Effect of Ambient Temperature and Humidity on Combustion and Emissions of a Spark-Assisted Compression Ignition Engine

Increasing fuel economy and emissions requirements motivate the development of advanced combustion strategies. Using dilute mixtures and elevated compression ratios, homogeneous charge compression ignition (HCCI) engines benefit from reduced throttling losses, higher ratio of specific heats, and reduced combustion temperatures and can thus achieve better efficiencies and lower engine-out emissions than conventional spark-ignited gasoline engines [1,2]. Despite these benefits, the timing of the auto-ignition event and therefore the combustion phasing are more difficult to control in an HCCI engine. The operating range is also more limited, constrained at low loads by bulk gas quenching and at high loads by in-cylinder pressure limits [3].

Spark-assisted compression ignition (SACI) offers a more practical and direct method of controlling combustion phasing and provides a means of extending the high load limit of HCCI while maintaining high thermal efficiency [4]. SACI combustion first consumes a portion of charge by a spark-ignited flame; then, an auto-ignition event consumes the rest of the charge. SACI bridges the gap between dilute, unthrottled HCCI combustion, and stoichiometric part-throttle spark ignition (SI) operation [5]. Also, SACI offers the potential for smooth combustion mode transitions during transient engine operation due to the hybrid nature of SACI [6].

Several studies have investigated the effects of temperature and humidity on SI engine combustion and emissions [7–13] and even their effect on aggregate emissions at the traffic level [14]. The effects of temperature and humidity on HCCI engine combustion and emission have also been investigated [15–18]. The impact of ambient conditions on SACI combustion, however, has not been thoroughly investigated. An exception is a recent study by Mendrea et al. [19], which investigates the effects of ambient air temperature and humidity on crank angle of 50% mass fraction burned (CA50) and the allowable range of CA50 in a production-feasible SACI engine.

This study extends the SACI ambient condition investigations described in Ref. [19] by performing additional engine dynamometer tests for investigating how temperature and humidity affect rate of heat release profiles, fraction of flame propagation versus auto-ignition, exhaust gas recirculation (EGR), residual gas fraction, and engine-out CO, NO_x, and hydrocarbon emissions from stoichiometric SACI combustion.

Experiments were performed using a 2.0 L inline four-cylinder engine based on General Motors Ecotec LNF engine that was modified to enable multimode operation including HCCI, SACI, and SI. Key engine specifications are provided in Table 1. The engine used a custom cylinder head, pistons, and a central-mounted (spray guided) direct injection (DI) fuel injection system using Bosch HDEV5 six-hole high-pressure fuel injectors. The valvetrain package includes dual independent variable valve timing (VVT) controlled by DENSO electric camshaft phasers in conjunction with a two-step Delphi variable valve lift (VVL) mechanism. A high-pressure EGR system was used to move exhaust gas through an EGR cooler and EGR valve to the intake manifold. The cooler maintained EGR temperature at approximately 100 °C using engine coolant.

Engine control was via a Bosch MED17.3.2 ECU and ETAS ES910.3 prototyping interface. An AVL AC dynamometer system was used to maintain engine speed and measure engine load. Cylinder pressure was measured using Kistler 6125 C pressure transducers. Engine-out lambda was measured with a Bosch LSU4.9 wide range oxygen sensor and ETAS LA4 lambda module.

The fuel utilized in this study was Chevron Phillips UTG96 gasoline fuel. Selected key fuel specifications, taken from the Chevron Phillips certificate of analysis, are listed in Table 2.

EGR was calculated from the molar ratio of intake and exhaust gas CO₂ concentrations. An AVL SESAM i60 emissions bench was used to measure intake and exhaust concentrations. The bench uses an NDIR for intake CO₂ concentration and an FTIR for exhaust CO₂, CO, NO_x, and individual hydrocarbon species. An AVL combustion air system controlled intake temperature and humidity.

Four ambient conditions, points A, B, C, and D, are chosen to investigate the effects of ambient temperature and humidity (Table 3). Points A–C have very similar absolute humidity and different temperatures, while points B and D have the same temperature and different absolute humidities.

Air from the combustion air unit passes through the turbocharger compressor and intercooler and then enters the engine's intake manifold. Temperature and relative humidity after the combustion air unit, temperature after the intercooler, dew point, and absolute humidity (which are the same at both locations) are shown in Table 3. In the figures and discussion that follow, temperature after the intercooler (°C) and absolute humidity (g/m₃) are the selected reference variables (e.g., the variables shown on the horizontal axis of figures) and are referred to as “temperature” and “humidity.”

The engine is operated in negative valve overlap (NVO) mode to enable SACI combustion. The SACI operating range for this engine is from 1400 rpm to 3600 rpm and from 2 bar to 6 bar brake mean effective pressure (BMEP). The operating condition selected for this study is 2000 rpm and 4 bar, which is in the middle of the SACI speed and load range for this engine and is a representative point for SACI combustion [19]. While running the various ambient conditions, the engine's fueling rate and throttle position were fixed, thus lambda and BMEP vary slightly between points. Other engine actuator settings are also held constant. The settings of actuators are listed in Table 4. Error bars shown in following figures represent the repeatability of measurement of 300 sequential cycles.

Rate of heat release analysis and residual gas estimation is performed using three pressure analysis (TPA) [20]. In this technique, measured instantaneous intake, exhaust, and cylinder pressures are used in conjunction with measured cylinder head port flow data and intake and exhaust valve profiles to determine the contents of the cylinder (air flow, residual content, and fuel quantity) at the start of combustion. The Woschni heat transfer correlation is used [21]. By calculating the maximum of the second derivative of the rate of heat release, the auto-ignition point is identified [22]; and thus, the burned fraction of flame propagation is calculated. This method is deemed sufficient for obtaining trendwise estimates of auto-ignition timing for SACI combustion where there is significant flame propagation and auto-ignition [23].

To investigate the effect of ambient temperature on SACI combustion, rate of heat release analysis is performed for conditions A–C, which use the same actuator settings and the same absolute humidity. As shown in Fig. 1, as ambient temperature increases, the maximum rate of heat release increases significantly and occurs earlier. From the mass fraction burned profiles, which are shown in Fig. 2, it can be seen that burn duration decreases and auto-ignition timing occurs earlier as ambient temperature is increased.

Combustion phasing, as described by CA50, is shown in Fig. 3. As temperature is increased from 20 °C to 41 °C, CA50 is advanced by 5.6 crank angle degrees. Also, it can be seen from Fig. 3 that as temperature increases, a smaller fraction of the fuel is burned by flame propagation and a larger fraction through auto-ignition. The earlier and faster combustion increases maximum pressure rise rate (Fig. 4). Higher temperature leads to higher charge temperature during compression, which promotes the auto-ignition process. According to the results calculated from TPA method, the in-cylinder charge temperature at IVC increases from 456 K to 468 K when the ambient condition changes from condition A to condition C. The air density decreases as temperature increases, which leads to the reduction of inducted air mass at a fixed throttle position. Less air and the same fuel shift lambda slightly rich (lambda decreases from 1.008 to 0.988 as temperature increases from 20 to 41 °C) and causes a slight decrease in BMEP as shown in Fig. 4.

Due to the higher fraction of heat release by the auto-ignition process, the maximum pressure rise rate increases from 2.6 to 5.4 bar/CAD as temperature increases (Fig. 4). Figure 5 shows that total EGR increases slightly as temperature is increased. The increase is driven by increased internal EGR. Earlier combustion phasing reduces cylinder pressure and increases cylinder density during the exhaust stoke, which causes less exhaust mass to be pushed out before exhaust valve closing (EVC) and more internal EGR to be trapped.

Increased fraction of auto-ignition tends to reduce COV, while increased total EGR tends to increase COV. Figure 6 shows the net effect of this, in which COV first decreases and then increases with increasing temperature. As temperature is increased, BSFC is improved by more advanced combustion phasing, but it is deteriorated by lower BMEP (from 3.98 bar to 3.83 bar) and slightly richer mixture. The overall effect is a slight increase from 284.9 g/kWh to 290.1 g/kWh in BSFC (Fig. 6). In a production application, lambda would remain stoichiometric by closed-loop control and BMEP would remain fixed by driver command, so in this scenario BSFC would most likely not increase.

In terms of emissions, CO increases and then decreases as temperature is increased, as shown in Fig. 7. Advanced combustion phasing increases peak charge temperature and the dissociation of CO₂ into CO; slightly richer operation also increases CO. But increasing the fraction of fuel burned through auto-ignition likely leads to more complete combustion and lower CO, consistent with the reduction in CoV of IMEP.

Higher peak charge temperature leads to higher NO_x, which is illustrated in Fig. 7. NO increases significantly as charge temperature increases while NO₂ stays essentially constant (Fig. 8).

Various exhaust hydrocarbon species were measured as temperature was increased: CH₄, C₂H₂, C₂H₄, C₂H₆, C₃H₆, and C₄H₆. From Fig. 9, it can be observed that each hydrocarbon only varies slightly as temperature is increased, indicating that temperature and auto-ignition fraction do not have an appreciable effect on combustion efficiency for these conditions.

To investigate the effect of humidity on SACI combustion, rate of heat release curves and mass burned profiles of operating points with the same temperature and different humidities are calculated using TPA and shown in Figs. 10 and 11. Higher humidity decreases the rate of specific heats and increases the heat capacity of the charge, which results in lower compression and combustion temperatures, respectively. This delays the auto-ignition event, thereby lowering the rate of heat release and lengthening combustion duration.

Engine performance characteristics from the humidity tests are shown in Table 5. Higher humidity causes slower flame propagation and delayed auto-ignition, which retards CA50 by 13 crank angle degrees. Higher humidity causes a lower fraction of flame propagation and a higher fraction of auto-ignition combustion. Also, these factors lead to less stable combustion, resulting in higher COV of IMEP and reduced maximum rate of pressure rise. The reduction of inducted air mass resulting from the displacement of intake air by water vapor results in a slight reduction of BMEP and a slightly richer mixture. Due to the delay of combustion phasing, reduction of BMEP, and a slightly richer mixture, BSFC increases. Internal and external EGR rates decrease. External EGR decreases slightly due to a slight increase of intake manifold pressure caused by higher water concentration, and due to hotter and less dense exhaust gas. Later combustion phasing increases cylinder temperature during expansion and exhaust, which leads to less internal EGR trapped in NVO duration.

In terms of emissions, dissociation of CO₂ is reduced due to lower combustion temperature, which leads to less CO as shown in Fig. 12. Also, less NO_x is created due to lower maximum combustion temperature. However, NO₂ rises with the increase of humidity (Fig. 13). CH₄ and C₂H₂ decrease slightly, while C₂H₄, C₂H₆, C₃H₆, and C₄H₆ increase slightly, as shown in Fig. 14.

To compare how ambient conditions affect SACI, HCCI, and SI combustion, SACI data obtained from this research are compared with HCCI and SI results from three other studies [13,15,17]. Even though operating conditions and engine configurations are different, the comparison provides insight into the relative significance of ambient condition on various combustion modes.

Data shown in Fig. 15 indicate that the effect of temperature on combustion phasing is stronger on SACI combustion than on HCCI combustion. Similar to that, the effect of humidity on SACI is also more significant than on HCCI, which is shown in Fig. 16. The higher sensitivity of SACI combustion is believed to result from its combustion phasing, as described in the following paragraphs.

SACI tests in this study have CA50 at 10–16 deg aTDC. Hotter temperature and lower humidity tend to speed up flame propagation combustion and auto-ignition combustion, which advance combustion phasing and bring combustion closer to TDC. Combustion that takes place close to TDC occurs at higher temperature and pressure due to compression from the piston, which further speeds up combustion and advances it even more. So, there is the direct effect of higher temperature, and the indirect effect of combustion occurring closer to TDC.

HCCI tests performed by Andreae et al. [15] have CA50 at 0–6 deg ATDC, which is more advanced than in the SACI tests. Hotter temperature and lower humidity speed up combustion, which advances combustion phasing. But in these HCCI test cases with earlier combustion phasing, an advancement of the start of combustion, which occurs before TDC, away from TDC, moves the start of combustion away from TDC, to a point where the piston has not yet completed its full compression of the charge. The middle and later parts of combustion move closer to TDC, but the amount of compression done by the piston at TDC versus a few degrees after TDC is not very different. Compression temperature has a parabolic shape in an engine with a crank slider mechanism and temperature is relatively flat near TDC. Thus, for the HCCI cases, higher temperature advances combustion, and there is no strong indirect effect as there is in the SACI cases.

In terms of emissions, Figs. 17–19 show how brake-specific CO, NO_x, and HC emissions from HCCI [17] and SACI and SI [13] combustion are affected by temperature. BSCO emissions from HCCI are lower than from the other combustion modes, possibly because the HCCI cases are run at a lean condition that tends to reduce BSCO. For the HCCI and SI cases, higher temperature advances combustion, increases peak combustion temperature, and decreases BSCO. For the SACI cases, higher temperature initially increases BSCO and then slightly decreases BSCO. The increase may be a result of increased dissociation of CO₂ into CO and a slightly richer mixture, while the decrease may be due to more auto-ignition promoting more complete combustion.

SI has higher BSNO_x than SACI and HCCI because the SI cases have less dilution (from burned gases or air) and therefore have much higher combustion temperatures. As temperature increases, BSNO_x increases at a similar rate for HCCI, SACI, and SI combustion.

Higher temperature tends to reduce BSHC for all the three combustion modes, but the lower peak combustion temperature of HCCI causes its BSHC emissions to be much greater than from SACI and SI.

This study investigates the effects of ambient temperature and humidity on combustion and emissions from a stoichiometric SACI engine. The effects of temperature and humidity on combustion phasing, maximum pressure rise rate, fraction of auto-ignition combustion versus flame propagation, COV of IMEP, BSFC, and CO, NO_x, and hydrocarbon emissions are analyzed.

At the medium speed and load SACI test point selected, increasing air temperature from 20 °C to 41 °C advances combustion phasing, increases maximum pressure rise rate, causes a larger fraction of the charge to be consumed by auto-ignition (and a smaller fraction by flame propagation), and increases BSNO_x. Increasing humidity from 32% to 60% retards combustion phasing, reduces maximum pressure rise rate, increases COV of IMEP, reduces BSNO_x, and increases BSFC.

A comparison of the SACI results from this study to HCCI results from another study shows that the effects of temperature and humidity on CA50 are stronger for SACI as compared to HCCI, likely due to different combustion phasing used by the two combustion modes.

Higher temperature increases BSNO_x from SACI, HCCI, and SI. The level of BSNO_x from SACI is similar to HCCI and much lower than SI. Temperature decreases BSHC from SACI, HCCI, and SI. The level of BSHC for SACI is similar to SI and much lower than HCCI. Increasing temperature initially increases and then decreases BSCO from SACI. This differs from HCCI and SI, where BSCO decreases with temperature.

These data indicate that successful implementation of SACI combustion in passenger cars requires a combustion control strategy that considers ambient temperature and humidity.

This material is based upon the work supported by the Department of Energy (National Energy Technology Laboratory) under Award No. DE-EE0003533. This work was performed as a part of the ACCESS project consortium (Robert Bosch LLC, AVL, Inc., Emitec, Inc., Stanford University, and University of Michigan) under the direction of PI Hakan Yilmaz and Co-PI Li Jiang and Oliver Miersch-Wiemers, Robert Bosch LLC.

The authors sincerely thank Tom Marino (Roush, Inc.) for his help with engine and test cell setup, Patrick Gorzelic (University of Michigan), Yusuf Zeynel Abidin Akkus (Robert Bosch LLC), George Lavoie (University of Michigan), and Robert Middleton (University of Michigan) for their helpful discussion, and Michael Mosburger (Robert Bosch LLC) for his support of the study.

This report was prepared as an account of work sponsored by an agency of the U.S. Government. Neither the U.S. Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the U.S. Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the U.S. Government or any agency thereof.

BMEP =

brake mean effective pressure

BSFC =

brake-specific fuel consumption

CAD =

crank angle degree

CA50 =

crank angle of 50% mass fraction burned

CI =

compression ignition

COV =

coefficient of variation of IMEP

DI =

direct injection

EGR =

exhaust gas recirculation

EVC =

exhaust valve closing

GDI =

gasoline direct injection

HCCI =

homogeneous charge compression ignition

IMEP =

indicated mean effective pressure

NVO =

negative valve overlap

SACI =

spark-assisted compression ignition

SI =

spark ignition

SOI =

start of injection timing

TDC =

top dead center

TPA =

three pressure analysis

TWC =

three-way catalytic converter

VVL =

variable valve lift

VVT =

variable valve timing