Effects of Pre-injection on Combustion Characteristics of a Single-cylinder Diesel Engine

Transcription

1 Proceedings of the ASME 2009 International Mechanical Engineering Congress & Exposition IMECE2009 November 13-19, Lake Buena Vista, Florida, USA IMECE IMECE Effects of Pre-injection on Combustion Characteristics of a Single-cylinder Diesel Engine M. MITTAL, G. ZHU, T. STUECKEN, and H.J. SCHOCK Automotive Research Experiment Station, Michigan State University, East Lansing, MI-48824, USA mittalma@msu.edu ABSTRACT Multiple injections used for diesel engines, especially preand post- injections, have the potential to reduce combustion noise and emissions with improved engine performance. This paper outlines the combustion characteristics of a singlecylinder diesel engine with multiple injections. The effects of pre-injection (multi-injection) on combustion characteristics are presented in a single-cylinder diesel engine at different engine speeds and load conditions. A common rail fuel system with a solenoid injector, driven by a peak and hold circuit, is used in this work. This enables us to control the number of injections, fuel injection timing and duration, and the fuel rail pressure that can be used to optimize the engine combustion process (e.g., eliminate engine knock). Mass fraction burned and burn durations are determined by analyzing the measured in-cylinder pressure data. Results are compared with the cases when no pre-injection was used, i.e. only main injection, at the same engine speeds and load conditions. In each study, different cases are considered with the variation in main injection timing. It is found that at full-load condition and lower engine speeds pre-injection is an effective method to alter the engine burn rate and hence to eliminate knock. 1. INTRODUCTION The fuel injection system is one of the most important engine sub-systems that affect the performance of diesel engines. In recent years, several studies have been performed to investigate the effects of fuel injection systems on diesel engine performance and emissions. Research has shown that by using the technique of fuel pre-injection, diesel engines have the potential to reduce the NO x levels and combustion noise, (Durnholz et al., 1994). The authors outlined the potential improvements in combustion and exhaust emission characteristics through the application of injection rate shaping and pilot fuel injection. They also discussed the potential of optimizing engine behavior through variation of injection system characteristics. Mallamo et al. (2002) studied the influence of different multiple injection strategies on the emissions, combustion noise and brake specific fuel consumption of a small non-road diesel engine prototype equipped with a common rail fuel injection system. They found that multiple injection strategies can be very effective in reducing particulate matter and noise levels without increasing fuel consumption. Carlucci et al. (2003) used a common rail injection system in their work to study the effects of pilot injection parameters, i.e. injection timing and duration, at different operating conditions in an in-line four-cylinder diesel engine. Park et al. (2004) presented the effects of multiple-fuel injection strategies on power improvement and emissions of a small diesel engine equipped with a high pressure common rail fuel injection system. They found that pilot-injection reduced the ignition delay of main injection and contributed to the improvement of power output by controlling the premixed combustion. The authors also showed in their study that postinjection is an effective tool in completing the oxidation process and reducing the particulate emissions. The objective of this paper is to investigate combustion characteristics in a single-cylinder diesel engine with multiple fuel injections. A common rail fuel injection system is used in this work. Pre-injection is shown to reduce the engine heat release rate by altering the mass fraction burned; therefore, it has the potential to eliminate engine knock. In the following sections, a brief description of the experimental setup is first outlined, followed by the results of various experiments 1

2 performed at different engine speeds and load conditions. Finally, concluding remarks are summarized from this work. 2. EXPERIMENTAL SETUP The engine used in this experiment is a single-cylinder diesel engine equipped with a common rail fuel system. The experimental setup is shown in Fig. 1. Specifications are listed in Table 1, with a bore diameter of 95 mm and a stroke length of 105 mm. The common rail fuel system, consisting of a Bosch fuel pump driven by an AC motor, was used to maintain the fuel rail pressure at 18,000 psi or 21,000 psi. The fuel rail pressure was controlled by a closed loop controller using both pressure relieve solenoids on fuel pump and rail. A Kistler pressure sensor was installed on the fuel rail for pressure feedback. A production Bosch fuel injector was used for all the experiments performed at different loads, 100 (full-load) and 70 percent (part-load) of IMEP, and engine speeds of 1200 and 2000 rpm. Specifications Measure Bore (mm) 95 Stroke (mm) 105 Compression ratio 18.0 Stroke/Bore ratio Connecting rod length (mm) 176 Table 1. Engine specifications In-cylinder pressure data were recorded with one degree of crank angle resolution. The Kistler pressure transducer (piezoelectric type) was used in this work with the measurement range varying from 0 to 250 bars. For each case 300 consecutive cycles were recorded; then, this in-cylinder pressure data was used to evaluate the engine performance. Mass fraction burned (MFB) and burn durations are determined using the Rassweiler and Withrow method (Rassweiler and Withrow, 1938). A linear model for the polytropic index during the combustion process is used to evaluate the pressure change due to the volume change (Mittal et al., 2009). MFB is a measure of the fraction of thermal energy released due to combustion of air-fuel mixture inside an engine cylinder, with respect to the total energy released at the end of combustion during a cycle. This also signifies how fast the chemical energy is released. The first derivative of MFB can be treated as the rate of heat release of the combustion process (Zhu et al., 2003). One may observe higher values of heat release rate when engine is operated with knock due to higher rate of change of mass fraction burned profile compared to the similar case without knock. The detection of knock and its classification has been the subject of discussion for many years (Burgdorf and Denbratt, 1997). Mittal et al. (2007) recorded in-cylinder pressure data with a resolution of 10 samples per crank angle degree. They used Fast Fourier Transform and band pass filtering approach to process the in-cylinder pressure data to determine the knock intensities for each cycle. It is to be noticed that the resolution of in-cylinder pressure data recorded in this work is only 1 sample per crank angle degree, which is insufficient to determine the knock frequencies. Therefore, this paper quantifies the effects of pre-injection on knock reduction by calculating the first derivative of MFB. Results are then compared with similar cases when no pre-injection was used. Fig. 1: Experimental rig 2

3 Due to the fact that derivative operation is very sensitive to noise, a two-way low pass filter is used to filter the in-cylinder pressure data (Zhu et al., 2003). Figure 2 shows the comparison between the experimental and filtered pressure data. advanced injection timings. It is a well known fact that in order to reduce engine NOx emissions; fuel injection timing has to be retarded close to TDC, which makes engine knock control difficult. Fig. 2: Experimental and filtered pressure data 3. EXPERIMENTAL RESULTS Combustion characteristics of a single-cylinder diesel engine without and with pre-injection are presented. Experiments are performed at full and part load conditions with engine speeds of 1200 and 2000 rpm. Figure 3 shows the averaged in-cylinder pressure data (upper graph), its mass fraction burned curve (middle graph) and the first derivative of MFB (lower graph) when no pre-injection was used, i.e. only main injection, at full-load condition with 1200 rpm engine speed. The averaged in-cylinder pressure data is the average of 300 consecutive cycles. To compare the different cases, we maintained a constant relative air-to-fuel ratio (λ), inverse of fuel-to-air equivalence ratio (φ), for each case and all the experiments are performed at rail pressure of 18,000 psi unless specified. As shown in the figure, the peak in-cylinder pressure is about 152 bars when main injection is at 15 crank angle degrees (CAD) before top dead center (BTDC), i.e. 15 BTDC case. The peak pressure reduces with the increase in main injection delay close to the top dead center (TDC) location, and it is about 87 bars for the case when main injection is at zero crank angle degree. As expected, with the advance of main injection timing the combustion process starts earlier; see MFB curves in the figure. The combustion starts at -10 and zero crank angle degrees for the cases when main injection is at 15 crank angle degrees BTDC (15 BTDC) and 0 BTDC, respectively. The burn duration is 35 CAD for 15 BTDC and reduces to 32 CAD when main injection is at TDC (0 BTDC case). The peak value of the MFB first derivative increases with the increase in main injection delay (with the exception of 3 BTDC case). This shows that the engine has higher knock tendency when main injection is close to TDC when compared to the cases with Fig. 3: Averaged in-cylinder pressure, MFB and its first derivative without pre-injection at full-load and 1200 rpm Figure 4 shows the averaged in-cylinder pressure, mass fraction burned and its first derivative when pre-injection was used at full-load condition with 1200 rpm engine speed. Compared to Fig. 3, the only difference in this test is that preinjection is used whereas the same total fuel quantity is maintained in both the tests. Pre-injection was initiated at 90 crank angle degrees before top dead center with pulse width of 200 µs. As shown in the figure, the peak in-cylinder pressure is about 165 bars when main injection is at -15 CAD (15 BTDC). As expected, the peak pressure reduces with further delay in main injection timing and it is about 98 bars for the case when 3

4 main injection is at TDC (0 BTDC). It is to be noticed that the peak values are significantly higher when compared to the similar cases without pre-injection. This is due to the higher incylinder temperature with pre-injection (see Park et al., 2004). As expected, when pre-injection is used the combustion process starts earlier and the burn duration increases when compared to the cases without pre-injection. The start of combustion takes place at -15 CAD and the burn duration is 39 CAD for 15 BTDC case. Note the burn duration of 35 CAD for the similar case without pre-injection. The burn durations are about 40 and 42 CADs for 6 BTDC and 0 BTDC cases, respectively. The peak values of heat release rate are lower with pre-injection when compared to similar cases without preinjection; see lower graph of Fig. 3 for comparison. The lower values clearly show that knock tendencies are reduced with pre-injection. This is due to the fact that by injecting a small amount of fuel before the main injection, it is possible to reduce the combustion noise by shortening the ignition delay of main injection in a diesel engine (Uekusa et al., 2005). Note that pre-injection is substantially different than simply advancing the main injection timing. The main difference is that advancing the main injection timing leads to high NOx emissions and increases engine combustion noise (knock) but pre-injection prepares gas mixture for the main injection that reduces the engine NOx emissions with low combustion noise (Cem et al., 2007). Also, one can observe an early small peak due to pre-injection in the first derivative of MFB. Figure 5 shows the effects of both pre-injection and main injection timing on mean of indicated mean effective pressure (IMEP), evaluated using 300 consecutive cycles. As shown in the figure, the mean IMEP increases from 0 to 6 BTDC and then decreases with the advance of main injection timing for both the tests with and without pre-injection (PI). The mean IMEP values are higher with pre-injection. This is due to the fact that in-cylinder temperature and pressure are increased by adoption of pre-injection. 1.36E E E+06 IMEP (Pa) 1.30E E E+06 without PI with PI 1.24E+06 Main Injection (BTDC) 1.22E Fig. 5: Mean IMEP at full-load and 1200 rpm engine speed, with and without pre-injection COV in IMEP without PI with PI 0.5 Fig. 4: Averaged in-cylinder pressure, MFB and its first derivative with pre-injection at full-load and 1200 rpm Main Injection (BTDC) Fig. 6: COV imep at full-load and 1200 rpm engine speed, with and without pre-injection 4

5 Figure 6 shows the effects of pre-injection on cyclic variability using coefficient of variation (COV) in IMEP, COV imep, Equation 1. This defines the cyclic variability in terms of indicated work per cycle (Heywood, 1988). It is evident that cyclic variability reduces with pre-injection. COV imep σ imep = x100 (1) IMEP timing at -17 CAD (17 BTDC) and -2 CAD (2 BTDC), respectively. The combustion process starts at -11 CAD, and the burn duration is 37 CAD for 17 BTDC. Burn duration decreases with the increase in main injection delay. It is 28 CAD when main injection starts at -2 CAD. The peak values of heat release rate increase when compared to similar cases at full-load condition (see Fig. 3 for comparison). It is as high as 0.1 for the case when main injection is started at -2 CAD. This shows that knock tendency increases at part-load condition. Fig. 7: Averaged in-cylinder pressure, MFB and its first derivative without pre-injection at part-load (70 % IMEP) and 1200 rpm Figure 7 shows the averaged in-cylinder pressure, mass fraction burned and the first derivative of MFB when no preinjection was used at part-load condition with 1200 rpm engine speed. Similar to the full-load condition, the peak in-cylinder pressure decreases with the increase in main injection delay and it is about 143 and 92 bars for the cases with main injection Fig. 8: Averaged in-cylinder pressure, MFB and its first derivative with pre-injection at part-load (70 % IMEP) and 1200 rpm Figure 8 shows the averaged in-cylinder pressure, mass fraction burned and its first derivative when pre-injection was used at part-load condition with 1200 rpm engine speed. Compared to Fig. 7, the only difference in this test is that preinjection is used whereas the same total fuel quantity is maintained in both the tests. In each case, pre-injection was 5

6 initiated at 90 crank angle degrees before TDC with pulse width of 150 µs. As shown in the figure, no significant difference can be observed in peak in-cylinder pressures when compared to the similar cases without pre-injection. The peak in-cylinder pressure is about 143 and 94 bars with main injection timing at -17 CAD and -2 CAD, respectively. Although the burn duration with pre-injection is slightly higher compared to the similar cases without pre-injection, no significant difference can be observed in heat release rate. This shows that the effects of pre-injection are not significant to reduce knock at part-load condition at 1200 rpm. Figure 9 shows the averaged in-cylinder pressure, mass fraction burned and the first derivative of MFB at full- and part-load conditions with 1200 rpm engine speed. The rail pressure used in each case is 21,000 psi. This figure shows the effects of pre-injection by comparing the similar cases when no pre-injection was used. As shown in the figure, the peak incylinder pressure is slightly higher and the burn duration increases with the pre-injection. However, no significant difference can be observed in the peak values of heat release rate. This shows that the effects of pre-injection, at higher rail pressure, are not significant to reduce the engine knock. Figure 10 shows the averaged in-cylinder pressure data, mass fraction burned and its first derivative at both full- (left graphs) and part- (right graphs) load conditions when engine is operated at 2000 rpm. The rail pressure used in each test is 18,000 psi. The effects of pre-injection are presented and compared with the similar cases when only main injection was used, i.e. no PI. As shown in the figure, peak in-cylinder pressure reduces with the increase in main injection delay; however, the peak values are only slightly higher with preinjection at both the load conditions. Also, no significant difference can be observed in mass fraction burned and its first derivative curves when pre-injection effects are compared with no PI cases. This shows that at higher engine speeds the effects of pre-injection are not significant even at full-load condition. 4. CONCLUSION An experimental study is performed to investigate the combustion characteristics of a single-cylinder diesel engine. Experiments are performed to study the effects of pre-injection at different loads and engine speeds. Mass fraction burned, burn duration and heat release rate are presented and compared with the similar cases when only main injection was used. Results show that at 1200 rpm engine speed, pre-injection is an effective tool to alter the mass fraction burned and hence to eliminate the knock at full-load condition. The effects of preinjection are not significant at part-load condition. Similarly, it is found that pre-injection effect reduces at higher rail pressure. At higher engine speed of 2000 rpm the effects of pre-injection are even not significant at full-load condition for the cases studied. Therefore we can conclude that pre-injection effects are more significant at full-load condition and lower engine speeds to reduce the engine knock. ACKNOWLEDGMENTS This project was partially supported by the US Army under a SBIR grant to Mid Michigan Research, Okemos, Michigan. REFERENCES [1] Durnholz M., Endres H., and Frisse P., 1994, Preinjection a measure to optimize the emission behavior of DI-diesel engine, SAE Paper No [2] Mallamo F., Badami M., and Millo F., 2002, Analysis of multiple injection strategies for the reduction of emissions, noise and BSFC of a DI CR small displacement non-road diesel engine, SAE Paper No [3] Carlucci P., Ficarella A., and Laforgia D., 2003, Effects of pilot injection parameters on combustion for common rail diesel engines, SAE Paper No Fig. 9: Averaged in-cylinder pressure, MFB and its first derivative at full- and part-loads with 1200 rpm; rail pressure 21,000 psi [4] Park C., Sanghoon K., and Bae C., 2004, Effects of multiple injections in a HSDI diesel engine equipped with common rail injection system, SAE Paper No

7 Fig. 10: Averaged in-cylinder pressure, MFB and its first derivative at 2000 rpm, (a) full-load, and (b) part-load [5] Rassweiler G.M., and Withrow L., 1938, Motion Pictures of Engine Flames Correlated with Pressure Cards, SAE Transactions, Vol 42. [6] Mittal M., Zhu G., and Schock H.J., 2009, Fast mass fraction burned calculation using net pressure method for real-time applications, Proc. IMechE, Part D: J. Automobile Engineering, 2009, 223 (3), [7] Zhu G., Daniels C.F., and Winkelman J., 2003, MBT timing detection and its closed-loop control using in-cylinder pressure signal, SAE Paper [8] Burgdorf K., and Denbratt I., 1997, Comparison of cylinder pressure based knock detection methods, SAE Paper No [9] Mittal V., Bridget M.R., and Heywood J.B., 2007, Phenomena that determine knock onset in spark-ignition engines, SAE Paper [10] Uekusa T., Nakada T., Ishikawa N., Ueda T., Fujino R., Brown D.B., Paratore M.J., and Ryan D.M., 2005, Emission reduction study for meeting new requirements with advanced diesel engine technology, SAE Paper [11] Cem S., Abdurrahman I., Metin E., Dogan S., and Ergun G., 2007, Control of NOx emissions from diesel engines by the optimisation of fuel injection strategies, International Journal of Vehicle Design, 45 (1-2), [12] Heywood J.B., 1988, Internal combustion engine fundamentals, McGraw Hill Book Company, ISBN X. NOMENCLATURE Symbols φ: Fuel-to-air equivalence ratio λ: Relative air-to-fuel ratio, inverse of φ Abbreviations 15 BTDC: Main injection timing at -15 CAD BTDC: Before top dead center CAD: Crank angle degree COV: Coefficient of variation IMEP: Indicated mean effective pressure MFB: Mass fraction burned PI: Pre-injection RPM: Revolutions per minute TDC: Top dead center 7