement of the turbine feeding linewas recently modified. In fact, during previousactivities, a connection between pulseamplitude and mean pressure level at the turbi-ne entry was observed (6). This was due to thecharacteristics of the pulse generator system.

In arder to determine the effect of pulseamplitude and mean value on turbine unsteadyflow performance, a new organization of thesupply circuit has been studied.

Tests with full, partial or unequaladmission can now be performed on single andtwo-entry turbines, both in steady and pulsatingflow conditions. Besides the pulse amplitude andphasing can be controlled and modified.

A broad experimental investigation wascarried out on a small automotive turbocharger,fitted with a waste-gate valve integrated in theturbine housing. Tests were performed in arderto analyse the influence of the main pulsatingflow parameters and of the waste-gate opening onturbine behaviour. The main eiements of the new turbine supply

circuit are shown in fig.1. The air filteringand pressure contro 1 devices are foiiowed by thefiow rate measuring station (fitted withpositive-displacement meters) and by the

A quasi-steady flow approach was used topredict turbine unsteady performance. Differentmethods far calculating overall pulsating mass

123C405/017 @ IMechE 1990

3 STEADY-FLOW PERFORMANCEelectrical air heater. Then a small reservoiracts as a damping element and flow distributor.For each turbine entry two independent ducts areprovided, in one of which a rotating valve isinserted. The pulse amplitude and mean pressurecontro l is got by modulating the steady andpulsating flow components through dedicatedvalves. Phasing between the two pressure pulsescan be easily modified (by changing the relativeangular position of the pulse generator rotors).

The volume between each rotating valve andthe corresponding turbine entry is limited toreduce damping effects (approximately 0.002 m3).

Steady-flow characteristics were measured forconstant values of turbine speed factor (n/~TS)and waste-gate opening.

In arder to define significant waste-gatepositions, a specific testing method wasdeveloped. FirstIy, the fIow through the turbinewheel was shut off and the pneumatic by-passcontro l was disconnected. The waste-gate massfIow characteristics were measured far severalpositions of the valve rod (9).

Some tests (developed at the Centro RicercheFiat) showed that the valve opening range inreal on engine operation was quite wide. FivefiKed waste-gate positions were then consideredin the investigation, corresponding to O, 25,50, 75 and 100 per cent of the total by-passflow area (A). To avoid inaccuracies due tovibrations of the valve plate, the waste-gateassembly was removed and the above valvepositions were replicated by equivalent nozzles,fastened to the turbine housing (fig. 3). Foreach waste-gate opening several values ofturbine speed factor were considered, rangingfrom 2500 to 5500 rpm/4K.

Since the DINE test rig operates at low airtemperature at the turbine inlet, the compressorsupply pressure is controlled to modify itspower absorption <.at constant rotational speed).By this technique it is possible to investigateturbine characteristics up to higher expansionratios, as in the case of "hot air"measurements.

Nevertheless turbine steady-flow curves wereoften extrapolated in the portion of the lowestmass flow rates far their use in quasi-steadyflow calculations (6).

In arder to prevent this appreciableinaccuracy (3), a new experimental device hasbeen developed. The system is similar to oneformerly proposed to get a better transientresponse of turbocharged Diesel engines (8).

Fig. 4 shows some constant-speed steady massflow characteristics vs. turbine expansionratio, far the different equivalent waste-gateopenings.

Turbine steady-flow curves were measuredaver a wide operating range. Higher expansionratios were reached by controlling thecompressor supply pressure, while the airinjection device allowed an appreciableextension of investigation almost to zero massflow (see the characteristics referred to O and25 per cent by-pass area in fig. 4).

The compressor housing was replaced by a newdedicated casing, fitted with three injectornozzles, fed with compressed air (fig. 2). Nomodifications of the turbocharger rotaryassembly were necessary.

The nozzle position was set so as to directthe air impulse to the compressor rotar blades,keeping the usual direction of rotation. Thepositive torque contribution on the turbineshaft can be modulated by adjusting the nozzlesupply pressure, and therefore the air flow ratethrough the injection system. By this impulsivedevice it was possible significantly to extendthe measurement of the turbine characteristicsnear zero mass flow conditions (see fig. 4).

The turbine ma 55 flow rate obviouslyresulted higher far larger wa5te-gate opening5,at the same expan5ion ratio. However a deepertheoretical and experimental analysis on thi55ubject is now in progress to point out thepo5sible flow interaction between the turbineimpeller and the wa5te-gate. The results of theinvestigation will be pre5ented in a futurepaper (9).

The investigation was carried out on aGarrett TBO25 turbocharger, usually coupled tosmall automotive Diesel engines (totaldisplacement about two litersJ. The turbochargeris fitted with a single entry nozzleless radiaIflow turbine. The waste-gate valve has its seatintegrated in the turbine exhaust housing.

4 UNSTEADV FLOW RESULTS

Measurements were performed both in steadyand pulsating flow conditions by the automaticdata acquisition system of the DINE test rig,controlled by an IBM-AT computer through apurpose-developed software (6, 7).

Measuring stations were located close tothe inlet and exit of the turbine, as shown infig. 1: the relevant pressure transducers weremounted near the duct walls, in arder to avoidsignal modifications due to the connecting pipesin unsteady flow conditions. The instantaneouswall static pressures at the turbine inlet andexit were recorded and stored by a digital highsampling frequency oscilloscope (2 MHz),integrated in the data acquisition system.

In a previous stage of the research programme(6), a generaI increase of mean turbineperformance when operating in unsteady flowconditions was observed, but it was impossibleto separate the effect of pulse amplitude andmean value as these quantities resultedcorreIated during the tests. The new arrangementof the turbine feeding line permitted toovercome this drawback.

The influence of the waste-gate valveopening was also considered, referring both toupstream and downstream pressure waves and toaverage pulsating turbine performance.

For each test the time-dependent pressure(static) diagrams at the turbine inlet (ps) andexit (P4) were evaluated by straingaugetransducersj this permitted to calculate theinstantaneous and the mean level of theexpansion ratio. The measured values of turbineinlet temperature, rotational speed and mass

@ IMechE 1990 C405/017124

flow rate were assumed as average levels in thepulse periodo Mean turbine torque and power werecalculated on the basis of measurements on thecompressor side of the turbocharger (6).

The influence of pulse frequency on turbineaverage pulsating performance wasn't evident,especiaIIy in the case of mass fIow (ref.fig.6b). This is probably connected to theobserved modifications of inIet pulse shape withfrequency, due to the wave action in the ducts.Nevertheless, particuIarly far turbine torque, ageneraI trend towards steady-fIow levels far thehighest pulse frequencies was observed, atconstant amplitude of inIet pressure diagrams(see fig. 6d).

At constant inlet pulse amplitude and meanvalue, significant modifications of the upstreamwaveforms were observed far different frequencylevels (fig.Sa). This aspect, already pointedout in previous investigations (6, 7), seems dueto the wave action in the turbine feeding line.A notable amplitude of the turbine downstreampressure signal was found in a few conditions,near the natural frequency of any portion of theexhaust turbine circuit (see, far example, thevalue of 110 Hz in fig. Sa).

The effect of inlet pulse amplitude onturbine mean unsteady performance, far constantvalues of other operating parameters, is shownin fig. 6a and 6c, respectively far mass flowand torque.

At each frequency level, lower differencesbetween the average pulsating values and thecorresponding steady flow performance weremeasured with increasing pulse amplitude. Forthe highest considered level of £1p3 (0.66bar) mean unsteady mass flow and torque oftenproved to be slightly below steady flow values.

The puise amplitude contro l actuated by theexperimentai system of the DINE test rig didn'taffect the shape of the pressure diagrams, as isshown in fig 5b far the same conditions offrequency, rotational speed and average inietpressure.

As regards the effect of the waste-gatevalve opening on the time-dependent pressuresignals, the wave form at the turbine entry wasalmost the same far different by-pass areas, asis apparent in fig. 5c, with reference tocorresponding operating conditions (i.e. far thesame levels of turbine rotational speed, pulsefrequency, amplitude and mean value).

The exhaust pressure diagrams, on thecontrary, presented a clear noise (the frequencyof which was estimated, through a Fourieranalysis, at about 3 kHz) in the case of waste-gate opening, more evident far higher by-passareas. This aspect, shown in fig. 5c, isprobably due to the interference between theflow components from the impeller and the waste-gate valve at the turbine exit. In authors'opinion this result is a consequence of theparticular geometry of the turbocharger, fittedwith a by-pass system integrated in the turbineex~aust housing without any mechanicalseparation between the two flow portions (seefig. 3).

The observed influence of inlet pulseamplitude on average turbine performance can beexplained if a quasi-steady behaviour of turbineoperating under pulsating flow conditions issupposed.

In fact, in the above assumption, anincrease of the expansion ratio oscillationamplitude, at the same mean value, causes lowerlevels of average performance, due to thevariable slope of the steady flowcharacteristic.

This conclusion was confirmed by the rise ofturbine mass flow and torque ratios (K)generally measured far higher mean inletpressures, at constant pulse frequency andamplitude.

As regards the effect of the waste-gatevalve opening, far small by-pass areas (25 percent of full opening>, the differences betweenthe mean pulsating and the corresponding steadyflow turbine performance and the influence ofpulse frequency and amplitude resulted similarto those observed far fully closed valveconditions.

Several tests were carried out in arder toinvestigate the influence of the main parametersof the pulsating flow on turbine mean

performance.Since the average exhaust pressure result

was affected by the operating conditions(particularly by the waste-gate opening>,turbine mean pulsating performance was relatedto steady flow results far constant values ofaverage expansion ratio and rotational speed.

With the above assumption, some significantresults are presented, referring to the ratiosKM and K~ between mean pulsating mass flow andtorque and the corresponding steady flow values.

A few tests performed in the case of largerwaste-gate openings showed a greater increase ofturbine unsteady performance, particularly fartorque. This might be connected to a flowmodification in the turbine exhaust mixingchamber when operating in pulsating flowconditions, caused by wave propagation throughthe waste-gate port in the case of largeopenings. The observed differences in outletmean pressure levels may be a signal of thisaction.

A deeper investigation on this matter isactually in progress at DINE, the results ofwhich will be covered by a future papero

5 QUASI-STEADY FLOW ANALYSIS

Following a previous work (6), an extensivestudy on quasi-steady flow techniques has beendeveloped. Instantaneous turbine mass flow andpower were evaluated according to the methoddescribed by Benson (3), and several procedureswere used to calculate overall performanceparameters.

Turbine unsteady mass flow and torqueresults Nere generally higher than the steadyflow levels, as was observed in a previousinvestigation on a different turbocharger (6),but performance increase proved to be slightlylower.

A typical result is reported in fig.6, wherethe influence of pulse frequency (f) and inletamplitude (~P:a) is shown. Three differentlevels of each of the above parameters areconsidered, relating to significant on-engineconditions. The situation of fully closed waste-gate valve is presented.

C405/017 @ IMed1E 1990

characteristics at increasing turbine expansionratio. Average experimental pulsating levels, onthe contrary, resulted generally higher than thesteady flow ones. This result may suggest agreater effect of the pulse intervals in whichthe instantaneous turbine expansion ratio isabove the mean level.

To take into account this aspect, a firstalternative procedure far calculating averageperforma,nce has been developed, based on theevaluation of turbine mean pulsating parametersas weighted average values aver the pulseperiodo The instantaneous expansion ratio (£~,)has been considered as weight in calculations.

Overall mass flow and power were thencalculated by the following relations:

Starting from pressure diagrams measuredupstream (ps) and downstream (P4) of theturbine, the pulse period was divided into Nequal intervals and the instantaneous expansionratio E~l (static/static) far each of them wasdetermined. The static inlet temperature Ts,was assumed to be related to the mean measuredvalue through the isentropic relationship (3):

- -TS1 = Ts ( PSl I P3 )'k-1)/"

N

E (~"..M".>t-lI I

M.. QSF =

The speed parameter n/~31 was evaluatedassuming the measured level of turbinerotational speed as a reliable mean value averthe pulse periodo Instantaneous levels ofturbine mass flowand power were determinedreferring to the relevant steady flowcharacteristics. As regards power, it wasassumed that the bearing losses were unchangedin pulsating flow operation far the same turbinerotational speed, oil pressure and temperatureat the turbocharger inlet (6).

Since measured turbine steady flow curveswere e~tended to very low e~pansion ratios, itwasn't necessary to e~trapolate the e~perimentalcharacteristics.

N

L Etl1-1

andTurbine overall performance parameters werethen evaluated following different methods, andaverage computed mass flow and power (MQSF andPQSF) were compared with the mean experimentalvalues (MNsF and PNs,,) through the infiuencefactors 1M and I,.. (1, 3, 6):

N

1: (e...P".)'-1

II

P.. QSF

N

1: Etii-1

MNSF

= MQElFPNBF

--"

PQBF

Il" =

Average mass flow and power were firstlycalculated as arithmetic mean values aver thepulse period, following the traditional methodreported by Benson (3), through the relations:

N

1: Mt.~-1

N

A different approach (proposed by Shamsi ina previous paper (10» is founded on thedetermination of a suitable mean value of theexpansion ratio which allows to evaluate turbinemean unsteady performance tram steady flowcharacteristics.

To this purpose a "mean equivalent expansionratio" (E" 8q) was introduced. In arder torelate this quantity to the pulsecharacteristics and to give a weight to eachinterval aver the period, the instantaneous massflow level (M".) was considered. The "meanequivalent expansion ratio" was then calculatedas a weighted average value:

IM" QBF =

andN

I: p",1-1 N

E (M".oE".)

IP" QBF = "-

Ne" ."

The results were substantially similar tothose previously obtained on a small IHIturbocharger turbine (6); computed results werealways lower than mean measured parameters, withaverage differences lower than 10 per cent farmass flow and slightly above this level farpower. Nevertheless it was noted that thedifferences between mean measured pulsatingvalues and the relevant steady flow levels weregenerally of the same magnitude (fig.6).

N

I: M"

III IIIOverall ma 55 flow M~ QSF and power p~ QSF

were then determined directly by reading outturbine 5teady flow characteri5tic5.

In arder to obtain computed values closer toaverage experimental turbine unsteadyperformance, two different procedures farcalculating overall levels have been developed,the results of which were compared with thoseprovided by the above method.

The traditional procedure always suppliesaverage performance results lower than thecorresponding steady flow values. This is due tothe reduction of the slope of the steady flow

A first comparative application of the abovemethods was developed with reference to theexperimental results previously obtained on asmall IHI turbocharger turbine (6).

The latter method produced the best resultsin alI operating conditions, with differencestram measured values generally lower than 5 percent both far mass flow and power. Largerdeviations were found far the traditionalprocedure, while th~ pressure ratio weighted-average method gave halfway results.

In fig.? the estimated influenc~ factors

@ IMechE 1990 C405/017126

values and convenient corrective factors have tobe introduced. Nevertheless more work isrequired to improve this technique.

vs. pulse frequencyare represented far thethree techniques, at constant turbine rotationalspeed and mean expansion ratio.

The traditional and the weighted-averagemethod always provided mean computed valueslower than measured mass flow and power, whilehigher levels were often obtained by using the"mean equivalent expansion ratio".

6 CONCLUSIONS

However it should be considered that in theabove application the test rig configuration (6)didn't allow to analyse independently theinfluence of pulse amplitude and mean valuewhich were correlated.

The new arrangement of turbine feeding linemade it possible to point out this effect onquasi-steady flow approximation.

The new arrangement of the turbine supply lineof the DINE turbocharger test rig has beendescribed in the papero By this facility testson single and two-entry turbines are nowpossible, with steady and pulsating flowadmission, and controlled flow characteristicsfar each turbine entry.

Additionally, a new experimental injectiondevice has been presented, which alIowsforinvestigation of turbine performance curvesclose to zero mass flow conditions withoutremoving or modifying the turbochargercompressor wheel.

The steady flow characteristics of a radiaIflow turbine tram a Garrett turbocharger fittedwith a waste-gate valve integrated in theturbine housing were investigated far differentwaste-gate apenings.

The described procedures were then appliedin the case of the unsteady flow tests performedon the Garrett TB025 turbocharger withcontrolled levels of frequency, pulse amplitudeand mean value. A typical result of thisextensive study is reported in fig. 8; theeffect of pulse amplitude (fig.8a) and frequency(fig.8b) is shown, far constant values ofother operating parameters.

The resulting influence factors proved to bemore affected by pulse amplitude than byfrequency (or shape). The lowest deviations trammean experimental values were obtained by theweighted-average procedure. This technique andthe traditional method results were almostunaffected by pulse amplitude and produced meancomputed parameters generally lower than themeasured ones. The procedure based on the "meanequivalent expansion ratio" didn't give goodresults, since large differences (up to lO percent far mass flow and to 20 per cent farpower) , increasing with pulse amplitude, werefound.

Pulsating flow tests showed an evidentmodification of pressure diagrams measuredupstream and downstream of the turbine far somepulse frequencies, due to the wave action in theducts.

By the new pulse generator system, periodicpressure fluctuations of different amplitudewere obtained at constant mean value; theirshapes proved to be quite similar far the samefrequency.

The time dependent pressure signals at theturbine inlet were unaffected by the waste-gateopening, while the outlet diagrams showed a highfrequency noise, probably due to the interactionbetween the flow components through the impellerand the by-pass porto

The influence of pulse frequency on turbinemean pulsating performance was not clear, but atendency to smaller differences tram steady flowvalues was generally observed far higherfrequency levels.

Inlet pulse amplitude proved to affectturbine average unsteady parameters. Thereduction of mean measured ma 55 flow and powerat increa5ing amplitude might 5ugge5t a quasi-5teady behaviour of the turbine operating underpulsating flow condition5.

A more thorough investigation on theinfluence of the waste-gate opening on meanturbine unsteady performance seems desirable inthe case of large by-pass areas. The study willconcern particularly the flow characteristics inthe turbine exhaust mixing chamber, referring topossible wave actions through the waste-gateporto

The extensive quasi-steady flow analysisperformed on two automotive turbochargerturbines suggests some generaI conclusions aboutthe reliability of the different calculationprocedures.

The traditional method gave good results inboth applications. Deviations tram the averageexperimental level (the amount of which wasgeneraIly about lO per cent) were unaffected bypuise amplitude, mean value or wave shape. Thisfact might suggest a quasi-steady flow behaviourof small radiaI turbines operating underpulsating flow conditions (at least as regardsmean resuIts (4, 5». Shifts tram measuredturbine overall performance might be reduced bymeans of appropriate corrective factors.

Lower errors were produced by the weighted-average method, although the results tramreievant influence factors were slightlyinfluenced by pulse amplitude. A betterapproximation in alI operating conditions mightbe achieved by proper corrective factors or,alternatively, by introducing more suitableweighting quantities, which take into accountthe variation of instantaneous mass flow andpower in the pulse period (and moreover anyphase difference tram the expansion ratiodiagram (4, 5»). Further developments are hopedto be obtained in this direction.

The method founded on the "mean equivalentexpansion ratio" didn't give good results, sinceit proved heavily affected by pulse amplitudein many operating conditions. In arder tooverçome this drawback, different weighting

An extensive application of the quasi-steadyfIow approach to caicuiate turbine puisatingperformance has been presented. Differentprocedures of integration aver the period ofinstantaneous quasi-steady values wereconsidered and applied in the case of tests ontwo different turbocharger radiaI turbines.

.In addition to a traditionai method ofcaicuiating overaii turbine performance, twoother procedures were proposed so as better toapproximate mean experimentai resuIts. In a

127rAn!\/n17 @ IMechE1990

first case, weighted average values wereconsidered, while a different system is based onthe definition of an appropriate turbine "meanequivalent expansion ratio".

The two proposed methods gave very goodresults in a first application, but results tramthe latter method were influenced by the inletpulse amplitude in a subsequent utilization.

Developments of the weighted-averageprocedure are actually being studied, while morework seems to be required to improve the latter

technique.

(7) ACTON, O. and CAPOBIANCO, M. Performanceof a radiaI fIow turbocharger turbine.Proc. Eight Conference on Fluid Machinery,Budapest, 1987, p. 3.

(8) WINTERBONE, D.E., BENSON, R.S., MORTIMERA.G., KENYON, P. and STOTTER, A. Transientresponse of turbocharged Diesel engines.SAE paper 770122.

(9) CAPOBIANCO, M., CIPOLLA, G., GAMBAROTTA, A.Effect of iniet puisating pressurecharacteristics on turbine performanceof an automotive wastegated turbocharger.SAE Internationai Conference, Detroit, 1990

(10) SHAMSI, S.S. Estimating the influence ofpulsating fIow conditions on the per-formance of a turbine. SAE paper 790068.

Many interesting suggestions far futureinvestigations are provided by the resultsreported here. Some possible developments of theDINE research programme on exhaust turbochargersare: the comparison between different turbinecontrai systems (such as variable geometry andwaste-gate valve) with particular reference topulsating flow operation, the theoreticalanalysis of wave action in the pipes of theturbocharger circuit, tests on twin-entryturbines with partial and unequal admissionconditions and, moreover, the investigation onthe effect of particular devices (such asparticulate traps) fitted into the turbinefeeding line.

NOMENCLATURE

Notation

fkntpAK

ACKNOWLEDGMENTS

MNpTe'T

.6.

This work was developed jointly and partlyfinanced by the Centro Ricerche Fiat ofOrbassano, Italy. The authors would like tothank Garrett S.A. far their technical suppor tand assistance during the investigation.

frequencyspecific heat ratiorotational speedtimepressurewaste-gate flow arearatio between mean pulsating and steadyflow performance (at the same meanexpansion ratio and rotational speedJinfluence factormass flow ratenumber of intervals in the pulse periodpowertemperature, pulse periodexpansion ratiotorqueoscillation amplitude

REFERENCESSubscricts

34itMNSFPQSFTT

turbine inletturbine exiti-th interval in the pulse periodturbinereferred to mass flow ratemean unsteady flow valuesreferred to poweraverage quasi-steady flow valuesstagnation conditionreferred to torqueaverage values

Performance Darameters

n/./T", turbine speed factor

M"../Ts

P3turbine flow factor

P3

p..

~.,= turbine expansion ratio

(1) BENSON R.S. and SCRIMSHAW, K.H. AnexperimentaI investigation of non-steadyfIow in a radiaI gas turbine. Proc. Instn.Mech. Engrs., 1965-66, vol.180, p.74.

(2) WALLACE, F.J., ADGEY, J.M. and BLAIR, G.P.Performance of inward radiaI fIow turbinesunder non-steady fIow conditions. Proc.Instn. Mech. Engrs.,1969-70, voI.184, p.183

(3) BENSON, R.S. Nonsteady fIow in a turbo-charger nozzieless radiaI gas turbine. SAENationai Combined Farm, Construction &IndustriaI Machinery and PowerpiantMeeting, Mi Iwaukee, 1974, paper 740739.

(4) DALE, A. and WATSON, N. Vaneless diffuserturbocharger turbine performance. IMechEConference on Turbocharging and Turbo-chargers, London, 1986, paper C110/86,pp.65-76.

(5) DALE, A., WATSON, N. and COLE, A.C. ThedeveIopment of a turbocharger turbine testfaciIity. IMechE Seminar on ExperimentaiMethods in Engine Research and DeveIopment,London, 1988, pp.75-83.

(6) CAPOBIANCO, M., GAMBAROTTA, A. and CIPOLLAG. Influence of the puisating fIowoperation on the turbine characteristics ofa smaii internaI combustion engineturbocharger. IMechE Conference on TheSmaii InternaI Combustion Engine, London,1989, paper C372/019, pp.63-69.

~ - Performance parameters were defined

referring to total/static conditions in steadyflow and static/static conditions in pulsatingflowoperation.

@ IMechE1990 C405/017128

Arrangement of the turbine feeding line on the turbocharger test rigFig 1

Fig3 Installation of wastegate equivalent nozzles on theturbine housing

Fig 2 Impulsive experimental device

129C405/017 @ IMechE 1990

3;6 3.6

3.2 3.2

2.8 2.8

2_4 2.4

2.0 2.0

1.6 1.6

1.2 1.2

0.8 0.8

0.4 0.4

0.0 0.01.0 4 1.8 2.2 2.6 3.0 1.0 4 1.8 2.2 2.6 3.0

Fig 4 Turbine steady flow characteristics

n".:JT;= 3500rp m/lK

f=50 Hz

P3=1~4bar A=O2.0 ~ flp3 =

-- ~~~~O~P4-

P3

1.5

1.0 ~

2.0 ~.t!p3 =0.50 barP3

1.5

1.0

~

-.~

2.0 ~/:,pJ =0.30 bar

I f=110Hz 1-I

P31.5 ~'---

P4ç~i

1 . O t:i:::.

0.0 0.5(a)

- L:J==l~~~~~ -- -1.00.0 0.5 1.00.0

(b)

0.5

(C)

1.0

Fig 5 Measured pressure diagrams upstream and downstream of the turbine(a) Influence of pulse frequency(b) I nfluence of pulse amplitude(c) Effect of wastegate opening

130 @ IMechE 1990 C405/017

06

03

00

0.9750 80

Cb)

1100.30 0.50

(o)

0.70

1.15

1.10 l*,i""

05 ~ ~~~ '*' "

, .[HZ]' '*

1.00

0.30 0.50

CC)

0.70 50 80

(d)

10

I nfluence of pulse frequency and amplitude on mean turbinepulsating flow performance

Fig 6

1.3

1.2

1.1

1.0

0.91.4

1.3

1.2

1 1

1.0

0.9

120 160o 40 80

Fig 7 Application of different quasi-steady flow methods to an I H Iturbocharger turbine

131C405/017 @ IMechE 1990

2

1.0

0.8

1.2

1.0

0.8

0.30 0.50

(O)0.70 50 80

Cb)

10

Fig8 Application of different quasi-steady flow methods to a Garrettturbocharger turbine(a) Influence of pulse amplitude(b) Frequency

@ IMechE 1990 C405/017132