2003-01-1068

The Impact of Exhaust Gas Recirculation on Performance and Emissions of a Heavy-Duty Diesel Engine Timothy Jacobs, Dennis Assanis, and Zoran Filipi Automotive Research Center The University of Michigan

Copyright © 2003 Society of Automotive Engineers, Inc.

ABSTRACT This work studies the complex interactions resulting from the application and control of Exhaust Gas Recirculation (EGR) on a production heavy-duty diesel engine system, and its effectiveness in reducing NOx emissions. The coupling between EGR, the Variable Geometry Turbocharger (VGT) and the EGR cooler critically affects boost pressure, air/fuel ratio (A/F), combustion efficiency and pumping work. It is shown that EGR provides an effective means for reducing flame temperatures and NOx emissions, particularly under low A/F ratio conditions. However, engine thermal efficiency tends to decrease with EGR as a result of decreasing indicated work and increasing pumping work. Combustion deterioration is predominant at higher load, low speed and low boost conditions, due to a significant decrease of A/F ratio with increasing EGR. For conditions allowing the VGT to maintain high enough boost and hence A/F ratio, efficiency losses with increased EGR are largely attributed to increased pumping work. Finally, the total system heat rejection increases significantly due to EGR cooling.

INTRODUCTION

EPA started imposing air emission regulations on heavyduty engines in 1985, to take effect in 1991, and then more stringent regulations in 1994. However, these initial regulations could be met with optimized combustion strategies, and improved combustion chamber design. EGR became a necessary component on heavy-duty diesel engines with the implementation of the 2004 regulations (accelerated to 2002 for six major manufacturers affected by the Consent Decree) where NOx release is restricted to 2.5 g/bhp-h. Nevertheless, introducing EGR effectively into the combustion chamber of a multi-cylinder engine remains a considerable challenge. External EGR, using piping to route the exhaust gas to the intake system where it is inducted into the succeeding cycles, has emerged as the preferred current approach. However the high efficiency of a state-of-theart turbocharger often establishes conditions where the intake manifold pressure is higher than the exhaust manifold pressure. Consequently, an auxiliary device, such as the Variable Geometry Turbine (VGT) is needed to increase the backpressure above the intake manifold pressure and allow flow in the proper direction. The work presented here is motivated by the need to better understand the issues associated with practically coupling an EGR system with a VGT on a heavy-duty diesel engine.

Exhaust Gas Recirculation (EGR), has long been of interest to engine designers, researchers, and regulators. EGR was originally considered as a method While EGR’s ability to alter combustion and reduce to alter combustion [1]* and suppress knock in spark diesel engine emissions has been previously ignition engines [2]. Considerable interest in EGR for demonstrated [e.g., 6-8], the experimental research was gasoline engines developed shortly after 1955 when conducted with highly controlled introduction of typical Haagen-Smit [3] successfully demonstrated the combustion products (not actual exhaust) into the intake dependency of smog on combustion-generated stream. Therefore, other engine parameters, such as hydrocarbons and oxides of nitrogen. Five years later, air-flow, boost pressure, and exhaust pressure could be Kopa and other researchers, demonstrated that EGR held constant during the studies. In current practical could in fact lower the concentration of NOx in the applications of EGR on production engines, these exhaust gas [4,5]. Recently, EGR has emerged as a variables are not constant with changing levels of EGR. necessary means to meet the United States In particular, the issues related to applying EGR in a Environmental Protection Agency (EPA) nitric oxide multi-cylinder engine rather than a laboratory single(NOx) regulations for heavy-duty diesel engines. *Numbers in brackets indicate references included at the end of the paper.

cylinder, are much more complex, due to the coupling between events in the cylinder and the performance of the turbocharging system. This work examines the interactions resulting from the application and control of EGR on a practical heavy-duty diesel engine system, with the aim of understanding their impact on reduction of NOx emissions, as well as reasons for potential negative side-effects. Since both (EGR and VGT) impact the “intake charge side” of the engine considerably, altering one or the other (or both) will have dramatic and sometimes surprising effects on engine performance and combustion. In particular, EGR’s effect on combustion can differ, depending on other system issues, such as the ability of the turbocharging system to preserve the desired A/F ratio, or the effectiveness of EGR cooling. There might be a fuel economy penalty associated with deteriorated combustion. The pumping work associated with driving the EGR flow might be as significant or more dominant under certain conditions. The goal of this experimental study is to distinguish between in-cylinder processes and system effects related to EGR, and to assess their respective contributions to potential performance and efficiency penalties. The nitric oxides trends will be examined carefully to determine the primary factor/factors responsible for their reduction, such as thermal, dilution and dissociation effects. Furthermore, videoscope images obtained in the combustion chamber will be used to assess the effects of decreased flame temperatures with increasing residual. Other emissions, such as CO and unburned hydrocarbons (HC), will also be studied in the light of VGT’s ability (or inability) to maintain a constant air / fuel ratio. Finally, practically utilizing EGR in a production engine has implications on the engine/vehicle heat rejection. Cooled EGR, versus non-cooled EGR, improves fuel economy and performance, but considerably increases engine system heat rejection. Distinguishing between the system heat rejection and combustion chamber heat rejection allows better interpretation of combustion trends, as well as better assessment of possible durability implications. To this end, both system energy balance and detailed wall surface temperature measurements will be analyzed. This paper begins with a discussion of current methods to drive EGR using VGTs, and describes the method studied in this work. The subsequent section provides details about experimental methodology and engine instrumentation. This is followed by a thorough experimental analysis of the performance and fuel economy implications associated with practically implementing EGR with VGT. Next, the emissions trends are examined and the primary factors responsible for reduction of nitric oxides are identified based on a combination of traditional measurements and in-cylinder

temperature contours extracted from flame images. Finally, the paper addresses issues of total system heat rejection with EGR and offers conclusions of the experimental study.

METHODS FOR RECIRCULATING EXHAUST GAS IN A TURBOCHARGED CI ENGINE One of the more challenging aspects of EGR is introducing it into the combustion chamber of a multicylinder engine. There are two “types” of EGR; internal and external. Internal EGR uses variable valve timings or other devices to retain a certain fraction of exhaust from a preceding cycle. External EGR uses piping to route the exhaust gas to the intake system, where it is inducted into the succeeding cycles. While internal EGR provides very short response time, its practical application is not possible until camless technology becomes widely available. Furthermore, internal EGR cannot be cooled, whereas external EGR can. Cooling the residual improves fuel economy and engine performance and enables further reduction of NOx. Therefore, external EGR has emerged as the preferred type of EGR for heavy-duty diesel engines and is used in this study. However, application of external EGR opens up several system issues that need to be addressed. In particular, the high efficiency of state-of-the-art turbochargers often establishes conditions where the intake manifold pressure is higher than the exhaust manifold pressure. If residual is to be flown from the exhaust manifold directly to the intake manifold (the so called short-path EGR), the exhaust backpressure must be higher than the intake. Hence, an auxiliary device is needed to increase the backpressure above the intake manifold pressure. Figure 1 illustrates an engine system configured for short-path external EGR through the use of a VGT auxiliary device, commonly used in North America. The reader is referred to Arnold, et al. [9] for more information on VGTs. In general, the VGT is capable of changing its geometry to obtain different turbine flow areas, and thus different boost pressures. Vanes in the nozzle upstream of the turbine open (or close) enhancing (or restricting) the flow. Originally conceived to reduce turbocharger lag in transients, the VGT also has the ability to assist or enable the external EGR flow, since changing its flow characteristics will affect the pressure ratio between the exhaust and intake sides. Other methods of flowing external EGR include piping the EGR after the turbine into the fresh air intake upstream of the compressor (long-path EGR). This method would prevent excessive back pressure of the exhaust to flow EGR. However, although EGR fundamentally contains CO2 and H2O, in reality its composition includes soot, as well as SO2, SO3, and at certain temperatures and pressures, sulfuric acid. The aluminum components of the compressor and

intercooler would probably not sustain long life if exposed to contamination with corrosive compounds and particles. Other variations involve extracting the exhaust downstream of the VGT and using an EGR pump. While these alterations appear attractive from a performance standpoint, complications regarding control, durability, and packaging are discouraging wider application.

In te rc o o le r

E n g in e

P am b

T

Engine tests were conducted at 1200 RPM, mid load (to simulate a moderate acceleration in urban driving conditions), 1200 RPM, low load (to simulate low speed steady urban driving), and 1800 RPM, mid load (to simulate high speed steady driving). Each operating point was tested as a steady-state condition. The EGR valve acted solely as an on-off valve. At each condition, the VGT vane position was altered to adjust the pressure difference across the engine required for obtaining the desired EGR rate (as measured with an orifice in the EGR pipe). As a result, the boost pressure altered as vane position changed. The fuel rate and the injection timing remained unchanged. However, A/F changed with changes of both the air flow and EGR flow. As exhaust pressure, boost pressure, and A/F changed, the torque level changed.

C

F re s h A ir EGR C o o le r

Figure 1: Layout of the short-path EGR System with a VGT and an EGR Cooler.

Figure 1, illustrating the system studied in this work, helps identify the control issues associated with using the VGT to drive EGR. The engine’s Electronic Control Module (ECM) can effectively respond to a single feedback signal (e.g. residual flow rate); however, two goals must be met related to controlling the EGR rate. The primary goal is to flow enough EGR to adequately reduce NOx emissions. The secondary goal is to maintain high enough boost required for the engine’s efficient operation, while preserving desired EGR flow. Therefore, a considerable amount of knowledge must be built regarding the total system operation and control to properly use a VGT to drive EGR. One particular publication by Pfeifer, et al. [10] attempts to develop a simulation-based model to predict a suitable control algorithm for optimizing boost pressure and EGR rate. The experimental research presented in the following sections offers insight into the system issues and consequences of complex interactions between intake and exhaust components.

Several variables were measured on a time basis, in particular: speed, torque, intake manifold pressure, exhaust manifold pressure, air flow, fuel flow, coolant flow, intake manifold temperature (controlled by air to water intercooler), exhaust manifold pressure, engine coolant temperature difference, exhaust CO/CO2 (NDIR analyzer), exhaust O2 (paramagnetic analyzer), NO and NOx (chemiluminescence analyzer), and gaseous hydrocarbons (hot FID). Signals measured on a quarterdegree crank-angle basis were: cylinder pressure (cylinders 1, 3, and 6), fire-deck surface temperature and corresponding reference temperature (heat flux probe in cylinder 3), and strain on the cylinder 6 unit injector rocker arm (for post-processing fuel injection pressure and timing). EGR rate was measured and provided as output by the ECM. The video images were collected at different engine operating conditions from cylinder 6 with an AVL 513D engine videoscope. The images show an 80 degree view field, looking straight into the combustion chamber (through the cylinder head) near the outer bowl region of the piston. Figure 2 illustrates the viewing angle of the videoscope more clearly.

Back Port

Front Port Injector

EXPERIMENTAL METHODOLOGY The experimental data were acquired at the University of Michigan’s Automotive Research Center on a heavy-duty Detroit Diesel Series 60 CI engine rated for 1580 ft-lbs torque at 1200 RPM, and 440 hp at 1500 RPM (constant to 1800 RPM). The engine has six in-line cylinders, and a total displacement of 12.7 L. It was retrofitted with a VGT, turbine-mounted EGR valve, EGR cooler, EGR mixer, and upgraded ECM capable of controlling both the VGT and EGR flow rate.

Piston at TDC Cylinder

Figure 2: Schematic illustrating the view angle of the videoscope probe installed in cylinder 6 of the heavy-duty Diesel engine.

The temperature profiles have been processed by the AVL 513D Thermovision software. The flame images were recorded using the two-color pyrometry method, with a CCD camera. The Thermovision software analyzes the wavelengths of the two colors, and based on the radiation intensity of these wavelengths, determines a corresponding flame temperature.

nearly constant, indicating little degradation of combustion. However, for both of the latter cases, Net IMEP steadily decreases with increased EGR, indicating an increase in pumping losses. 0.41

1200, 50% 1800, 50% 1200, 20%

EFFECTS OF EGR ON FUEL ECONOMY AND PERFORMANCE Previous studies [11,12] have shown that using EGR in order to reduce NOx emissions generally degrades engine performance. This section quantifies the effect of EGR on the fuel economy and performance of a representative heavy-duty diesel engine through experimental measurements. All results pertain to three reference conditions: 1200 RPM - 50% load, 1200 RPM 20% load and 1800 RPM - 50% load. Results are analyzed to determine the specific physical reasons for degraded performance.

Thermal Efficiency (%)

0.4 0.39 0.38 0.37 0.36 0.35 0

5

10

15 EGR Rate (%)

20

25

Figure 3: Brake Thermal Efficiency versus EGR Rate for three operating conditions.

11 10

Engine System Performance and Distribution of Efficiency Losses IMEP (bar)

Figure 3 illustrates a decrease of engine thermal efficiency as a function of the amount of EGR for all three cases investigated.

9 8 7

Gross IMEP Net IMEP

6

1200, 50% 1800, 50% 1200, 20%

5 4

Indicated mean effective pressure (IMEP) trends, shown in Figure 4, allow better understanding of the relative contribution of each of the factors (decreased combustion work versus increased pumping work). The Gross IMEP and the Net IMEP shown in Figure 4 are obtained by processing cylinder pressure data. For the 1200 RPM - mid load case (highest IMEP level), both Gross IMEP and Net IMEP decrease with increasing EGR. The decreasing Gross IMEP indicates a decrease in combustion work. For the 1800 RPM case, as well as the 1200 RPM - low load case, Gross IMEP remains

3 0

5

10

15

20

25

EGR Rate (%)

Figure 4: Indicated Mean Effective Pressures versus EGR Rate for three operating conditions (both gross and net are shown).

To further quantify the effect of pumping work, Figure 5 shows the Pumping Mean Effective Pressure (PMEP) plotted versus EGR Rate. Note that negative PMEP refers to the piston doing work on the gas. 0

1200, 50% 1800, 50% 1200, 20%

-0.1 -0.2 PMEP (bar)

Clearly, the deterioration of brake thermal efficiency with increasing EGR is non-negligible, and it is more pronounced at 1200 rpm / mid load conditions than at the other two sets. The two main causes for decreasing brake thermal efficiency are attributed to decreased combustion work (i.e. indicated work) and increased pumping work (assuming that friction remained constant). The decreased combustion work is the consequence of combustion degradation due to lower combustion temperatures and changes in A/F ratio. Flowing EGR reduces the concentration of oxygen, hence directly affecting fuel-air mixture composition in a CI engine. This is in contrast to a premixed SI engine, where EGR replaces a unit of air and fuel with an equal unit of burned exhaust products, hence keeping the oxygen to fuel ratio unaltered. The increased pumping work results from VGT settings increasing exhaust manifold pressure in order to force the flow of residual from the exhaust to the intake side.

-0.3 -0.4 -0.5 -0.6 -0.7 -0.8 -0.9 0

5

10

15

20

25

EGR Rate (%)

Figure 5: Pumping Mean Effective Pressure versus EGR Rate for each of the three operating conditions.

PMEP increases drastically with the increase of the direct consequence of turbine geometry required to provide adequate flow of residual. provides more details, illustrating variations of

EGR, as changes Figure 6 both the

intake manifold pressure and the pressure difference between the exhaust and the intake manifold, as a function of exhaust manifold pressure.

Intake Manifold Pressure (kPa)

EGR Flow Rate

1200, 50% 1800, 50% 1200, 20%

218 198

35 30 25

178

20 158

15

138

10

118

5

98 98

118

138

158

178

198

218

Manifold Pressure Difference (kPa)

40

238

0 238

Exhaust Manifold Pressure (kPa)

For the 1200 RPM - mid load condition, over 60% of the decrease in thermal efficiency is due to degraded combustion, with the balance attributed to the increased pumping work associated with driving the EGR. In contrast, for the 1800 RPM and 1200 RPM - low load cases, more than 75% of the thermal efficiency reduction is attributed to increased pumping work. Figure 8 shows a similar histogram, but for the increase in the EGR rate from 10% to 20%. Contrasts are even more pronounced then in the previous comparison. The drop in thermal efficiency at 1200 RPM - mid load is due almost entirely to degradation of combustion, whereas almost the entire loss of thermal efficiency at 1200 RPM - low load and 1800 RPM- mid load can be attributed to increased pumping work.

Figure 6: Intake Manifold Pressure versus Exhaust Manifold Pressure. Also shown is the pressure difference between the exhaust manifold and the intake manifold.

1

For all three operating conditions, variations of VGT settings provide steady increase in pressure difference across the engine. In other words, exhaust manifold pressure is higher than the intake manifold pressure, and the increased difference between the two allows higher flow of EGR. Boost pressure increases as well, but the rate of increase is significantly smaller than the rate of increase of exhaust back pressure. In addition, the boost pressure curve for the 1200 RPM - mid load case tapers off, and actually starts to decline for higher levels of EGR. While Figures 4 and 5 have illustrated changes in gross and net work, further analysis is required to assess the relative contributions to the loss of engine thermal efficiency. Hence, changes in gross IMEP and PMEP were calculated as EGR rate increased. Dividing the difference in gross IMEP by the decrease in net IMEP indicates the percentage of thermal efficiency loss due to degraded combustion. Similarly, dividing the difference in PMEP by the decrease in net IMEP indicates the percentage of thermal efficiency decrease due to increased pumping work. Distributions of losses, as EGR is changed from 0% to 10%, are given in Figure 7. 1 Fraction of Net Work Loss

0.9 0.8 0.7 0.6 Pumping Loss Combustion Loss

0.5 0.4 0.3 0.2 0.1 0 1200, 50%

1800, 50%

1200, 20%

Figure 7: Fraction of net work loss contributed to either an increase in pumping work or a decrease in combustion (gross) work when increasing EGR from 0% to 10%.

Fraction of Net Work Loss

0.9 0.8 0.7 0.6 Pumping Loss Combustion Loss

0.5 0.4 0.3 0.2 0.1 0 1200, 50%

1800, 50%

1200, 20%

Figure 8: Fraction of net work loss contributed to either an increase in pumping work or a decrease in combustion (gross) work when increasing EGR from 10% to 20%.

In summary, the behavior of the system and the distribution of losses contributing to the reduction of thermal efficiency with elevated EGR are markedly different at 1200 RPM - mid load, than at other two operating regimes. To explain this, one must consider interactions in the engine system, i.e. the relationship between in-cylinder processes and manifold conditions driven by the VGT’s operating characteristics. In-Cylinder Processes EGR routes exhaust gas from preceding engine combustion cycles into the combustion chamber for succeeding combustion cycles. Therefore, the initial composition of the succeeding cycle’s mixture contains concentrations of burned combustion products, i.e. residual. These products primarily include carbon dioxide (CO2) and water (H2O), and much smaller concentrations of carbon monoxide, nitric oxides, hydrocarbons, particulates, sulfur dioxides, sulfates, etc. As the concentrations of the exhaust species in the intake charge increase, the concentrations of oxygen decrease. The ultimate effect on in-cylinder composition and total mass depends on the system response, examined in the previous sub-section. A closer look at Figure 6 reveals that boost actually decreases for 1200 RPM mid-load as EGR rate increases from 10% to 20%,

due to insufficient enthalpy of the exhaust gas. This is expected to significantly alter the A/F ratio and combustion. In contrast, boost does maintain an increasing trend for the same level of EGR increase at 1800 RPM. Increased boosting comes with its price in terms of pumping work, but should provide better conditions for combustion.

intercooler), as well as increased internal residual fraction (due to increased exhaust manifold back pressure, see Figure 6). A sense of increased initial temperature can be obtained from Figure 17. 6.5

Figure 9 illustrates changes of A/F ratio for the three reference operating points. The slopes of the two lines obtained at 1200 RPM are similar, leading to overall reduction of A/F ratio of 24% (low-load operation) and 31% (mid-load operation) over the complete EGR range. However, the curve corresponding to mid load operation is at the much lower level, and the A/F ratio drops to only 20 for 20% EGR. This near-rich A/F ratio is a major factor contributing to the deteriorated combustion efficiency with EGR at 1200 RPM - mid load. Boost pressure trends at 1800 RPM - mid load allow maintenance of higher A/F ratio values throughout the range, thus preventing deterioration of combustion. 50

A/F Ratio

45 40

24% Decrease

35

18% Decrease

30 25

1200, 50% 1800, 50% 1200, 20%

20

31% Decrease

15 0

5

10

15 EGR Rate (%)

20

Ignition Delay (deg)

6 5.5 5

1200, 50% 1800, 50% 1200, 20%

4.5 4 3.5 3 0

5

10

15

20

25

EGR Rate (%)

Figure 10: Ignition delay for each engine condition versus EGR Rate.

Increased gas temperature has a tendency to shorten ignition delay. Secondly, the A/F ratio decreases with increasing EGR. This is confirmed by Figure 9 for all operating conditions. A decreasing A/F ratio has a tendency to shorten ignition delay, as discussed by Assanis et al. [15] and Xia et al. [16]. Other effects, such as cylinder pressure also impact ignition delay, but to a lesser extent. In summary, the increased temperature and reduced A/F ratio seem to offset the dilution effect on ignition delay observed by Ladommatos et al. [14], hence the ignition delay remains generally unchanged with increasing EGR.

25

Figure 9: Variations of Air / Fuel Ratio with EGR for each engine operating condition.

When EGR in a diesel engine displaces a unit of fresh air with an equal unit of burned exhaust products, it not only alters A/F ratio, but causes a dilution effect, as described by Ladommatos, et al. [6-8]. By reducing the oxygen concentration, the mixing time between the direct-injected fuel and the fresh oxygen increases. This is expected to increase the ignition delay and reduce the burn rate once diffusion combustion starts, in case all other parameters are kept constant. However, variations of ignition delay as EGR increases, given in Figure 10, show no apparent increase in ignition delay for any of the tested conditions. These measurements contradict the general expectations described above and results published by Nitu, et al. [13] and Ladommatos, et al. [14]. The apparent insensitivity of ignition delay to the residual fraction is the consequence of realistic conditions occurring in the multi-cylinder engine. Firstly, the gas temperature in the cylinder increases somewhat with increasing EGR. This is the consequence of increased intake temperature (due to residual flow not being cooled down enough to match the air temperature after the

Finally, the combined effects of flowing EGR in the multicylinder engine system on in-cylinder pressure histories for 1200 RPM and 1800 RPM - mid load operation are presented in Figures 11 and 12, respectively. At 1200 RPM – mid load, as EGR rate increases from 10% to 20%, the peak pressure, as well as the integrated area of the high-pressure loop, decreases significantly (see Figure 11). In contrast, at 1800 rpm – mid load, Figure 12 shows that as EGR rate increases, the boost pressure and the peak pressure increase as well, resulting in high-pressure p-V loops that appear to be almost identical in shape and just translated upwards. Consequently, negative side-effects of increased EGR on indicated efficiency appear closely related to the overall A/F level. In other words when the engine operates with overall low A/F ratios, relative changes of mixture composition due to EGR have much more impact. If the turbocharging system is able to provide sufficiently increased boost, e.g. at 1800 RPM, hence maintaining higher A/F ratio levels, the combustion deterioration is practically negligible. This observation, as well as apparent insensitivity of the ignition delay to EGR variations, seems to indicate that the dilution effect of EGR has a much smaller role.

100

Compared to the other two conditions, the initial emission level and the rate of NOx reduction are much greater at 1200 RPM - mid load. Hence, low A/F ratio conditions are more sensitive to EGR, and both efficiency and nitric oxides emission trends would benefit from improved boosting technology at low speed - mid load.

0% 10% 20%

90 Pressure (bar)

80 70 60

Increasing EGR Decreasing Boost

50 40 30 20 150

200

250

300 350 Volume (cm^3)

400

450

500

Figure 11: Pressure versus Volume for 1200 RPM, mid load condition with varying levels of EGR, and boost pressure.

100

Increasing EGR, Increasing Boost

90

20% EG 10%

80 70

1600

0%

1400

60

Nitric Oxides (ppm)

Pressure (bar)

The traditional assumption is that increasing engine boost, and subsequently peak cylinder pressures results in increased NOx emissions. The curves given in Figure 13 show that the effect of residual far exceeds the potential impact of increased boost on NOx. Furthermore, the slopes of the trend-lines corresponding to engine operation at 1800 RPM - mid load and 1200 RPM - low load are almost the same, thus indicating that there is no apparent relationship between increased pressure in the cylinder at 1800 RPM and changes in NOx emissions.

50 40 30 20 150

200

250

300 350 Volume (cm^3)

400

450

500

1200, 50% 1800, 50% 1200, 20%

1200 1000 800 600 400 200

Figure 12: Pressure versus Volume for 1800 RPM, mid load condition with varying levels of EGR and boost pressure.

0

5

10

15

20

25

EGR Rate (%)

Figure 13: Nitric Oxide emissions versus EGR Rate.

In conclusion, careful optimization of system parameters timing is required in order to minimize fuel economy penalties associated with the application of EGR on realistic, multi-cylinder turbocharged diesel engines with VGT. Combustion deterioration is the predominant reason for efficiency losses under low speed - mid load conditions where relatively low boost pressure levels might lead to critically low A/F values. For conditions characterized by higher overall A/F ratio, e.g. low speed low load and mid speed (and high boost) - mid load, most of the fuel economy deterioration can be attributed to the increase in pumping work resulting from restrictive VGT setting required to produce desired pressure drop across the engine. Careful matching and optimized variable geometry turbine control should be considered concurrently in order to effectively address both of the issues. In addition, re-optimized injection timing can alleviate part of the penalty stemming from low A/F ratio operation.

EFFECTS OF EGR ON EMISSIONS Effectiveness of EGR on NOx Reduction The trends of NOx emissions with EGR rate are shown in Figure 13. For all three operating conditions, NOx concentrations decrease with increasing EGR.

Discussion of NOx Reduction Mechanisms and Modeling Guidelines Published literature [6-8] cites three mechanisms via which EGR affects combustion, and hence NOx: formation and reduction: •

Dilution Mechanism: The potentially increased mixing time and longer burn duration caused by EGR’s dilution effect result in lowered flame temperatures.

•

Thermal Mechanism: The increased heat capacity of an EGR-laced mixture results in lowered flame temperatures.

•

Chemical mechanism: Increased dissociation from the more complex EGR molecules (such as CO2 and H2O) result in lowered flame temperatures.

In our study, combination of classical measurements, such as pressure-based diagnostics, and advanced incylinder visualization techniques, such as the videoscope and two-color pyrometry, has been used to assess the relative importance of those flame temperature reducing

mechanisms, and also to guide future modeling efforts. In turn, the mechanisms associated with NOx formation and destruction strongly depend on flame temperature [17]. In addition, for NOx formation to occur, high concentrations of nitrogen and oxygen must also be present. The combustion inside a diesel engine provides both these essential conditions. Figure 14 illustrates the very heterogeneous burning environment in the heavyduty DDC diesel engine combustion chamber, with bright flame regions naturally bordering dark regions containing mostly air (i.e. oxygen and nitrogen). This image was obtained at 1200 RPM - 30% load using the AVL videoscope, and it was captured at four degrees CA after TDC. Based on the rate of heat release analysis corresponding mass fraction burned was roughly 65 %. The flame temperature contour plot, shown in Figure 15, was obtained by processing the videoscope image using the two-color pyrometry method. The plot corresponds to a different combustion cycle compared to the flame image shown in Figure 14, but at the same crank angle and under the same operating conditions. The hottest spots are close to the periphery of the flame region, thus being in contact with the unburned charge with the highest concentration of nitrogen and oxygen, resulting in high NOx formation. The sets of flame temperature contours shown in Figure 16 corroborate the dependence of flame temperatures on EGR rate. Each set contains a sequence of five images collected in the interval between 1 deg CA BTDC and 3 deg CA ATDC. Four sets are presented, corresponding to operation with no EGR, and 10%, 20% and 35% EGR. Comparison of the sets of flame images in Figure 16 clearly shows that, as the EGR rate increases, flame temperatures decrease significantly. As an example, peak flame temperature at TDC without EGR is 2700 K, while recirculation of 20% of the residual reduces the peak flame temperature to only 2400 K. Hence, EGR does have a critical impact on lowering the temperature of burning. It should be noted that the effect of EGR on flame temperatures should not be confused with bulk gas temperature trends, as will be shown in the latter part of this section. Clearly, the thermal EGR mechanism plays a dominant role in the observed reduction in flame temperatures, and hence NOx emissions, with increasing EGR rate. Residual contains increased concentrations of carbon dioxide and water. Both these molecules have higherthan-air heat capacities at typical combustion temperatures (with H2O being much higher than CO2). During the combustion process, these molecules must attain the same temperature as that of the flame front, which requires pre-heating by the flame. With the higher

heat capacity of the mixture, more energy is required to pre-heat the incoming mixture, thus lowering the flame temperature. Ladommatos et al. [18] demonstrated that, although CO2 concentrations are roughly twice that of water vapor, the increased H2O contributes the largest to the thermal effect due to its significantly higher heat capacity (on a mass basis). On the other hand, lower concentrations of nitrogen and oxygen could alter the NOx formation rates. Durnholz et al. [19] indicated that the reduced probability that fuel and oxygen molecules meet (EGR dilution effect) and the higher heat capacity of the working fluid (EGR thermal effect) influence the reduction of NOx. However, the significance of the dilution effect of EGR remains an open question. Does the reduced temperature caused by the dilution effect of EGR reduce NOx more than the reduced oxygen caused by the same effect? Our results indicate a very small effect of dilution on ignition delay (see Figure 10) and a very strong effect of EGR on flame temperatures, thus suggesting the predominance of the thermal effect. This is in agreement with Mitchell et al. [20], who argued that the flame temperature has more influence on NOx formation than the available oxygen. Note that the chemical mechanism involving dissociation of EGR species ultimately lowers the flame temperature, as well. Again, EGR increases the concentrations of CO2 and H2O, and energy is required to dissociate these molecules during the combustion process (particularly H2O, which has a highly endothermic dissociation mechanism). The energy is naturally obtained from the high temperature flame front. Although the dissociation effect is non-negligible, it is considerably less influential than the thermal effect [18]. Even though this work focuses on an experimental investigation, modeling issues are considered by assessing how valuable the measurements are for model development. A predictive, physically-based model for NOx formation, implemented in the engine system simulation, could contribute significantly to the advanced development and evaluation of strategies for reducing NOx emissions. A particular aspect that seems to be critical in any analysis of NOx emissions is the gas temperature during combustion. Previous discussion in this section and results shown in Figure 16 indicate the strong dependence of NOx emissions on local flame temperatures. However, engine cycle simulations often utilize the bulk gas temperature to predict NOx emissions. However, the bulk mean gas temperature and flame temperature do not necessarily correlate very well.

Figure 15: Flame temperature contours extracted from a flame image obtained in a heavy-duty Diesel engine operating at 1200 RPM – 30% load.

10 % 35 %

20 %

EGR Rate

0%

Figure 14: Flame image from a heavy-duty Diesel engine operating at 1200 RPM – 30% load. Bright regions indicate flame areas.

-1

0

1 2 Engine Crankangle (deg aTDC)

3

Figure 16: Flame temperature images obtained using the two-color methodology indicating decreasing flame temperature with increasing EGR rate at 1200 RPM.

As an illustration, the bulk gas temperature profile at the 1200 RPM, mid-load condition is shown in Figure 17 for the two extreme EGR cases studied (0% and 20%). Clearly, the mixture temperature actually increases as EGR increases. Since this temperature calculation was based on the ideal gas law (pV=mRT), this could only happen if either cylinder pressure increased, or cylinder trapped mass decreased, or initial temperature increased. Based on Figure 11, pressure decreased with increased EGR. However, as indicated in Figure 18, intake temperature increased as EGR increased beyond 7%, due to the inability of the EGR cooler to lower EGR temperature below the coolant temperature (which is generally 80° to 90° C). The initial slight decrease in intake temperature is the result of decreased boosting and improved EGR cooling when the flow rate of residual is low. The trapped mass did indeed decrease by 16% for the extreme case of 20% EGR.

1900

Increasing EGR Decreasing Trapped Mass

20% EGR

1700

0% EGR

1500 1300

In addition to NOx, emissions, unburned hydrocarbon (HC) and carbon monoxide (CO) emissions were also measured and plotted as a function of EGR rate in Figure 19. In accordance with Figures 7 and 8, only the 1200 RPM - mid load condition clearly indicates combustion degradation, with dramatically increased CO and HC levels above 10% EGR. The other two conditions, characterized by higher overall A/F levels, demonstrate almost no change in HC or CO levels with EGR.

1100

500 300

350

400 Crank Angle (deg)

70

450

500

Figure 17: Mixture temperature at 1200 RPM, mid load for two extreme EGR conditions versus engine crank angle.

The continuous decrease in the trapped mass with increasing EGR in Figure 18 is related to the fact that at 1200 RPM - mid load, the VGT cannot provide enough boost pressure to compensate for the effect of the residual mass replacing the air (this is effectively confirmed by Figure 9). In summary, increased intake temperature leads to increased gas temperatures during compression, while reduced trapped mass and A/F ratio explain the high fluctuation in the bulk gas temperature during combustion. 21

55

20 50

19.5 19

45

18.5

40

18 35

17.5 17

Intake Temperature (C)

60

20.5

30 0

5

10

15

20

25

EGR Rate (%)

Figure 18: Trapped Mass and Intake Temperature versus EGR Rate for 1200 RPM, mid-load condition.

0.6

1200, 50%, HC 1800, 50%, HC 1200, 20%, HC 1200, 50%, CO 1800, 50%, CO 1200, 20%, CO

60 50 40 30

0.5 0.4 0.3 0.2

20

0.1

10 0

Carbon Monoxide (% vol)

Trapped Mass decreases 16% as EGR rate increases 20%

900 700

Trapped Mass (g/cycle)

Effect of EGR on Other Emissions

Unburned Hydrocarbons (ppm)

Mixture Temperature (K)

2100

The higher bulk gas temperatures during combustion seem to contradict the NOx reduction measured with 20% EGR. This finding could incorrectly suggest that a reduction in oxygen supply drives the NOx reduction, while in fact flame temperatures which decrease significantly with 20% EGR (see Figure 16) are responsible. Consequently, a proper NOx formation modeling approach must be based on flame temperature computations and not bulk mean gas temperatures. Furthermore, advanced diagnostic techniques, such as two-color pyrometry or laser diagnostics, are required for model development and validation.

0 0

5

10

15

20

25

EGR Rate (%)

Figure 19: Unburned hydrocarbons and carbon monoxide versus EGR rate for the three studied operating conditions.

EFFECTS OF EGR ON SYSTEM HEAT REJECTION One of the secondary, and yet important aspects of applying cooled EGR on a heavy duty diesel engine, is the increased heat rejection from the engine. In order to withstand the corrosive nature of diesel exhaust, the cooler should be constructed of corrosion-resistant material, such as stainless steel. Hence, both packaging and cost considerations require that the cooler be made as small and efficient as possible. The easiest and most widely considered way to implement cost-effective EGR cooling is to use the engine coolant to reduce the exhaust gas temperature. The high flow rate of engine coolant coupled with its high heat capacity allow for the design of a compact EGR cooler. However, more heat must be rejected from the vehicle’s radiator, which

In order to better understand the system heat rejection issues, an energy balance was performed and wall temperatures were measured for various EGR levels. Figure 20 illustrates the engine energy balance versus EGR rate. The y-axis indicates the fraction of total fuel energy attributed to the power at the flywheel, heat rejection, and exhaust energy. The power is determined from brake torque and speed measurements. The heat rejection to coolant is calculated by measuring the coolant temperature difference across the engine and the coolant flow rate. The exhaust energy is calculated by speciating the exhaust, measuring the exhaust temperature (out of the turbocharger), and measuring the exhaust flow rate (assumed as mass of air plus mass of fuel). The enthalpies of the exhaust species are determined from a lookup table. The combustible constituents of EGR (such as HC or CO) were neglected in the computation of fuel energy input.

Exhaust Energy

0.8 0.7 0.6

As an example of decreased heat transfer to the walls, Figure 21 shows surface temperature measurements at the fire-deck under 1200 RPM - mid load operation with both 0% and 20% EGR. The fire deck surface temperature decreases as EGR rate increases. The amplitude of the temperature fluctuation due to combustion is lower as well, indicating lower rates of heat transfer to the wall. This can be correlated with the previous discussion of combustion degradation under these particular operating conditions. 225 220 215

0% EGR

210

20% EGR

205 200 195 190

0.5 0.4 0.3

Heat Rejection

0.2 0.1 0

Power

0

5 Power

185 550

600 650 Cylinder 3 Crank Angle (deg)

700

Figure 21: Cylinder 3 Fire Deck Surface Temperature at 1200 RPM, mid-load for 0% EGR and 20% EGR. 10

15 20 25 EGR (%) Power + Qcool Power + Qcool + Qexh

30

Figure 20: Energy Balance of Diesel engine versus EGR Rate for 1200 RPM, low-load operation.

As EGR rate increases, the relative mechanical energy at the flywheel decreases. This decrease is the result of less efficient energy conversion. The heat rejection fraction increases as a direct consequence of using engine coolant to cool the residual, as illustrated by the increasing gap between the bottom and the middle lines. The exhaust energy also decreases with increased EGR, primarily as a result of lower exhaust flow rate, since a fraction of exhaust energy is recirculated. This is often complemented with a decrease of exhaust temperature. However, under certain conditions the exhaust back pressure may actually lead to higher exhaust temperatures, partly offsetting, but not reversing the main heat rejection trend. Finally, heat losses to unknown sources, indicated by the gap between the top curve and the limit (1 or 100%) increase slightly with increasing EGR. This is possibly due to increased losses to surroundings from the piping that routes hot residual to the cooler. It is important to emphasize that, while the total system heat rejection increases with EGR, the heat rejection from the combustion chamber does not necessarily increase. In fact, under certain conditions it might

The surface temperature behaves differently at 1800 RPM - mid load, as shown in Figure 22. While introducing 20% EGR still reduces the amplitude of the temperature fluctuation during combustion, the peak temperature is practically the same as in the case of no EGR. This is the consequence of higher surface temperatures during compression, driven by increased boost, as well as higher temperatures of the coolant leaving the EGR cooler and entering the cooling jackets. Hence, system effects are more than offsetting the impact of somewhat deteriorated combustion on heat transfer and leading to overall higher combustion chamber wall temperatures. 176 174 Surface Temperature (C)

Fraction Total Energy

1 0.9

decrease, due to lower combustion temperatures. In some other cases, the increased mixture temperature throughout the cycle due to insufficient residual cooling will prevail and cause increased convection to combustion chamber walls.

Surface Temperature (C)

necessitates a larger radiator and more “under-the-hood” airflow.

172 170 168 166

20% EGR 0% EGR

164 162 160 158 550

600 650 Cylinder 3 Crank Angle (deg)

700

Figure 22: Cylinder 3 Fire Deck Surface Temperature at 1800 RPM, mid-load for 0% EGR and 20% EGR.

pumping work. Increased pumping work is the result of VGT’s actions necessary to maintain the pressure drop across the engine required for driving the EGR flow. Hence, there are significant opportunities for reducing the fuel economy penalty through optimization of design and control of the turbocharging and EGR system.

SUMMARY AND CONCLUSIONS An experimental study was conducted to investigate the practical application of EGR on a production-type heavyduty diesel engine, and its effect on performance, emissions and heat rejection. The study addressed EGR effects on in-cylinder processes, such as combustion efficiency, flame temperatures and NOx formation, as well as system level effects, such as the variations of pumping work due to changing VGT flow characteristics or the impact of EGR cooling on heat rejection. Quantification of the sources of system inefficiencies shed an improved understanding of the fuel economy penalty associated with EGR, thus providing direction for future work on optimizing the engine charging system design and its control. In particular, the study examined in detail operation with various levels of EGR under three characteristic operating conditions: low speed/low load, low speed/midload and mid speed/mid-load. The following conclusions can be drawn: •

In-cylinder flame temperature contours, obtained through application of videoscope technology and two-color pyrometry, reveal a distinct decrease in flame temperature as EGR increases. This happens despite the fact that bulk gas temperature may actually increase with increasing EGR, due to increasing intake charge temperature and decreasing trapped mass. Hence, significant reductions of nitric oxides with increasing EGR can be correlated with flame temperatures effects. Correct modeling and validation of the combustion flame temperature is therefore a prerequisite for accurate predictions of NOx formation, as opposed to using bulk gas temperature.

•

Engine thermal efficiency tends to decrease with EGR as a result of decreasing indicated work and increasing pumping work. However, re-optimized injection timing could offset part of the negative impact of changing EGR rates.

•

The somewhat increased temperature and reduced A/F ratio seem to offset the dilution effect on ignition delay, hence the ignition delay remains generally unchanged with increasing EGR for given test conditions.

•

The total system heat rejection increases significantly due to EGR cooling. However, the heat rejection from the combustion chamber may increase or decrease. The former is attributed to increased bulk gas temperatures. The latter is observed in cases with deteriorated combustion, leading to reduced peak combustion temperatures.

ACKNOWLEDGEMENTS The rate of NOx reduction with increased EGR is much steeper under low A/F ratio conditions. Increased boost provided by VGT at higher engine speed does not appear to have a negative effect on NOx emissions.

•

•

•

Combustion deterioration is predominant at higher load/low speed and low boost conditions, due to a significant decrease of A/F ratio with increase of EGR. This is accompanied by increased CO and unburned HC emissions. For conditions allowing the VGT to maintain high enough boost and hence A/F ratio, combustion deterioration with increased EGR is minimal and efficiency losses are largely attributed to increased

The authors would like to acknowledge the technical and financial support of the Automotive Research Center (ARC) by the National Automotive Center (NAC), and the U.S. Army Tank-Automotive Research, Development and Engineering Center (TARDEC). The ARC is a U.S. Army Center of Excellence for Automotive Research at the University of Michigan, currently in partnership with the University of Alaska-Fairbanks, Clemson University, University of Iowa, Oakland University, University of Tennessee, Wayne State University, and University of Wisconsin-Madison. Dr. Walter Bryzik and Dr. Peter Schihl of TARDEC provided valuable guidance and significantly influenced the scope of this research. Additionally, the authors wish to thank Detroit Diesel Corporation, and in particular Tim Schafer and Nabil Hakim for their financial, hardware and technical support throughout this study. Finally, Kevin Morrison and Pin Zeng of the University of Michigan are thanked for their assistance in the engine test cell.

REFERENCES [1] Berger, L.B., M.A. Elliot, T.C. Holts, and H.H. Schrenk, “Diesel Engines Underground – II. Effect of Adding Exhaust Gas to Intake Air.” U.S. Dept. of Int., Bur. Of Mines, R.I. 3541, pp. 7867, Nov. 1940. [2] Miller, A.E., S. Sullis, “Recycling Exhaust Gas for Suppressing of Knock in I.C.E.” Proc. Am. Petroleum Inst., Sec. 111, p. 151, 1953.. [3] Haagen-Smit, and M.M. Fox, “Automobile Exhaust and Ozone Formation,” SAE Golden Anniversary Meeting, Detroit, Michigan, January 1955.

[4] Kopa, R.D., and H. Kimura, “Exhaust Gas Recirculation as a Method of Nitrogen Oxides Control in rd an Internal Combustion Engine,” APCA 53 Annual Meeting, Cincinnati, Ohio, May 1960. [5] Yee, S.Y., and W. Linville, “The Effect of Exhaust Gas Recirculation on Oxides of Nitrogen,” rd ACPA 53 Annual Meeting, Cincinnati, Ohio, May 1960. [6] Ladommatos, N., S.M. Abdelhalim, H. Zhao, and Z. Hu, “The Dilution, Chemical, and Thermal Effects of Exhaust Gas Recirculation on Diesel Engine Emissions – Part 1: Effect of Reducing Inlet Charge Oxygen,” SAE Paper 961165, International Spring Fuels and Lubricants Meeting, Dearborn, Michigan, 1996. [7] Ladommatos, N., S.M. Abdelhalim, H. Zhao, and Z. Hu, “The Dilution, Chemical, and Thermal Effects of Exhaust Gas Recirculation on Diesel Engine Emissions – Part 2: Effects of Carbon Dioxide,” SAE Paper 961167, International Spring Fuels and Lubricants Meeting, Dearborn, Michigan, 1996. [8] Ladommatos, N., S.M. Abdelhalim, H. Zhao, and Z. Hu, “The Dilution, Chemical, and Thermal Effects of Exhaust Gas Recirculation on Diesel Engine Emissions – Part 3: Effects of Water Vapour”, SAE Paper 971659, International Spring Fuels and Lubricants Meeting, Dearborn, Michigan, 1997. [9] Arnold, S., K. Slupski, M. Groskreutz, G. Vrbas, R. Cadle, S. Shahed, “Advanced Turbocharging Technologies for Heavy-Duty Diesel Engines,” SAE Paper 2001-01-3260, 2001. [10] Pfeifer, A., M. Smeets, H.O. Herrmann, D. Tomazic, F. Richert, and A. Schlober, “A New Approach to Boost Pressure and EGR Rate Control Development for HD Truck Engines with VGT”, SAE Paper 2002-010965, SAE International Congress and Exposition, 2002. [11] Baert, R.S.G., D.E. Beckman, and A. Veen, “Efficient EGR Technology for Future HD Diesel Engine Emission Targets”, SAE Paper 1999-01-0837, SAE International Congress and Exposition, Detroit, Michigan, 1999.

[12] Hawley, J.G., F.J. Wallace, A. Cox, R.W. Horrocks, and G.L. Bird, “Reduction of Steady State NOx Levels from an Automotive Diesel Engine Using Optimised VGT/EGR Schedules,” SAE Paper 1999-010835, SAE International Congress and Exposition, Detroit, Michigan, 1999. [13] Nitu, B., I. Singh, L. Zhong, K. Badreshany, N. Henein, W. Bryzik, “Effect of EGR on Autoignition, Combustion, Regulated Emissions and Aldehydes in DI Diesel Engines,” SAE Paper 2002-01-1153, SAE International Congress and Exposition, 2002. [14] Ladommatos, N., S. Abdelhalim, H. Zhao, Z. Hu, “Effects of EGR on Heat Release in Diesel Combustion,” SAE Paper 980184, SAE International Congress and Exposition, 1998. [15] Assanis, D. N., Filipi, Z. S., Fiveland, S. B., Syrimis, M.,”A Predictive Ignition Delay Correlation Under Steady-State and Transient Operation of a Direct Injection Diesel Engine,” Proceedings of 1999 ASME-ICE Fall Technical Conference, Paper # 99-ICE-231, Vol. 332, Ann Arbor, MI, 1999. [16] Xia, Y. Q., and Flanagan, R. C., “Ignition Delay A General Engine/Fuel Model,” SAE Paper 870591, 1987. [17] Heywood, John B., Internal Combustion Engine Fundamentals, McGraw-Hill, New York, 1988. [18] Ladommatos, N., S.M. Abdelhalim, H. Zhao, and Z. Hu, “The Dilution, Chemical, and Thermal Effects of Exhaust Gas Recirculation on Diesel Engine Emissions – Part 4: Effects of Carbon Dioxide and Water Vapour”, SAE Paper 971660, International Spring Fuels and Lubricants Meeting, Dearborn, Michigan, 1997. [19] Durnholz, M., G. Eifler, and H. Endres, “ExhaustGas Recirculation – A Measure to Reduce Exhaust Emissions of DI Diesel Engines,” SAE Paper 920725, 1992. [20] Mitchell, D.L., J.A. Pinson, and T.A. Litzinger, “The Effects of Simulated EGR via Intake Air Dilution on Combustion in an Optically Accessible DI Diesel Engine,” SAE Paper 932798, 1993.