Printer Friendly

Optical investigations of soot formation mechanisms and possible countermeasures on a turbocharged port fuel injection SI engine.


Despite the known benefits of direct injection (DI) spark ignition (SI) engines, port fuel injection (PFI) remains a highly relevant injection concept, especially for cost-sensitive market segments. Since particulate number (PN) emissions limits can be expected also for PFI SI engines in future emission legislations, it is necessary to understand the soot formation mechanisms and possible countermeasures. Several experimental studies demonstrated an advantage for PFI SI engines in terms of PN emissions compared to DI. In this paper an extended focus on higher engine loads for future test cycles or real driving emissions testing (RDE) is applied. The combination of operating parameter studies and optical analysis by high-speed video endoscopy on a four-cylinder turbocharged SI engine allows for a profound understanding of relevant soot formation mechanisms. For selected operating points, engine operating parameters such as injection timing, inclination of a charge motion flap, and engine coolant temperature were varied. Furthermore, the impact of two different spray layouts on the mixture formation was evaluated. Parameter sets showing significant reduction of PN emissions were subsequently analyzed using high-speed video endoscopy. Optical access to both the intake port as well as the combustion chamber allows visualization of fuel transport mechanisms leading to diffusion flames and soot emissions respectively. In summary, this study shows that port fuel injection at high engine loads can lead to significant PN emissions. Three locations within the combustion chamber could be identified as sources for diffusion flames leading to particulate emissions. The governing parameters allowing substantial reduction of PN emissions at these locations were found to be the injection timing and the charge motion.

CITATION: Schueck, C, Koch, T, Samenfmk, W., Schuenemann, E. et al., "Optical Investigations of Soot Formation Mechanisms and Possible Countermeasures on a Turbocharged Port Fuel Injection SI Engine," SAE Int. J. Engines 9(4):2016.


For spark ignition engines, port fuel injection still holds a significant market share of 54 % next to direct injection in the US [1]. Based on an analysis by Bosch for 2015, around 70 % of world-wide gasoline passenger car sales, including light commercial vehicles below 6 t, were still equipped with port fuel injection [2]. PFI offers an attractive cost-benefit ratio in terms of achievable emission targets combined with the simplicity of the fuel system. In the context of real driving emissions testing operating points with higher engine load will present a challenge for both DI and PFI engines in regards to pollutant emissions. Especially the particulate emissions from DI SI engines are known to raise with increasing load [3,4]. Currently there is no legislation limiting the particulate number emissions of PFI engines [5_,6,7,8]. It can be expected though, that PN emissions will be regulated for PFI in future legislations as well.

When comparing PN emissions of an engine with both port fuel injection and direct injection, previous studies demonstrated a significant advantage for PFI [9]. In [10] PN concentrations were reported to be lower by one order of magnitude using PFI at an indicated mean effective pressure (IMEP) of 8 bar and 1500 rpm. Nevertheless, the presence of particulate emissions from PFI engines has been reported in several previous studies [4,11,12,13,14,15]. The particulate mass (PM) emissions from a boosted PFI SI engine were investigated by empirical relations of opacimeter measurements and two-color pyrometry in [16], giving clues to the intake valves being an important source for diffusion flames.

The primary objective of this work is the identification of soot formation mechanisms within the combustion chamber using high speed video endoscopy. With the PCO Dimax cameras, which allow the acquisition of color images, diffusion flames leading to soot formation can be recognized as flames ranging from yellow to red in color. This color range of soot luminosity can be explained by Planck's law for radiance from a black body, even though soot is not a perfect black body, considering its lower emissivity ([epsilon] [??] 1). The maximum spectral energy density emitted at typical soot temperatures during combustions in SI engines occurs at wavelengths in the near-infrared part of the spectrum. But the emission emerging in the visible spectrum, especially in the red and yellow, is also significant enough to be recorded by the cameras. The distinction from the mostly blue (and UV) chemoluminescence of a homogeneous pre-mixed flame is unambiguous. The physical background for the incandescence of soot particles can be found in more detail in [17,18].

The work presented in this paper intends to obtain an overview for both the influencing parameters and the root causes for PN emissions, specifically from a PFI engine. The identification of PN-critical operating points was the first step in this context. Detailed engine operating parameter studies - in particular injection timing, charge motion, coolant temperature, and fuel injector variations - were performed considering the effects on PN emissions. Injection timing and the intensity of the charge motion are known influencing parameters on PN emissions for DI SI engines [19]. Finally the distinct sources for diffusion flames for PFI were visualized by optical access via endoscopes to both one combustion chamber and one intake port using PCO Dimax high-speed cameras. This will help to understand the underlying soot formation mechanisms of a combustion with mixture preparation by port fuel injection.


The investigations were carried out on a 2.0 1 four cylinder turbocharged SI engine with port fuel injection. The fuel injectors are located in the intake manifold, each closely positioned near the interface to the intake ports of the cylinder head. Detailed specifications of the research engine can be found in Table 1. Throughout all measurements RON 95 fuel with an ethanol content of 10 % was used. The engine was equipped with indicating pressure sensors. Gaseous emissions were acquired by a Horiba Mexa 7100D. The particulate number emissions were measured by a Condensation Particulate Counter (CPC) from TSI, which is integrated in a Horiba Mexa 2100 SPCS. The built-in volatile particle remover (VPR) reduces the risk of measuring small liquid droplets. The particles counted by the SPCS range in sizes from 23 nm to 3 urn. The cut-off curve for small particles is defined by counting efficiencies of 50 % [+ or -] 12 % at 23 nm and > 90 % at 41 nm [20]. Depending on the level of PN emissions total dilution ratios of 1500 or 3000 were applied. The location of the measurement probe for PN is downstream of the catalyst and upstream of the final muffler.

Optical access to the combustion chamber of cylinder 4 was prepared for an endoscope and a fiber optic light guide, the latter to provide illumination for images before the start of combustion. That way the fuel transport mechanism into the combustion chamber can be visualized. A second set of endoscope and light conducting fiber was integrated into the intake manifold to give insight to the intake port of the same cylinder as shown in Figure 1:

a. Intake valve (IV)

b. Outlet valve

c. Sparkplug

d. Tumble plate

e. Charge motion flap (CMF)

f Fuel injector (Bosch EV14)

g. Endoscope into combustion chamber

h. Connector for cooling system of endoscope

i. Endoscope into intake manifold / intake port

j. Additional access into port (used for additional light source)

k. Optic light guide into intake port

j. Optic light guide into combustion chamber

The viewing angles of both endoscopes are sketched at the bottom of Figure 1. Special attention was also paid to the stiffness of the mounts holding the high-speed cameras. Severe vibrations during engine operation can present a risk of damage to the cameras. The trigger signals of both cameras were recorded by the indicating system. This allows a synchronization of the simultaneously recorded high-speed videos of both cameras with the measurement data from the indicating system. Each recorded frame can be assigned to a certain degree of crank angle. The rate of image acquisition by the PCO Dimax high-speed cameras was set to a framerate of 6000 frames/s.

The highest possible engine load and speed, which could be measured, was limited by the temperature at the tip of the combustion chamber endoscope. Therefore, full-load measurements were possible up to the engine speed of 1500 rpm. Operating points which were analyzed with optimized air and fuel path parameters are listed in Table 2.

Two different fuel injectors were used during this study. Both are Bosch EV14 fuel injectors with extended tips and compact injector bodies. Injector A was designed for minimized fuel wall films in combination with a tumble plate and a charge motion flap (CMF). The second injector B takes advantage of the space gained by removing the tumble plate. The different spray layouts are illustrated in Figure 2.

The gamma angle describes how far the spray coils are tilted relative to a virtual plane defined by the mid-axis of the injector. From the perspective of Figure 2, this plane appears as a white dashed line going through the injector body.


Identification of PN Relevant Operating points

In order to determine the engine operating points, which are critical in terms of PN emissions, steady-state measurements in warm operating temperature conditions were performed over a wide range of the engine map without the endoscopy system. This also helped to gain a more global assessment regarding PN emissions from a t/c engine using PFI. Since this was a first step of the investigations, no specific optimizations were applied and the charge motion flap was left deactivated. Spark timing was set up for minimal fuel consumption and for avoiding knocking combustions. Knocking was monitored using the indicating system. A closed-valve injection (cVI) strategy was applied over the whole map. The PN concentration results of this map are summarized in Figure 3.

For engine loads up to brake mean effective pressures (BMEP) of about 10 bar the particulate concentrations measured are low. In this area of manifold pressures below 1000 mbar the PN concentrations scarcely exceed le+04 #/[cm.sup.3]. Just above BMEP > 10 bar, where manifold pressure is increasing, PN emissions raise strongly by finally 3 orders of magnitude at 2000 rpm and full-load. It can be concluded that for PFI operating points with high manifold pressures are especially critical in terms of PN emissions.

Engine Operating Parameter Studies

For a more detailed understanding of the influencing parameters on PN emissions with PFI, load sweeps were performed. Two injection strategies and both charge motion flap positions were investigated using injector A, and the tumble plate in the position shown in Figure 2 on the left. For the closed-valve injection strategy, shown by all dark-blue lines in Figure 4, the end of injection (EOI) was set to 300 [degrees]CA before intake valve closes (IVc). The open-valve injection strategy (oVI) was applied by EOI of 60 [degrees]CA before IVc. A logarithmic scale was used for the PN concentrations on the y-axis.

The highest PN concentrations were found to occur for the case of closed-valve injection and deactivated charge motion flap. As previously seen for the engine map, PN emissions raise rapidly with increasing load. Switching to open-valve injection, still with the charge motion flap deactivated, leads to an improvement regarding PN concentration. The level of PN emissions decreases even further, when the CMF is activated. Combining the open-valve injection strategy with the enforced tumble by the charge motion flap yields the lowest PN emissions. The described effects apply best for the engine speeds of 2000 and 3000 rpm. At the engine speed of 1500 rpm only the high loads show these trends for oVI and activated CMF. The conventional cVI strategy seems to work well for low engine loads at 1500 rpm, where PN emissions are generally very low.

Since the injection strategies showed a significant effect on PN, variations of injection timings were carried out. EOI sweeps with both positions of the CMF are plotted in Figure 5 using injector A. The x-axis represents the crank angle for the end of injection relative to IVc. PN concentrations for an early injection are found on the far left of the diagram, for example. Smaller increments of injection timing were chosen in the region just before and during intake valve open. The duration of the injection is shown only for the latest theoretically possible EOI = 0 [degrees]CA before IVc, for a visual comparison with the event length of the intake valve.

Both curves plotted in Figure 5 represent mean values for PN concentrations from up to three EOI sweeps. If the tumble is not enforced, given by CMF = 0, PN emissions increase for open-valve injection by nearly one order of magnitude. The same measurements repeated with activated CMF = 1 leads to an improvement for PN of around two orders of magnitude. A small benefit by tumble enforcement is also visible for the cVI case. An explanation of such opposing effects regarding PN emissions upon different CMF positions for the oVI strategy was one of the goals for the high-speed video endoscopy measurements.

High-Speed Video Endoscopy

Minimal Soot Emissions Operating Point - Low Engine Load

Before investigating the soot formation mechanisms at high engine loads, an operating point with very good emission results was measured. The objective was to understand how the apparently good mixture formation works for low engine loads, where PN emissions are very low.

Figure 6 shows the injection and intake stroke phase relevant for the fuel transport into the cylinder at 1750 rpm and IMEP = 7.6 bar. Important parameter settings were: usage of injector A without tumble plate, CMF = 0, and cVI set to EOI = 300 [degrees]CA b. IVc. The images illustrate an example of ideal mixture preparation for port fuel injection:

a. First droplets of fuel jet approach intake valves.

b. Fuel accumulates at closed intake valves. Injection has ended.

c. IVo: back flow of burnt gas from combustion chamber into intake port starts due to negative pressure gradient over the IVs.

d. Build-up of back flow still ongoing, facilitating evaporation of fuel droplets and fuel wall films on IVs and intake port.

e. Back flow moved most liquid droplets away from IVs.

f Transition of back to forward flow.

g. Forward flow is accelerating. Most fuel droplets evaporated by now.

h. Few left-over fuel droplet pass IV

One rather large fuel droplet was marked with a white circle to demonstrate the process of back flow and forward flow after IVo. This back flow of hot exhaust gas entering the intake port supports several elements of the mixture preparation: the evaporation and strip atomization of fuel films, the fuel-air mixture process, the break-up of fuel droplets, and generally the evaporation of left over liquid fuel droplets. Since the fuel amount at this operating point with low engine speed and low engine load was small, and since the low manifold pressure allowed an extensive usage of the exhaust back flow, minimal soot emissions were the result. A PN concentration of 2.05e+03 #/[cm.sup.3] is in the region of fresh ambient air in the test cell.

The described back flow effect can only occur as long as there is a negative pressure gradient over the intake valves ([p.sub.INM] < [p.sub.cyl] at IVo). This pressure gradient depends on the manifold pressure and on the camshaft timings, as they influence the cylinder pressure at the time of IVo. For increasing loads, hence increasing manifold pressures, this back flow effect will diminish and eventually disappear. A gas exchange analysis calculated from the pressure traces of the indicating system using the analysis tool BeCAT (Bosch engine Combustion Analysis Tool) is shown in Figure 7. The pressure trace data was acquired simultaneously with the video images at 1750 rpm and IMEP =7.6 bar, which are shown in Figure 6.

The large valve overlap of 36 [degrees]CA and the low intake manifold pressure [p.sub.INM] of about 900 mbar allow the mentioned back flow of exhaust gas into the intake port.

The most interesting combinations of injection timing and CMF positions regarding PN emissions were investigated by recording high-speed videos for several operating points.

Soot Formation from "Liner Wetting"

Images from the high-speed video of the combustion chamber, on the left side of Figure 8, show the fuel transport into the cylinder at the "low-end-torque" (LET) operation point. The white arrows in the image at 80 [degrees]CA after top dead center (TDC) help to visualize the effect of fuel spray being deflected by the intake valve towards the cylinder liner. This region is where the thin diffusion flames come from, rather late at 110 [degrees]CA after ignition on the right hand side. These diffusion flames are moving further towards the center of the combustion chamber, which is shown at 134 [degrees]CA after ignition. This phenomenon of apparent wall wetting of the liner by the oVI followed by diffusion flames was seen repeatedly at this operating point without the tumble enforcement by CMF.

The same operating point with open-valve injection was run again with the CMF activated. Even though it is hardly visible on the left side of Figure 9. the stronger tumble helps to reduce the liner wetting, which occurred for oVI without tumble enforcement at this operating point. No diffusion flames are visible anymore with activated CMF. The PN concentration recorded simultaneously with the high-speed video is more than ten times lower than with CMF deactivated. This is in good agreement with the results from the EOI sweep in Figure 5.

One representative engine cycle was selected for each image series discussed in Figure 8 and Figure 9. In order to visualize the mean behavior of the soot formation at each operating point a statistical analysis was carried out. For that purpose the presence of red diffusions flames was checked for each pixel of all video frames acquired at 110 [degrees]CA a.ign. The result is a spatially resolved frequency of occurrence for diffusion flames out of the 20 cycles, which were recorded. In Figure 10 these statistical results were combined with one background image of a single cycle at the same engine position.

Since no diffusion flames were recorded during all 20 cycles, the statistical analysis result is blank in Figure 10b where the charge motion flap was active (CMF =1).

Soot Formation at Intake Valve Crevice

The dominating soot formation mechanism for closed-valve injection can be studied at the engine operating point 2000 rpm and IMEP =15 bar. In the left column of images in Figure 11 the opening of the intake valve and the fuel transport are shown. Due to the spray layout of injector A, most of the fuel entering the combustion chamber passes the intake valve on the top side closest to the spark plug (white arrow). Even though fuel impingement directly onto the piston surface can be observed at 25 [degrees]CA a. TDC (white circle), no diffusion flames originating from the piston surface occur. This is different from the behavior known from direct injection, where pool-fires are a typical result of fuel impingement onto the piston.

Instead severe diffusion flames originating from the crevice of the intake valve seat were found. An explanatory model is that the liquid fuel film, which was most likely created during the closed-valve injection and during the event of intake valve opening, did not evaporate in time. With the evaporation of the fuel film still ongoing as the flame-front passes the intake valve, the locally rich conditions for soot formation by diffusion flames are given.

Also this operating point with closed-valve injection using the same injector A was repeated with the only difference being the activated CMF. For the intake stroke on the left column of Figure 12 similar effects were seen as with CMF = 0. A difference to the previously discussed effects is that even at 72 [degrees]CA a. TDC fine fuel spray parcels were washed off the intake valve. The results in terms of soot emissions are visible on the right column of Figure 12. Compared to Figure 11 the diffusion flames are not only less intense, but also the measured PN concentration is less than half of the value from without the CMF. The reason for this improvement is the reduced cross-sectional area of the intake port, by the closure of the CMF. The thereby increased velocity of the intake air passing the intake valves helps to evaporate left-over fuel films by more efficient strip atomization.

In addition to the two selected characteristic cycles shown in Figure H and Figure 12, the average behavior for the operating point of 2000 rpm and IMEP =15 bar can be found in Figure 13. The engine position for the statistical analysis of the soot formation was chosen to be 15 [degrees]CA a.ign. for this operating point, as the first diffusion flames appeared at this timing. This helps to identify the spatial origin of the diffusion flames, which is the intake valve crevice as seen before in the single cycle image series.

Even lower PN emissions were measured at 2000 rpm and IMEP = 15 bar for the oVI strategy combined with activation of the CMF yielding PN of 2.74e+05 #/[cm.sup.3]. No more diffusion flames were visible at all for the latter case, neither near the intake valve seat nor on the cylinder liner. It can be derived, that oVI and a tumble enforcement by a charge motion flap both reduced possible fuel wall films. Having seen the soot formation mechanisms, it is easier to understand the ranking of PN emissions found during the load sweeps and EOI sweeps.

Soot Formation in Gas Phase

A third source for soot emissions was observed at the operating point of 3000 rpm and IMEP =13.6 bar. Sporadically occurring small red spots, most likely originating from locally rich zones are shown at 36 and 51 [degrees]CA after ignition in Figure 15 g) and h).

The random distribution of these small diffusion flames within the gas phase of the combustion chamber is even more obvious in the illustration of the statistical analysis shown in Figure 14.

For the measurement described in Figure 14 and Figure 15, the charge motion flap was left deactivated, the tumble plate was removed, and injector B was installed, as its spray was designed to fully utilize the space of the intake port. This operating point and setup was chosen to simulate the scenario of a naturally aspirated engine at higher engine speed and "full-load", where a lack of charge motion can result in inhomogeneities in the fuel air mixture, which might lead to elevated pollutant emissions. By using a slightly boosted operating point the challenge of mixture preparation was intensified. The injection timing of EOI = 144 [degrees]CA b. IVc was chosen to be a combination of open and closed-valve injection in order to take advantage of the backflow, at the instant of IVo.

The intake stroke with this injection strategy is shown in Figure 15 a) - d). Image a) shows the progress of the injection at the time of IVo. The back flow of hot exhaust gas into the intake port can be seen at 30[degrees] CA a. IVo in image b), as the intake valve shaft becomes visible again. As the forward flow of the intake air accelerates in images c) and d), the domination of the air flow becomes visible, since the spray is moved towards the top left side of the intake port, which is different from the spray target region at the closed intake valve from image a).

The backflow into the intake port, which was discussed above and shown as spray effects in Figure 15 b) can also be seen in the gas exchange analysis in Figure 16, even though it is much less intense compared to N = 1750 rpm and IMEP = 7.6 bar (Figure 7).

Despite the usage of the exhaust back flow and the optimized spray targeting, the homogenization of the fuel-air mixture is not sufficient, resulting in the PN emissions visualized in Figure 15 g) - h). An increase of the charge motion is necessary to encounter this problem.

Comparison of Spray Layouts

More load sweeps at 3000 rpm were performed after the installation of injector B and the removal of the tumble plate, in order to compare its performance regarding soot emissions to injector A. The main influencing parameters for PN emissions, injection timing and CMF inclination, were varied for each measurement. Even without the tumble plate being installed, the CMF could still be controlled to be activated (CMF = 1) or deactivated (CMF = 0). All load sweeps set up with closed-valve injection (EOI = 300 [degrees]CA b. IVc) are plotted on the left side of Figure 17. An obvious difference between injector A and B occurred for the case without the CMF and cVI. The wider spray of injector B yields lower PN emissions than injector A.

The reason for this behavior is most likely the narrow spray cone of injector A. As seen at 2000 rpm in Figure 11. the fuel wall film build up during closed-valve injection can lead to significant diffusion flames from the intake valve crevices. The wider spray of injector B

will distribute the fuel to a larger area, which can evaporate more easily in the time available until the start of combustion. Better evaporation conditions were apparently given by activating the CMF. PN results are very similar for both injectors for CMF = 1, as the disadvantage of injector A is compensated by the higher intake air velocity, facilitating the evaporation process at the intake valve crevice area.

When applying the open-valve injection strategy (EOI = 50 [degrees]CA b. IVc) the differences between the two injectors are very small. This is true for both positions of the CMF. The predominance of the air mass flow over the fuel spray reduces the influence of the spray layout for open-valve injection at high air mass flows. This phenomenon was also seen in Figure 15c. The activation of the CMF reduces wall films on the intake valves as well as on the cylinder liner, as seen before, leading to a significant decrease of PN emissions. The small PN benefit of injector B at low engine loads in the region of [10.sup.3] #/[cm.sup.3] should not be overestimated.

Generally the same optimal parameters apply for injector B as found for injector A regarding PN emissions: The combination of open-valve injection and tumble-enforcement by the charge motion flap.

Condition of cold engine

With injector A and the tumble plate installed, EOI variations were carried out. Both charge motion flap positions were tested at a coolant temperature of around 35 [degrees]C. The operating point was 1750 rpm and IMEP = 7.6 bar, which has been measured before at warm engine condition, where PN emissions were extremely low. Three distinctive effects are noticeable in Figure 18.

For late open-valve injections, fuel wall films in the combustion chamber cause high PN emissions, due to the slow evaporation from the cold liner for example. The effect of fuel droplets being carried onto the liner was seen before in Figure 8.

When the back flow effect around IVo was exploited optimally a local minimum was reached in the PN trace. The back flow of hot exhaust gas into the intake port supports droplet break-up and fuel evaporation leading to reduced wall films. This back flow effect was visualized for the same operating point in warm condition in Figure 6. The injection timing at the local PN minimum of EOI =162 [degrees]CA b. IVc cannot be recommended for series calibration though, as a wide and stable calibration windows is needed.

The activation of the CMF enhances the fuel evaporation at these difficult conditions. Smaller droplets by secondary droplet atomization, due to the higher intake air velocity, can evaporate more easily in the combustion chamber. The enforced tumble also improves the homogenization of the air-fuel mixture. Overall the activation of the CMF helps to reduce the PN emissions significantly at cold engine operation.

Images via high-speed endoscopy were acquired in the combustion chamber for the worst case of EOI = 50 [degrees]CA b. IVc and CMF = 0 and also for the best case with regards to PN at EOI =162 [degrees]CA b. IVc, which is referred to as roughly the timing of intake valve opening ("iVO"), and CMF = 1 (see Figure 19). For the case of oVI and deactivated CMF the known sources for diffusion flames were observed, namely the intake valve crevice in Figure 19a and the cylinder liner in Figure 19b next to some more inhomogeneities. Activating the CMF and using the optimized injection timing of EOI = 162 [degrees]CA b. IVc significantly smaller diffusion flames are seen in Figure 19c. Soot emissions from the liner are likely to have reduced too, as they did not move into the field of view anymore (Figure 19d) This seems to be in good agreement with the other PN sources found at warm engine conditions.

At this operating point diffusion flames emerge fist at the intake valve crevice followed by those from the piston liner, approximately 35 [degrees]CA later. Considering this gap in time, two statistical evaluations were carried out for this operating point for both soot formation locations in Figure 20.

More work will be needed to understand the problems and possible countermeasures for PN emissions at the cold engine conditions for port fuel injection. Other sources for soot emissions might occur, when a completely cold engine, immediately after a cold start, is analyzed. The exhaust valves, the piston and also the intake valves are expected to be much colder in this case, compared to the steady-state conditions at low coolant temperature. The highly dynamic operation during the start and engine run-up will present another challenge for the mixture preparation.


Particulate number emission measurements were carried out on a turbocharged PFI SI engine. A large number of operating points was measured under steady-state, warm operating temperature conditions. The mixture formation mechanism leading to very low PN emissions at low engine loads, hence low manifold pressures, was found to be the usage of exhaust backflow into the intake port, facilitating the evaporation of fuel droplets and fuel wall films. The area of PN-critical operating points was identified to be the region of high engine load, especially the boosted area. The combination of engine operating parameter studies, and high-speed video imaging, revealed that injection timing and the tumble enforcement by a charge motion flap (CMF) are the governing parameters in terms of PN emissions. Open-valve injection combined with a strong tumble by the CMF lead to the lowest PN emissions for the majority of operating points. An overview of the soot formation mechanisms found in this study and possible countermeasures are summarized in Figure 21.

Besides the shown beneficial effects in terms of PN emissions, the activation of the CMF can lead to increased fuel consumption. This is due to higher losses during the gas-exchange, as the cross sectional area of the intake ports is reduced. Depending on the engine operation point this downside can be compensated again, for example by reduced duration of the combustion (MFB5 % - MFB90 %) or better spark efficiency by a knock limit shift due to the increased tumble. A detailed discussion of these effects would lead beyond the scope of this paper, though.

The mechanisms and measures were found for steady-state engine operation at four operating points under warm engine conditions. It was also shown, that low engine temperature can lead to PN emissions even at low engine loads, which were uncritical for the warm engine.

More work will be needed in the field of "completely cold engine" - directly after the cold-start. Also the soot formation during catalyst heating as well as for dynamic engine operation shall be analyzed in following investigations.


[1.] U.S. EPA, ,,Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 Through 2015. Trends Report," EPA-420-R-15-016, 2015.

[2.] Bosch GmbH Robert, ,,Powertrain Scenarios," Customer presentation, 2015.

[3.] Kufferath, A., Wiese, W., Samenfink, W., Dageforde, H., et al., ,,Assessment of feasible system solutions for future particle emission requirements," Fuel Systems for IC Engines Conference, 2015.

[4.] Wang, Y., Zheng, R., Shuai, S., Qin, Y. et al., "The Impact of Fuel Properties from Chinese Market on the Particulate and VOCs Emissions of a PFI and a DIG Engine," SAE Technical Paper 2016-01-0838. 2016. doi :10.4271/2016-01-0838.

[5.] European Commission, ,,Commission Regulation (EU) No. 136/2014," 2014.

[6.] California Air Resources Board, ,,Attachment A-l Final Regulation Order," California Code of Regulations, 2012.

[7.] Environmental Protection Agency, "Control of Air Pollution From Motor Vehicles: Tier 3 Motor Vehicle Emission and Fuel Standards; Final Rule," Rules and Regulations 79(81), 2014.

[8.] Central Environment Council (Japan), ,,Future Policy for Motor Vehicle Emission Reduction (Eighth Report)," 2005.

[9.] Hemminger, S., ,,Optimierung der Gemischbildung und Verbrennung von Hochleistungsottomotoren zur Einhaltung zukiinftiger Emissionsgrenzwerte," Ph.D. thesis, University of Stuttgart, 2014.

[10.] Su, J., Lin, W., Sterniak, J., Xu, M. et al., particulate Matter Emission Comparison of Spark Ignition Direct Injection (SIDI) and Port Fuel Injection (PFI) Operation of a Boosted Gasoline Engine," Journal of Engineering for Gas Turbines and Power 136(9):91513-91513, 2014.

[11.] Merola, S., Vaglieco, B., and Di Iorio, S., "Nanoparticles at Internal Combustion Engines Exhaust: Effect on Urban Area," SAE Technical Paper 2006-01-3006. 2006. doi:10.4271/2006-01-3006.

[12.] Chan, T, Meloche, E., Kubsh, J., Rosenblatt, D. et al., "Evaluation of a Gasoline Particulate Filter to Reduce Particle Emissions from a Gasoline Direct Injection Vehicle," SAE Int. J. Fuels Lubr. 5(3):1277-1290, 2012. doi :10.4271/2012-01-1727.

[13.] Merola, S. and Vaglieco, B., ,,Optical investigations of fuel deposition burning in ported fuel injection (PFI) spark-ignition (SI) engine," Energy 34(12):2108-2115, 2009, doi: 10.1007/sl2239-009-0063-8

[14.] Li, Y., Xue, J., Johnson, K., Durbin, T. et al., "Determination of Suspended Exhaust PM Mass for Light-Duty Vehicles," SAE Technical Paper 2014-01-1594. 2014. doi:10.4271/2014-01-1594.

[15.] Merola, S., Sementa, P., Tornatore, C, Carmignani, L. et al., "Fuel Injection Effect on In-cylinder Formation and Exhaust Emission of Particulate from a 4-Stroke Engine for 2-Wheel Vehicles," SAE Technical Paper 2010-01-0354. 2010, doi:10.4271/2010-01-0354.

[16.] Merola, S., Sementa, P., Tornatore, C. and Vaglieco, B., ,,Effect of fuel injection strategies on the combustion process in a PFI boosted SI engine," International Journal of Automotive Technology 10(5):545-553, 2009, doi: 10.1007/sl2239-009-0063-8

[17.] Zhao, H. and Ladommatos, N., "Optical diagnostics for soot and temperature measurement in diesel engines," Progress in Energy and Combustion Science, 24(3):221-255, 1998

[18.] Stojkovic, B., Fansler, T, Drake, M. and Sick V, "High-speed imaging of OH* and soot temperature and concentration in a stratified-charge direct-injection gasoline engine," Proceedings of the Combustion Institute, 30(2):2657-2665, 2005, doi:10.1016/i.proci.2004.08.021

[19.] Bertsch, M., Koch, T, Velji, A., and Kubach, H, "Thermodynamic and Optical Investigations on Particle Emissions in a DISI Engine at Boosted Operation," SAE Int. J. Engines 9(1):154-170, 2016, doi:10.4271/2015-01-1888.

[20.] Horiba, ,,MEXA-2000SPCS series," data sheet accessed Jul. 2016


Dipl. Ing. Claudius Schtick

Bosch Engineering GmbH

Robert-Bosch-Allee 1

74232 Abstatt, Germany



The authors would like to thank the students Le Wang, Moritz Schock, Miguel Pacavita and Markus Lindner for their enthusiasm, effort and research spirit.


PFI - Port fuel injection

PN - Particulate number emissions

RDE - Real driving emissions

MEP - Ministry of Environmental Protection of China

IMEP - Indicated mean effective pressure

PM - Particulate matter

CPC - Condensation Particulate Counter

VPR - Volatile particle remover

t/c - Turbo charged

CMF - Charge motion flap

LET - Low end torque

N - Engine speed

CA - Crank angle

EOI - End of injection

TOI - Time of ignition

a.ign -AfterIgnition

EVo - Exhaust valves open

EVc - Exhaust valves closed

IVo - Intake valves open

a.IVo - After intake valves open

IVc - Intake valves closed

b.IVc - Before intake valve closed

a.TDC -After top dead center

oVI - Open-valve injection

cVI - Closed-valve injection

BMEP - Brake mean effective pressure

IVL - Intake valve lift

Claudius Schueck Bosch Engineering GmbH

Thomas Koch KIT Karlsruhe Institute of Technology

Wolfgang Samenfmk and Erik Schuenemann Robert Bosch GmbH, Gasoline Systems

Stephan Tafel and Oliver Towae Bosch Engineering GmbH

Table 1. Specifications of the research engine used for this study.

Displacement                         2000 cc

Compression ratio                    11.5
Number of cylinders / no. of valves  4/4
Injection system                     port fuel injection
Fuel system pressure                 7 bar
Charging system                      turbocharged (t/c)
Exhaust Valves                       cam phasing
Intake Valves                        2-step valve lift, cam phasing
Charge motion device                 2-step charge motion flap (CMF)

Table 2. Engine operating points for high speed video endoscopy and
simultaneous PN measurements.

N [rpm]  IMEP [bar]  [T.sub.coolant]
                     [[degrees]C]     Comment                  Injectors

1500     12.7        92               full load / LET           A
2000     15.0        92               high PN emissions         A
3000     13.6        92               higher engine speed       B
1750      7.6        35&92            cold engine / part-load   A
COPYRIGHT 2016 SAE International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Schueck, Claudius; Koch, Thomas; Samenfmk, Wolfgang; Schuenemann, Erik
Publication:SAE International Journal of Engines
Article Type:Report
Date:Dec 1, 2016
Previous Article:The effects of cylinder deactivation on the thermal behaviour and performance of a three cylinder spark ignition engine.
Next Article:A study isolating the effect of bore-to-stroke ratio on gasoline engine combustion chamber development.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters