Testing and qualification of two-stage turbocharging systems

Gerhard Fitzky, Mihajlo Bothien, Simon Zbinden, Ennio Codan, Stefan Vgeli ABB Turbo Systems Ltd., Baden, Switzerland

ABSTRACT The progressive lowering of emission limits for internal combustion engines aims, among other things, at a reduction in NOx formation. Miller valve timing is often chosen as the preferred option for reducing the NOx ratio by lowering combustion temperature. In order to avoid a drop in the mass and pressure of air in the cylinder due to Miller timing, higher charging pressure is needed. In the case of medium and large internal combustion engines, the necessary charging pressures can only be achieved economically by the use of two-stage turbocharging. In recent years ABB Turbo Systems Ltd. has developed products designed especially for two-stage turbocharging solutions. Rig testing of such complex systems plays an important role during the design and development phase. Requirements include the testing of thermodynamic characteristics and qualification of the two-stage system to ensure reliable operation on real engines. For the qualification tests new procedures were required that account of the implications of operation under higher pressures as well as the interconnection of two turbochargers. The new MNCH two-stage turbocharger test rig is described in the present paper. Some rig features are highlighted and the range of test possibilities is described. Measurement techniques used to record the thermodynamic behaviour of two turbochargers closely connected in series are presented in connection with what are, in part, new thermodynamic definitions. The paper addresses these issues and solutions are discussed. NOMENCLATURE p T Pressure Temperature Mass flow rate Isentropic enthalpy head Enthalpy Entropy Efficiency Pressure ratio Standard deviation Stress Mean value Total value Subscripts i o LP HP eq m Com Exp C T TS 2st a inlet outlet Low-pressure High pressure Equivalent Mean value Compression Expansion Compressor Turbine Turbocharging system two-stage Amplitude

 m hs

h s Superscripts (overbar) *

79

10.1243/17547164C0012010006 1. INTRODUCTION

1.1 General The turbocharged internal combustion engine is the most efficient single cycle machine for converting thermal energy into mechanical energy. Nevertheless, its development will require a further technological milestone with the planned tighter emissions limitations for large engines coming into force in the middle of this decade. The Miller process, involving the early closure of the inlet valve in order to allow the charge air to expand and cool and thus reduce pressure and temperature at the beginning of the compression stroke, is already applied in a moderate form on large 4-stroke engines. The advantage of the reduced compression temperature for diesel engines is a reduction in NOx emissions without any efficiency penalty and for gas engines a shift in the knock limit with corresponding power and efficiency improvement. Recent studies and experimental verification (1, 2) have shown that the application of an even stronger Miller process offers much greater potential. However, since every increase in the Miller effect requires a higher boost pressure and a higher turbocharging efficiency, the limits of single-stage turbocharging are being approached. Two-stage turbocharging provides a route to much higher pressure ratios and efficiencies. In connection with extreme Miller timing and other optimisations on the engine side, they open extended scope for improvements in terms of engine efficiency, power density and emissions. For these reasons it can be expected that two-stage turbocharging, which was recently introduced to the automotive market, will soon also become a standard feature on large engines. The first applications will probably be in constant speed engine applications such as large gas engines and medium-speed diesel engines in power plants. Applications at variable engine speed and load for marine propulsion systems or on locomotives are further candidates; first studies are underway for high-speed four stroke as well as low-speed two stroke engines. A prerequisite for the introduction of two-stage turbocharging systems is the availability of turbochargers developed and optimised for the specific application. The HP turbocharger, in particular, represents a new development, because the pressure ratio is moderate while absolute pressure and power density are extremely high. During its development it was soon obvious that it would be necessary to run it on a two-stage turbocharging test rig. 1.2 Requirements of the test rig Relying on similarity rules, it was possible to develop the two-stage system using conventional atmospheric test rigs for the individual stages. The influence of the absolute pressure on the HP-turbocharger could, to some extent, be measured on a pressurized test rig. This solution was tested at an early stage of the development process but the lack of flexibility, the higher costs for energy and the fact that it was impossible to study the interactions between the stages were serious limitations. The best way to measure the performance of a two-stage system is on a dedicated test rig. First engine applications are expected to be in the range of pressure ratios between 6.5 and 8, with the expected temperature range in the range of 500 to 600 C. With respect to future development, the design pressure ratio has been set at 12 with temperatures up to 650 C. A test rig for two-stage turbochargers can reproduce most steady-state operating points on the engine during normal operation using a combustion chamber. Since the pump action of the displacement machine is missing, additional energy sources are needed to run it only at the operating point with very low or negative pressure difference. The characteristics of the system, volume flow and efficiencies over pressure ratio can be measured in a realistic environment. The interaction between the 2 stages can only be studied qualitatively, because the connecting parts cannot be exactly the same as on the engine. Some of the control options that will be 80

10.1243/17547164C0012010006 available on the engine can be tested on the test rig. However, it is completely impossible to realistically simulate the transient phases on the engine due to the differences in behaviour between an engine and burner test rig. On the other hand, it is possible to measure system performance with higher accuracy than on the engine since the system layout can be adapted to the measuring equipment used. Additionally, the flexibility achieved with various control devices allows the system to be operated in almost the whole range of the component maps, which would not be possible on the engine. This aspect is very important for the qualification process, for which unusual operating conditions must also be run in order to check component behaviour under extreme conditions. In the following, the different aspects of design, control, measurement and evaluation will be discussed in more detail. Some special aspects of the qualification process are also highlighted. 2. CONFIGURATION OF THE MNCH TEST RIG

The development of two-stage turbocharging requires appropriate test rigs. As a result, it was necessary to design and construct a new test rig, named MNCH. 2.1 High flexibility Two-stage turbocharging makes increased demands on a test rig due to the higher pressures and more complex control and monitoring requirements. Due to these requirements, flexibility was an important issue during the design phase. The new test rig MNCH can be used for both single-stage and two-stage turbocharging. Furthermore, the flow rate can be varied over a wide range. Due to the wide range of power outputs, special actuators were installed in the combustion chamber. The different mounting configurations of the turbocharger meant that there was also a need for flexibility. Consequently, a new exhaust system was constructed. Instead of a fixed coupling with one mounting flange, a variable coupling with two flanges with a 90 offset was designed. One of these mounting flanges allows for infinite lateral adjustment. 2.2 Layout The layout of the test rig is shown in Figure 1. Figure 2 shows the assembly of the two turbochargers on the test rig. The test rig is usually operated in a closed cycle. Thus, the compressed air (of both the low- and high-pressure compressor) is used for combustion. The hot gas generated in the combustion chamber drives both the high- and low-pressure turbine. After passing through the turbines the gas is expelled through the exhaust system. The test rig allows for the turbochargers to be throttled collectively or for the lowpressure turbocharger to be throttled alone. This is achieved with a throttle installed downstream of each compressor. Some of the air after the high-pressure compressor can be bypassed directly to the exhaust system. To allow a wider operating range for the two-stage turbocharging, both the high-pressure and the low-pressure turbine can be bypassed. The high temperatures and high flow velocities on the turbine side make it necessary for the throttle and butterfly valve to be designed accordingly.

81

10.1243/17547164C0012010006

HP wastegate LP wastegate

High pressure turbocharger Intercooler

Low pressure turbocharger

Figure 1: Flow scheme of the two-stage test rig MNCH

HP-TC

LP-TC

Figure 2: View of the two-stage test rig MNCH 2.3 Supply system During operation the turbochargers are supplied by a lubrication system dimensioned to cover the turbochargers operational point. To avoid damage to the turbocharger in the event of the oil pump failing, an emergency oil system has been installed. Furthermore, the combustion chamber has to be supplied with different media, i.e. fuel, atomised air to disperse the fuel and propane gas to support ignition. The mass flow of the fuel is regulated by a bypass valve and a speedcontrolled pump.

82

Chimney

Blow out pipe

Auxillary blower Mass flow device

10.1243/17547164C0012010006 2.4 Control system The huge amount of data collected during operation of the turbocharger is reprocessed by the control system to provide the test engineer with an overview of relevant data, including information on the turbocharger itself, the lubrication system, the air system etc. If a safety critical operating point is reached the turbocharger is automatically shut down by the control system. However, the control system allows better reproducibility of tests. 2.5 Data acquisition Relevant thermodynamic data concerning the turbocharger, such as temperatures and pressures, are recorded by a separate system and analyzed online. Thus, the current operational point of the turbocharger can be calculated and visualized in its operating map. Further thermodynamic information on each turbocharger and on the complete supercharging system can also be determined and displayed, e.g. turbocharger efficiency. Together with the control system, data acquisition enables the test rig to be operated safely by only one test engineer. For cycle tests unattended operation is possible. 3. MODE OF TWO-STAGE TURBOCHARGER OPERATION ON TEST BED

The MNCH test rig allows qualification tests to be performed in addition to thermodynamic tests. In the case of qualification tests, it must be possible to run the test rig at chosen points on an operation line as well as near this line. Different control devices, like bypasses, were installed to achieve this aim. An example of the extension of the operation line is given in Figure 3.

Pressure ratio

opened LP-wastegate

closed LP-wastegate

Intake volume flow

Figure 3: Influence of a LP-wastegate on the HP-compressor operation line By installing a low-pressure wastegate (see Figure 1) the high-pressure turbocharger can be driven at higher rotational speeds and therefore higher pressure ratios. Figures 4 and 5 show the possibilities for shifting an operation point within the compressor characteristic in the two-stage operation mode for the high-pressure and the low-pressure turbocharger (Figures 4 and 5 respectively). Different control devices are required for the various movements, as depicted in Figure 1.

83

10.1243/17547164C0012010006 A combination of control devices is also possible. Opening the blow-out valves and increasing fuel pressure causes the operation point to shift in the direction of the choke limit.

Reference operation point: - opened throttle valves - closed blow-out valves - closed HP- wastegate Increasing cooling capacity of intercooler Increasing of fuel pressure or opening of LP-wastegate Throttling Opening of blow-out valves Reducing of fuel pressure Opening of HP-wastegate

Pressure ratio

Intake volume flow

Figure 4: Influence of different control devices on the position of the operation point for the high-pressure turbocharger

Reference operation point: - opened throttle valves - closed blow-out valves - closed LP- & HP- wastegate

Pressure ratio

Increasing of fuel pressure Throttling or opening of HP-wastegate Reducing of fuel pressure or closing of LP-wastegate Increasing cooling capacity of intercooler

Opening of blow-out valve Intake volume flow

Figure 5: Influence of different control devices on the position of the operation point for the low-pressure turbocharger

84

10.1243/17547164C0012010006 4. THERMODYNAMIC MEASUREMENT TECHNIQUE AND CALCULATION

4.1 Special thermodynamic definitions for two-stage turbocharging An important purpose of the two-stage test rig is the evaluation of the thermodynamic performance of two-stage turbochargers arranged in series. The low-pressure turbocharger can be tested in single-stage operation with adequate accuracy. However, the high-pressure turbocharger must be tested in two-stage operation due to high-pressure influences such as Reynolds-effects and leakage. Also, the interaction of the low-pressure and high-pressure turbochargers, i.e. the mutual influence due to flow effects, intercooling and the design of the connecting pipes, can only be evaluated in two-stage operation. It is therefore of interest to draw a thermodynamic balance of each single turbocharger and of the two serially connected turbochargers. Figure 6 shows the h,s-diagrams for the two-stage compressor and turbine process.

C o,LP

Ti,HP

Enthalpy h

C o,HP

Enthalpy h

Ti,LP To,HP

C i,HP C i,LP Entropy s Entropy s To,LP

Figure 6: h,s-Diagrams with isobars for two-stage turbocharging with intercooling between the compressors; left: compressors; right: turbines The following conventional definition of turbocharger efficiency is commonly used for single-stage evaluation, where the enthalpy change h is a function of temperature and pressure ratio:
* * pCo  Co hs ,Com m TCi , p * Ci = * * pTi  Ti hs , Exp m TTi , p To

TC

(1)

For this purpose, the values of the total inlet and the static outlet pressures and the total inlet temperature of an ideally uniform flow at the system boundaries in the transverse planes at the turbocharger flanges must be known. The energy balance of two-stage charging systems is calculated in accordance with single-stage charging by viewing the whole system as a black box. The efficiency of such balancing is called Equivalent Turbocharging Efficiency:
* * pCo , HP  Co , HP hs ,Com TCi , m , LP p * Ci , LP = p*  Ti , HP hs , Exp TTi*, HP , Ti , HP m pTo , LP

TC ,eq

(2)

85

10.1243/17547164C0012010006

This definition of turbocharging efficiency which is relevant for the engine can be directly compared with the single-stage turbocharger efficiency. The equivalent efficiency includes the influences of intercooling, of the connecting pipes between the turbochargers and other flow effects carried from one turbocharger to the other, e.g. between the two turbines. Hence, this efficiency definition involves a strong interdependence of the interconnection pipe configuration and of the intercooling temperature level. The following definition of the Mean Turbocharger Efficiency characterizes the thermodynamic performance of the high-and low-pressure turbocharger only, i.e. without any other system components:

TC ,m

* p * , HP  Co , HP hs ,Com TCi m , Co , HP * pCi , HP = * * pTi  Ti , HP hs , Exp TTi , HP , * , HP m pTo , HP

* p * , LP +m  Co , LP hs ,Com TCi , Co , LP * pCi , LP * p +m  Ti , LP hs , Exp TTi*, LP , Ti , LP pTo , LP

(3)

The ratio of the equivalent and the mean turbocharger efficiencies can be referred to as the Two-stage Turbocharging System Efficiency. It includes only the influence of the intercooler and of the flow losses of the connecting pipes. Due to the effects of intercooling, this efficiency can reach values greater than 1.

TS , 2 st =

TC ,eq TC ,m

(4)

Table 1 shows, as an example, the measurement results of the same turbocharging system on the test rig and on the engine at comparable operating points. It should be noted that the geometry of the interconnecting pipes is not identical. Differences in the temperature level of the intercooler are negligible. Test rig Research engine 0.675 0.012 1.096 0.740 0.007

TC ,m TS , 2 st TC ,eq

0.690 0.005 1.049 0.724 0.007

Table 1: Comparison of different efficiencies between test rig and engine measurements In principle, the mean turbocharger efficiency should be the same on both a test rig and a research engine. However, the effects of aerodynamic interaction between the turbochargers and different measurement accuracies cause differences. The turbocharging system efficiency is different on a test rig and a research engine due to the different geometries of the connecting pipes and the intercooler. The differences in the equivalent efficiencies are the summation of all described effects. In spite of these differences it is advantageous to conduct thermodynamic measurements on a special test rig due to the different achievable accuracies of the thermodynamic measurements on the test rig and on the real engine, see Table 1. On the engine, the design of the charging system is very compact. Hence, a representative balancing of the turbochargers at their inlets and outlets will be insufficient. The reason is that accurate thermodynamic measurements of two stage turbocharging systems require uniform distributions of pressure and temperature in 86

10.1243/17547164C0012010006 the transversal flow planes on the one hand and attention to the aerodynamic interaction between the two turbochargers due to the compact design of the system on the other. Homogenization of the flow can only be achieved when there is a sufficient axial distance between the turbocharger flange and the measurement plane. Such long axial lengths are not normally found on the engine due to the need for compact design. A test rig can, in principle, be designed without having to consider the criteria of compactness. But with respect to the aerodynamic interaction between the turbochargers, excessively long connecting pipes are not permitted. In particular, aerodynamic interaction is considerable between the two turbines. At this location a more elaborate measurement technique is necessary to obtain representative values. The application of such sophisticated measurement techniques is only feasible on a test rig. 4.2 Thermodynamic measurement technique in non-uniform flow Accurate thermodynamic characteristics need representative mean values to determine the inlet and discharge gas state. These are achieved either by means of measurement traverse or when there are sufficient axial lengths to homogenize the pipe flow velocity and pressure distribution. Longer axial distances between turbocharger flanges and measurement planes are therefore common, as they avoid extensive measurement techniques. Typical distances between the turbocharger flange and measurement plane are between 0.5 and 1.5 times the pipe diameter. Beyond these distances homogenization is adequate and within these distances the flow loss and the heat loss are still negligible. An exception to this is the flow vortex, which needs a considerably longer path length to decay. The residual non homogeneity in the measurement plane is taken into account via multiple measurements of pressure and temperature at different locations uniformly distributed over the plane. The average of these single measurements represents the mean value which will be achieved with uniform flow. Test rig pressure and temperature are each measured at 8 different locations in the plane. Static pressure is determined at the pipe wall by using pressure taps. Rod shaped thermocouples, immersed in the flow at different depths, determine the temperature. Strong flow vortices and flow after changes in the flow direction cause greater deviations between the representative mean plane pressure and wall pressure values. Flow vortices occur mainly at the outlet of the volute of the radial compressor and of the mixed flow turbine. Flow analyses have shown that vortices and secondary flow effects at the compressor outlet are sufficiently decayed after a path length of a half pipe diameter. Non homogeneities at the turbine outlet are much more persistent. A flow straightener is thus often installed between the outlet flange and the measurement plane by extending the axial distance between the flange and the plane to 3 or 4 times the pipe diameter. A two-stage arrangement of two turbochargers on the test rig allows only short pipe lengths between the high-pressure turbine outlet and the low-pressure turbine inlet in order to achieve the desired aerodynamic interaction between the two turbines on test rig. Also, the measurement devices in this pipe, e.g. probes and thermocouples, must not disturb the flow field. Furthermore, wall pressure measurement is not capable of detecting the strong pressure profile in the turbine outlet. To solve these problems it was decided to use the second method of measuring a representative mean gas state value in addition to the wall pressure measurement: i.e. the measurement traverse method. For the total pressure measurement rakes, each equipped with 5 miniaturised Kiel probes, were installed at various equidistant radial locations. The outstanding advantage of Kiel probes compared to other total pressure probes is their complete insensitivity to the direction of flow within 50 degrees of pitch and yaw angle. This angular range covers all downstream flow angles occurring at all the different turbine operating loads.

87

10.1243/17547164C0012010006 Commonly, mean total pressure is calculated from the static wall pressure and the mean velocity head in the plane. If the special pressure rakes had not been available only this value would be used for calculating turbine and turbocharger efficiencies. In the whole operating range of the high-pressure turbine, the mean total pressure value detected with the total pressure rake is higher than the value calculated using wall pressure and the velocity head. The left-hand diagram in Figure 7 shows the relative difference in total pressure measured with the Kiel probes and with the pressure taps at the wall. The total pressure value measured with the Kiel probe rakes must be higher than the calculated total pressure since the mean rake value contains the absolute velocity component. The calculated value only comprises the axial flow component due to the velocity head. It is evident that such differences in measuring results lead to great differences in efficiencies. The right-hand diagram shows the difference between turbocharger efficiency calculated with, on the one hand, the total turbine outlet pressure measured with the Kiel probes and, on the other that determined with the wall pressure and the velocity head. Due to the higher pressure value obtained with the probes, the efficiency is between 0.8 and 1.6 percent higher than indicated by the wall pressure value. The higher sensitivity of efficiency to turbine outlet pressure at part load leads to greater efficiency differences than at full load.

((p*To,Probe / p*To,Wall) - 1)

0.015 ( TC,Probe - TC,Wall)

0.02

0.015

0.01

0.01

0.005

0.005

0 1 2 T* 3

0 1 2 T* 3

Figure 7: Comparison between mean total pressure calculated from the static wall pressure and velocity head (wall) and inside the flow field (probe) at the high pressure turbine outlet (left) and the impact on turbocharger efficiency (right) For discussion: In the diagrams shown above, the total pressure measured with the rakes at the high pressure turbine outlet takes into account the absolute velocity component. The circumferential component of the swirl flow downstream of the high pressure turbine cannot, as a rule, be used in the downstream low-pressure turbine. Thus, only the axial component is relevant for thermodynamic balancing of the high-pressure turbine. The circumferential component dissipates and cannot be utilized for the thermodynamic process in the low-pressure turbine. In general, it is the energy potential of the axial flow component which can be used in the lowpressure turbine. The total axial pressure can only be measured accurately using an elaborate measurement technique such as pneumatic multi-hole-probes. In view of the relatively small amount of useful knowledge gained, normal test rig operation does not allow the use of such time and cost intensive techniques. As a result, both measurement methods are used in parallel to determine high-pressure turbine efficiency - i.e. static wall pressure plus velocity head and absolute total pressure using a Kiel probe. 88

10.1243/17547164C0012010006

4.3 Measurement uncertainties The significance of a thermodynamic measurement is indicated by its estimated accuracy. Accuracy is influenced by the uncertainties of the measurement sensors, data processing, calibration and local averaging of the individual measured values. Internal studies have shown that a linear model of error propagation delivers results with sufficient accuracy. The several variables can be assumed to be sufficiently uncorrelated. Hence, the model of the Gaussian error propagation can be applied in its simplest form. For the standard deviation f of a characteristic f with n influence parameters xi the following general definition is used.

2 f =

f xi i =1 xi
n

(5)

The above mentioned individual influences result in the uncertainty of the efficiencies. The deviation due to averaging of the four or eight individual measured values in each measuring plane has a considerable influence on the summarized uncertainty of each efficiency. The more non homogeneous the flow characteristics in the measuring planes, the greater these averaging deviations. The recommendation VDI 2640 (4) proposes the following approach to quantify the deviation: half the maximum deviation occurring between the measured values of adjacent sensors in a plane is the averaging uncertainty of this variable in the plane. The diagrams in Figure 8 show the absolute uncertainty of the stated efficiencies as a function of the engine load. They show the extended uncertainty (2) with a confidence level of 95 percent. The uncertainties in the measurements performed on the engine are significantly higher than those on the test rig, due mainly to the larger averaging deviations. Also, the erratic behaviour of the efficiencies over the load spectrum when measured on the engine indicates the large deviations between the individual measured values in the measuring planes. Figure 9 shows the repeat uncertainties at 95% confidence level. Only random uncertainties of the data acquisition system are considered. Due to the similarity of the measurement equipment, these uncertainties are also quite similar. Summary: A two-stage test rig specifically for thermodynamic measurements allow both the turbocharger efficiency of each turbocharger and the different mean efficiencies of the system to be determined much more accurately than on the real engine. The equivalent efficiency depends to a considerable degree on the losses in the connecting pipes between the compressors and the turbines and on the level of the intercooling temperature drop. Thus, it is important to measure this efficiency on the engine unless the losses and intercooling temperature level are exactly reproducible on the test rig. However, when the mean turbocharger efficiency TC,m and the engine specific turbocharging system efficiency TS are known, the equivalent turbocharger efficiency TC,eq can be reliably determined.

89

10.1243/17547164C0012010006

Absolute uncertainty

Absolute uncertainty

0.05 0.04 0.03 0.02 0.01 0 0

Engine

0.05 0.04 0.03 0.02 0.01 0 0

Test rig

50 Load in percent

100

50 Load in percent

100

Figure 8: Absolute uncertainty of measured efficiencies on the engine (left) and on the test rig (right) at 95% confidence level

0.01 Repeat uncertainty

0.005

Repeat uncertainty

Engine

0.01

Test rig

0.005

0 0 50 Load in percent 100

0 0 50 Load in percent 100

Figure 9: Repeat uncertainty of measured efficiencies on the engine (left) and on the test rig (right) at 95% confidence level 4.4 First measurement results To obtain reliable information on thermodynamic behaviour, extensive tests were conducted on the MNCH test rig. Figure 10 shows the compressor map of the low-pressure and high-pressure compressors. An operation line is chosen which agrees with that recorded on the engine with early Miller Timing. The comparison of the marked operation points in the two maps indicates the different behaviour of the low-pressure and high pressure compressors. The lowpressure compressor continuously runs up to higher pressure ratios. The high pressure compressor reaches a certain level of pressure ratio and remains there. This should not distract attention from the fact that the energy conversion of the high pressure turbocharger still increases because of increasing compressor inlet pressure.

90

10.1243/17547164C0012010006

Total pressure ratio

Intake volume flow rate

Total pressure ratio Intake volume flow rate

Figure 10: Measuring points in the compressor maps of the low-pressure (left) and of the high pressure (right) turbocharger Figure 11 shows the mean TC,m and the equivalent TC,eq turbocharger efficiency and the test-rig-specific two-stage turbocharging system efficiency TS,2st. At part load the equivalent efficiency is lower than the mean efficiency due to the pressure losses in the interconnecting pipes between the two turbochargers. At full load the equivalent efficiency is higher because of the intercooling. Due to the intercooling the system efficiency increases above a value of one.

1.1 1 Efficiency 0.9 0.8 0.7 0.6 0.5 1 2 3 4 5 6 7 8

pCo,HP/pCi,LP
Figure 11: Thermodynamic test results of two-stage turbocharging obtained on the MNCH test rig

91

10.1243/17547164C0012010006 5. Qualification procedure and tests

5.1 Introduction The qualification system introduced by ABB Turbo Systems (for single-stage turbocharging) requires that each newly developed turbocharger has to pass a certain number of qualification tests before being released. Qualification tests are also applied in the new development of the high- and lowpressure turbochargers used in the two-stage turbocharging system. Furthermore, the interaction between both turbochargers plays an important role in the qualification process. The single-stage qualification procedure comprises up to 25 tests. The relevance of these tests to two-stage turbocharging must be validated. Some test procedures can be adapted while others need to be developed from scratch. Basic tests were performed prior to the development of the new test procedures. These basic tests were carried out both on a single-stage test rig which had been specially upgraded for higher inlet pressures and on the MNCH test rig. One aim of the qualification tests is to operate the turbocharger at conditions which, preferably, exceed those of the future application, i.e. at higher rotational speeds, pressures and temperatures. To achieve this goal, special control devices were installed on the MNCH two-stage development and qualification test rig at ABB Turbo Systems. A low-pressure wastegate allows the setting of certain operational points in the high-pressure compressor characteristic, points which it was impossible to approach without this control device - see the LP wastegate in Figure 1. It has yet to be decided whether each qualification test is to be performed in twostage operating mode (i.e. pressure at HP-compressor inlet higher than atmospheric pressure) or whether in the single-stage operating mode (i.e. atmospheric pressure at HP-compressor inlet). In some cases the single-stage operating mode delivers the same information or turns out to be an even stronger condition. A further link in the qualification test chain for a two-stage turbocharging system is on-engine tests at the engine builder. From these tests crucial information can be gained regarding the elements of a qualification test procedure on a turbocharger test rig. ABB Turbo Systems has always made wide use of these tests. The compilation of a whole qualification procedure for the two-stage system is still in progress. In the following section, some of the qualification tests already performed are detailed. 5.2 Blade vibration qualification One important qualification test is the blade vibration test, which is carried out on both the turbine and the compressor blading. 5.2.1 Turbine In a first step critical resonances are detected. In a second step an endurance test is performed to check these resonances with respect to the fatigue strength of the blades. This qualification test posed, among other things, the question of whether the resonance amplitudes in two-stage operation are higher than in single-stage operating mode. Figure 12 shows a comparison of the stress amplitudes due to vibration of the turbine blading in single-stage (atmospheric) operation (-symbol) and in twostage operation (x-symbol) of the HP-turbocharger. No general increase in amplitudes could be observed in two-stage operation. In single-stage and in two-stage operation amplitudes of the same level occur. By contrast, in single-stage operation the amplitudes can be higher or lower. On the one hand, fluid forces excite blade vibration. These forces are proportional to the level of supercharging. On the other hand aerodynamic damping attenuates blade vibration at the resonant frequency. Aerodynamic damping rises with the 92

10.1243/17547164C0012010006 density of the fluid (and hence with the level of supercharging). Which of these inversely proportional influences dominate in the two-stage operation mode certainly depends on excitability in the natural frequency range concerned.
1 Stress amplitude a/a,T,ref 0

Mean stress m/m,T,ref

Figure 12: Comparison of the load due to vibration of the turbine blading of the high pressure turbocharger. : atmospheric operation, : two-stage operation 5.2.2 Compressor During the stated preliminary tests described with the high pressure turbocharger, blade vibrations were measured at different compressor inlet pressures: atmospherically and at representative two-stage inlet pressure. Figure 13 shows the influence of the compressor inlet pressure of the high pressure compressor. The different symbols in each diagram indicate specific ranges of turbocharger speeds. As can be seen from Figure 13, the load profile of the compressor blades at different inlet pressures and for different resonances shows no clear trend.

1 Stress amplitude a/ a,C,ref

Atmospherically

1 Stress amplitude a / a,C,ref

pCi = 2 bar; TCi = 60C

Mean stress m/ m,C,ref

Mean stress m/ m,C,ref

Figure 13: HP-compressor blade vibration amplitudes, left: atmospheric test; right: supercharged test 5.3 Containment test Containment on the turbine side as well as on the compressor side is also a vital part of the qualification procedure. Accompanying burst simulations show that the pressure in the fluid channel has no decisive influence on the burst pattern. The forces (impact) interacting between the rotor parts and the casing parts dominate during the burst. For this reason the burst trials were performed atmospherically. 93

10.1243/17547164C0012010006

Figure 14: High pressure turbocharger after turbine side burst test. Figure 14 shows the result of the turbine-side containment test. The turbocharger was accelerated to its natural burst rotational speed. According to the specification sheet, no part escaped the metal casing. 6. CONCLUSION

Two-stage turbocharging provides an important contribution to the reduction of emissions in internal combustion engines. With the construction of the MNCH twostage turbocharger test rig, the foundation has been laid for the experimental development and qualification of the two-stage turbocharger and the two-stage turbocharging system. Important design targets of the test rig were: 1. Realistic engine-like operation of the turbocharging system, 2. High accuracy of thermodynamic measurements and 3. Flexibility and availability for basic tests. Authors 2010 7. References

1 Codan, E. & Ch. Mathey, 2007, Emissions a new challenge for turbocharging, 25th CIMAC World Congress in Vienna, Austria 2 Codan, E., S. Vgeli & Ch. Mathey, 2008, Hochdruckaufladung bei Gasmotoren, 13. Aufladetechnische Konferenz, Dresden (D). 3 CIMAC Recommendation No. 27, Turbocharging efficiencies Definitions and guidelines for measurement and calculations, May 2007 4 VDI/VDE 2640 Part 4 Guideline, Measurement of mean temperature in fluid flow, Oct 1983

94