The Effects of the Compression Ratio, Equivalence Ratio, and Intake Air Temperature on Ignition Timing in an HCCI Engine Using DME Fuel

Transcription

1 The Effects of the Compression Ratio, Equivalence Ratio, and Intake Air Temperature on Ignition Timing in an HCCI Engine Using DME Fuel (1) (1) (1) (1) (1) Keisuke HAMADA, Shun NIIJIMA, Kazunori YOSHIDA, Koji YOSHIDA, Hideo SHOJI, () (3) Kazuto SHIMADA, Kenji SHIBANO (1) () (3) Nihon University, Fuji Heavy Industries Ltd., SUBARU TECNICA INTERNATIONAL Copyright 5 Society of Automotive Engineers of Japan, Inc. and Copyright 5 SAE International Attention has recently been focused on homogeneous charge compression ignition combustion (HCCI) as an effective combustion process for resolving the essential nature of combustion. Meanwhile, dimethylether (DME) has attracted interest as a potential alternative fuel for compression ignition engines. Authors measured the combustion process of DME HCCI by using a spectroscopic method. A diesel engine was used as the test engine. The results of these analyses showed that changes in the compression ratio, intake air temperature and equivalence ratio influenced the ignition timing in the HCCI combustion process. This paper discusses these effects in reference to the experimental and calculated results. Keywords: Engine Combustion, Compression Ignition Engine, Flame / HCCI Engine, Dimethyl Ether 1. INTRODUCTION The depletion of energy sources has become an issue of global concern in recent years that cannot be ignored. While diesel engines achieve good fuel efficiency, it is difficult to reduce their nitrogen oxide (NOx) and suspended particulate matter (SPM) emissions simultaneously owing to the trade-off relationship between these pollutants with respect to the combustion gas temperature. Homogenous charge compression ignition (HCCI) engines have attracted attention as internal combustion (IC) engines potentially capable of resolving these energy and environmental protection issues. However, HCCI engines currently have a narrow range of stable compression ignition operation, and it is difficult to control the ignition timing in this combustion process. Dimethyl ether (DME) has drawn interest as a potential fuel for HCCI engines. (1) This fuel has a negative temperature coefficient region and characteristically displays a multistage heat release pattern owing to low- and high-temperature oxidation reactions. ()-(5) In this study, authors focused on elucidating the conditions of HCCI combustion based on measurement and analysis of radical behavior. Our aim was to use the resulting knowledge in an effort to expand the compression ignition region of HCCI combustion and to control the ignition timing. A spectroscopic method () was used to measure radical light emission behavior at wavelengths corresponding to formaldehyde (HCHO; characteristic spectrum of 395. nm) and the OH radical (3. nm), (7)-(11) which are characteristic intermediate products of compression ignition combustion. The results indicated that changes in the engine compression ratio, intake air temperature and equivalence ratio affected the ignition timing. This paper presents the experimental data along with the results of numerical calculations made for elementary reactions.. EXPERIMENTAL EQUIPMENT AND METHOD.1 EXPERIMENTAL METHOD FOR HCCI CONDITIONS The test engine used in this study was an air-cooled single-cylinder diesel engine, having the specifications given in Table 1. The configuration of the test equipment used is shown in Fig. 1. A crystal pressure transducer was installed in the top of the cylinder head to measure the cylinder pressure. A K-type sheath thermocouple was attached mm above the bottom of the cylinder head, and the temperature measured in that area was defined as the combustion chamber wall temperature. The intake air temperature and exhaust gas temperature were also measured using the same kind of K-type sheath thermocouple. A quartz observation window holder was attached to the top of the cylinder head for sampling the flame light in the combustion chamber. The flame light extracted through the quartz observation window holder was introduced into a polychromator SETC 5 1

2 Table1 Specifications of Test Equipment -Cycle Air-cooled Diesel Engine Number of Cylinders 1 Bore Stroke 7 mm Displacement 99 cm 3 Compression Ratio (Compression ratio change) 19:1, 1:1, 17:1, 1:1 15:1 1:1, 13:1, 1:1,11:1 Compression Ratio (Intake Air Heating) 1:1 Valve Arrangement OHV Type of Piston Reentrant Quartz Observation Window Cylinder center Intake Air Heat Release Rate [J/deg.] Pressure Transducer Dynamometer Personal Computer Laminar Flow Meter Surge Tank Q H max LTR Q L max. θ LS Engine Quartz Observation Window Optical Fiber Photomultiplier Crank Angle Pickup Low-Pass Filter θ LS : Onset time of LTR Q Lmax : Peak heat release of LTR θ HS : Onset time of HTR Q Hmax : Peak heat release of HTR L: Duration of LTR S: Amount of heat generated by LTR Crank Angle, θ [deg.] Fig. 3 Definitions of LTR and HTR Cylinder center K type sheath thermocouple Polychromator A-D Converter Fig. 1 Configuration of Test Equipment Fig. Configuration of Fuel Supply System L S θ HS Intake Air Heaters ( kw) HTR Control Unit Premixer Mass Flow Controler Throttle DME Gas Cylinder Cylinder Head Exhaust Gas via an optical fiber cable having a core diameter of 1 mm and separated into wavelengths of 395. nm and 3. nm. The wavelength resolution of the polychromator was. nm in terms of its half-width value. Following separation, the light at each wavelength was fed into a photomultiplier, and the respective output voltage was regarded as the light emission intensity. The test fuel used in the HCCI combustion experiments was DME, which was introduced into a premixer for formation of the premixed mixture. The premixer was installed approximately mm upstream of the intake valve. The nozzle provided in the center of the premixer was oriented in the opposite direction of the air flow so as to diffuse the fuel (1). The configuration of the fuel supply system is shown schematically in Fig.. A spacer was inserted between the cylinder and the crank case to vary the engine compression ratio ε while running the engine at a constant speed of N = 1 rpm. The compression ratio was varied in increments of one over a range of ε = 19~1:1. In the experiments, the equivalence ratio φ was varied in a range that allowed the engine to be operated at these compression ratios. The ignition timing and other parameters used in analyzing the heat release rate are defined in Fig. 3. The intake air temperature was varied over three levels of 3, 93 and 33 K. When the intake air was heated, the engine was operated at a constant speed (N = 1 rpm) and compression ratio (ε = 1:1), and the equivalence ratio φ was varied in a range of.3~.3. To heat the intake air, four 5-watt heaters (total of kw) were installed in the surge tank in the area indicated by the dashed lines in Fig.. The air temperature was controlled by using a voltage regulator to control the four heaters. The intake pipe was covered with insulation to prevent heat loss in the section from the surge tank to the intake port of the engine. A throttle was provided ahead of the intake port to maintain the charging efficiency at a constant level.. CHEMICAL KINETICS The oxidation reactions of DME were analyzed numerically, and the results were compared with the data measured experimentally with the test engine. The IC Engine software in Chemkin. was used to perform the calculations in the numerical analysis of the elementary reactions. The scheme used was the elementary reaction model proposed by Curran et al., (13) involving 33 elementary reactions and 7 chemical species. The operating conditions used in the calculations were the same as those used in the experiments. 3. EXPERIMENTAL RESULTS 3.1 RESULTS FOR VARYING EQUIVALENCE RATIOS The conditions for varying the equivalence ratio were set within a range where each compression ratio yielded sufficient torque and strong knocking noise did not occur. Figure shows typical waveforms measured for engine operation under DME-HCCI combustion at a constant engine speed N of 1 SETC 5

3 Light Emission Intensity T G [K] E HCHO P [MPa] rpm, compression ratio ε of 1:1 and a varying equivalence ratio φ of.,.3 and.3. The crank angle θ (deg.) is indicated along the horizontal axis. From the top, the figure shows along the vertical axis the cylinder pressure P (MPa), heat release rate (HRR, J/deg.) found geometrically from an indicator diagram, mean gas temperature T G (K) in the cylinder found with a state equation from an indicator diagram, and the light emission intensity E OH (V) at 3. nm, corresponding to the OH radical, and the light emission intensity E HCHO (V) at 395. nm, corresponding to HCHO, at the observation point in the combustion chamber. The equivalence ratio φ, charging Pressure H.R.R. 3 1 Gas Temperature.3 3. [nm]. (OH radical) [nm] (HCHO). H.R.R. [J/deg.] E OH Crank Angle, [deg.] T EX [K] T W [K] 1: HCCI : HCCI : HCCI Heat Release Rate [J/deg.] Fig. Measured Waveforms (varying equivalence ratio, φ) - TDC Crank Angle,[deg.] Combined Integrated 3rd stage Fig. 5 Measured Waveforms varying equivalence ratio, φ LOWHIGH efficiency η c (%), exhaust gas temperature T EX (K) and combustion chamber wall temperature T W (K) are shown below the crank angle as indicators of the engine operating state in this experiment. The HRR waveforms mainly show a pattern of two-stage heat release. This pattern of multistage heat release is characteristic of compression ignition combustion using DME. The first stage and second stages of heat release are indicative of low-temperature and high-temperature oxidation reactions, respectively. It is said that compression ignition reactions with DME theoretically show a pattern of three-stage heat release, consisting of first-stage low-temperature oxidation reactions that produce formaldehyde, second-stage high-temperature oxidation reactions that produce the CO radical and third-stage high-temperature oxidation reactions that produce CO. However, it can be observed that the second- and third-stage reactions are integrated into a single stage as a result of increasing the equivalence ratio. Figure 5 shows an example of experimental results in which that integrated pattern is clearly evident. Looking at the pressure and in-cylinder gas temperature waveforms in Fig., it is seen that the pressure and temperature at the onset of the low- and high-temperature oxidation reactions in each experiment were nearly equal. It is known that the lowand high-temperature oxidation reactions of HCCI combustion are initiated uniformly according to the temperature and pressure fields. The HRR waveforms indicate that the onset of the high-temperature oxidation reactions advanced to an earlier crank angle as the equivalence ratio was increased, though the onset time of the low-temperature oxidation reactions did not change. Low-temperature oxidation reactions begin when the in-cylinder gas reaches the necessary temperature for their initiation as a result of being heated due to compression by the piston and the transfer of heat from the combustion chamber walls, among other factors. It is therefore assumed that the temperature rise is not influenced by the equivalence ratio before the gas temperature reaches the level needed for the onset of the low-temperature oxidation reactions. Once the low-temperature oxidation reactions begin, they generate heat that raises the in-cylinder gas temperature and leads to the onset of the high-temperature oxidation reactions. Therefore, a higher equivalence ratio presumably promotes the low-temperature oxidation reactions, thereby quickening the progression to the high-temperature oxidation reactions. It is seen in Fig. that the peak temperature of the high-temperature oxidation reactions rose as the equivalence ratio was increased. That is attributed to greater heat release due to the presence of a higher density of fuel molecules per unit volume as the equivalence ratio was increased. The light emission intensity waveform of each intermediate product of combustion shows that the interval of light emission coincided with the period of high-temperature oxidation reactions in the HRR waveform. This indicates that the SETC 5 3

4 light emission behavior of the high-temperature oxidation reactions was reliably detected. It is seen from each light emission intensity waveform that the onset of the increase in the light emission intensity shifted to an earlier crank angle with an increase in the equivalence ratio φ. That is attributed to an earlier onset of the high-temperature oxidation reactions as the equivalence ratio φ was increased. In other words, it can be inferred that an earlier increase in the light emission intensity of the combustion intermediates occurs incidental to a quicker start of the high-temperature oxidation reactions, owing to the fact these reactions are intense processes accompanied by the emission of light. The higher heat release rate seen in the interval of the low-temperature oxidation reactions implies the passage of a low-temperature flame (cool flame), though no corresponding behavior is observed in the light emission intensity waveform. The extremely faint light emission of low-temperature flames would account for that tendency. (1) 3. RESULTS FOR VARYING COMPRESSION RATIOS Figure shows typical waveforms measured for engine operation under DME-HCCI combustion at an engine speed N of 1 rpm, equivalence ratio φ of. and a varying compression ratio ε of 19, 17 and 15:1. Shown below the crank angle are the charging efficiency ηc (%), exhaust gas temperature T EX (K) and combustion chamber wall temperature T W (K) as indicators of the engine operating state in this experiment. The HRR waveforms reveal that the onset of the low-temperature oxidation reactions shifted to a later crank angle accompanying a lower compression ratio (15). As the compression ratio is reduced, the compression pressure at the same crank angle is lower compared with that for a higher compression ratio. Accordingly, it takes a longer crank angle interval for the in-cylinder gas to reach the temperature level for the onset of the low-temperature oxidation reactions, which explains the shift to a later crank angle. The temperature rise originating from the low-temperature oxidation reactions is also delayed to the extent that the onset of these reactions shifts to a later crank angle, which means the start of the high-temperature oxidation reactions shifts to a later crank angle as a result. It is seen that the peak temperature of the high-temperature oxidation reactions was nearly the same for all three compression ratios. That can be explained by the fact that the heat release from the hot flames was virtually constant because the equivalence ratio was kept constant in this experiment. The light emission intensity waveforms of the combustion intermediates show that the onset of a sharp increase in light emission intensity shifted to a later crank angle as the compression ratio was reduced. This tendency can be attributed to the fact that an increase in the light emission intensity of intermediate products occurs incidental to the start of the high-temperature oxidation reactions. Light Emission Intensity P [MPa] T G [K] E HCHO RESULTS FOR VARYING INTAKE AIR TEMPERATURES Figure 7 shows typical waveforms measured for engine operation under DME-HCCI combustion at an engine speed N of 1 rpm, equivalence ratio φ of.3 and a varying intake air temperature T IN of 3, 93 and 33 K. The intake air temperature T IN (K), exhaust gas temperature T EX (K) and combustion chamber wall temperature T W (K) are shown below the crank angle as indicators of the engine operating state in this experiment. It is seen in the figure that onset of the low-temperature oxidation reactions shifted to an earlier crank angle accompanying a higher intake air temperature. Presumably, the in-cylinder gas reached the temperature needed for the initiation of the low-temperature oxidation reactions more quickly because a higher intake air temperature raised the initial gas temperature at the onset of the reactions. The temperature rise originating from the low-temperature oxidation reactions was also quickened to the extent that the onset of these reactions shifted to an earlier crank angle, which had the consequent effect of hastening the start of the high-temperature oxidation reactions. The light emission intensity waveforms of the combustion intermediates show that the light emission intensity began to rise sharply at an earlier crank angle as the intake air temperature was increased. An earlier rise in the light emission intensity of the intermediate products occurs incidental to Pressure H.R.R Gas Temperature 3. nm OH radical 395. nm HCHO Crank Angle, [deg.] H.R.R [J/deg.] T EX [K] T W [K] 1: HCCI : HCCI : HCCI Fig. Measured Waveforms E OH SETC 5

5 P [MPa] T G [K] Light Emission Intensity E HCHO aquicker start of the high-temperature oxidation reactions, which accounts for this light emission intensity behavior. The increase in the heat release rate during the interval of the low-temperature oxidation reactions is presumed to indicate the passage of a low-temperature flame (cool flame). However, no corresponding behavior is observed in the light emission intensity waveform under these engine operating parameters either because of the extremely faint light emission of low-temperature flames under these experimental conditions. 3.3 RESULTS FOR LOW COMPRESSION RATIO It is believed to be difficult to detect light emission from cool flames (low-temperature oxidation reactions) in HCCI combustion experiments conducted with DME as the test fuel. However, light emission presumably attributable to the passage of a cool flame was observed in the present study under experimental conditions of a low compression ratio (ε = 1:1) and a high equivalence ratio (φ =.5). The results measured for these conditions are shown in Fig.. Evidence of faint light emission (A in the figure) can be observed in the light emission intensity waveform for HCHO (395. nm) just before the emission intensity rises sharply. That period coincides with the manifestation of low-temperature oxidation reactions in the HRR waveform. On the other hand, there is no sign of light emission in the waveform of the OH radical. Accordingly, it is assumed that this detected light emission was from the low-temperature oxidation reactions. Pressure H.R.R. Gas Temperature - TDC Crank Angle, [deg.] 3. nm (OH radical) 395. nm (HCHO) T IN [K] T EX [K] T W [K] 1: HCCI : HCCI : HCCI Fig. 7 Measured Waveforms varying intake air temperature, T IN H.R.R. [J/deg.] E OH Mole Fraction T [K] P [MPa] G P [MPa] Light Emission Intensity E OH First Stage Heat Release 1, 1, A Temperature E-3 1E- 1E-5 1E- HCHO 1E-7 CO OH 1E- CO 1E-9 1E TDC 1 Crank Angle, [deg.]. Crank Angle, [deg.] Fig. Measured Waveforms (N = 1 rpm, φ=.5, ε= 1) Pressure H.R.R. 3. nm (OH radical) 395. nm (HCHO).3..1 Fig. 9 Results of Chemical Kinetics Simulation (N = 1 rpm, φ =.5, ε = 1) H.R.R. [J/deg.] E HCHO In order to examine this light emission further, Fig. 9 shows the chemical kinetics simulation results for numerical calculations of the elementary reactions. The conditions used in the simulation were an engine speed N of 1 rpm, compression ratio ε of 1:1, equivalence ratio φ of.5, initial temperature T of 37 K and initial pressure P of.1 MPa. The horizontal axis indicates the crank angle θ. Shown from the top along the vertical axis is the cylinder pressure P (MPa), heat release rate (HRR, J/deg.), in-cylinder gas temperature (K) and the mole fractions of the OH radical, HCHO, CO radical and CO. The vertical axis of the chemical species grpah is given in logarithmic notation. The HRR waveform shows a two-stage heat release pattern attributable to low- and high-temperature oxidation 1 1 H.R.R [J/deg.] SETC 5 5

6 reactions. At the point where the low-temperature oxidation reactions are manifested, the HCHO mole fraction increases gradually and accumulates until just before the appearance of the high-temperature oxidation reactions, and then it decreases suddenly once the latter reactions begin. This is a definite sign of the production of HCHO during cool flame reactions, and the waveform indicates that this light emission was detected in the test engine. On the other hand, the OH radical concentration is low compared with that of HCHO. The production of the CO radical that characterizes high-temperature oxidation reactions rises together with the extinction of HCHO and eventually the reaction proceeds to the production of CO. In other words, it is assumed that the sharp rise in light emission observed in the measurements made with the test engine represents the detection of the continuous light emission spectrum of CO-O radiation that occurred in this process. (1). ANALYSIS.1 HEAT RELEASE FROM LOW-TEMPERATURE OXIDATION REACTIONS The calculated heat release from the low-temperature oxidation reactions is shown in Fig. 1 along the vertical axis as a function of the equivalence ratio along the horizontal axis. Figure 11 shows the duration of the low-temperature oxidation reactions as a function of the equivalence ratio, and Fig. 1 shows the average heat release per degree from the low-temperature oxidation reactions as a function of the equivalence ratio. Three main factors are considered here that influence the rise of the in-cylinder gas temperature in the progression to the high-temperature oxidation reactions. Those factors are the temperature rise due to compression of the gas by piston motion, the temperature rise due to heat release from the low-temperature oxidation reactions, and the transfer of heat to the gas from the combustion chamber walls. Comparing the results in Fig. 1 for the range of compression ratios ε from 11:1 to 19:1, it is seen that the heat release from the low-temperature oxidation reactions increases with a lower compression ratio. The reason for that can be understood as follows. In the crank angle interval from 9 deg. before top dead center (BTDC) to TDC, the piston speed becomes slower at later crank angles, with the result that the gas temperature rises more slowly due to compression. Accordingly, the progress of the low-temperature oxidation reactions during that interval presumably accounts for the increased heat release. With regard to the change in the equivalence ratio, the heat release from the low-temperature oxidation reactions increases accompanying a larger equivalence ratio in the region of a relatively low equivalence ratio (denoted as A in Fig. 1). With a low equivalence ratio, heat release reactions proceed slowly, resulting in a longer reaction duration as seen in Fig. 11. As a result, compression of the gas by the piston and heat Heat Release from Low-Temperature Reactions, [J] Duration of Low-Temperature Reactions, [deg.] Average Heat Release from Low-Temperature Reactions [J/deg.] Equivalence Ratio, [-] Fig. 1 Heat Release from Low-Temperature Reactions Equivalence Ratio, [-] Fig. 11 Duration of Low-Temperature Reactions A B Equivalence Ratio, [-] Fig. 1 Average Heat Release from Low-Temperature Reactions Onset Time of Low-Temperature Reactions, [deg.b.t.d.c.] Equivalence Ratio, [-] Fig. 13 Onset Time of Low-Temperature Reactions SETC 5

7 transfer from the combustion chamber walls are the factors that have a larger influence on the temperature rise, and that of the heat release decreases. As the equivalence ratio increases, the low-temperature oxidation reactions are promoted, and compression and heat transfer from the combustion chamber walls have less influence on the gas temperature rise. Accordingly, the heat needed for the progression to the high-temperature oxidation reactions is obtained from the heat released by the low-temperature oxidation reactions, with the result that the total heat release presumably increases. It is seen in the region of a relatively high equivalence ratio (denoted as B in Fig. 1) that the heat release values remain nearly constant in spite of the increase in the equivalence ratio. In this regard, a look at Fig. 1 shows that the average heat release per degree continues to increase even in the region where the heat release values remain constant. It is assumed from this observation that the low-temperature oxidation reactions are promoted in this region. However, the reason that the heat release values are constant in the region of a relatively high equivalence ratio (B) is that the low-temperature oxidation reactions are sufficiently promoted. Consequently, the temperature for the initiation of high-temperature oxidation reactions is reached in a short period of time, and a progression to the high-temperature oxidation reactions takes place. As a result, the heat release values become nearly constant in the region of a relatively high equivalence ratio. Figure 13 shows the onset time of the low-temperature oxidation reactions along the vertical axis as a function of the equivalence ratio along the horizontal axis. It is seen in this figure that the onset time of the low-temperature oxidation reactions is nearly constant at the same compression ratio, regardless of the equivalence ratio. The presumable reason for that is because HCCI combustion starts when compression by the piston raises the temperature of the in-cylinder gas to the level where the low-temperature oxidation reactions begin. It is also seen that the compression pressure and in-cylinder gas temperature are lower at the same crank angle with a lower compression ratio, which explains why the low-temperature oxidation reactions begin at a later crank angle. A comparison of the results in Fig. 1 for different intake air temperatures indicates that the heat release during the low-temperature oxidation reactions tends to increase with an increasing equivalence ratio at each temperature level. In addition, as the equivalence ratio increases, the rate of increase in the heat release tends to become more gradual. The reason for that can be understood as follows. The increase in the equivalence ratio works to promote the reactions, with result that the in-cylinder gas reaches the temperature needed for the start of the high-temperature oxidation reactions and a progression to the latter reactions takes place. Looking at the results for each temperature level, it is seen that the heat release becomes relatively lower at higher intake air temperatures. That can be attributed to the following reason. As the intake air temperature Heat Release from Low-Temperature Reactions, [J] Duration of Low-Temperature Reactions, [deg.] Equivalence ratio,[-] T IN =3 [K] T IN =93 [K] T IN =33 [K] Fig.1 Heat Release from Low-Temperature Reactions varying intake air temperature, T IN T in =3 [K ] T in =93 [K ] T in =33 [K ] [-] Equivalence Ratio Fig.15 Duration of Low-Temperature Reactions varying intake air temperature, T IN rises, the in-cylinder gas temperature is higher at the start of compression. As a result, the temperature for the onset of the high-temperature oxidation reactions is reached with the small amount of heat released from the low-temperature oxidation reactions. It is clear from the results in Fig. 15 that the duration of the low-temperature oxidation reactions becomes longer as the intake air temperature rises. As the intake air temperature increases, the onset time of the low-temperature oxidation reactions shifts to an earlier crank angle. The shift of the onset time to an earlier crank angle means that the cylinder pressure is lower at the point when the low-temperature oxidation reactions begin. Although the gas temperature is high at the onset of the reactions, the lower pressure in the interval of the low-temperature oxidation reactions means that the reactions proceed slowly. That is why the duration of the reactions becomes longer.. COMPARISON OF OUTPUT FOR DME-HCCI AND DIESEL FUEL Figure 1 shows the crankshaft output Le along the vertical axis in relation to the equivalence ratio along the horizontal axis. For the sake of comparison, the test engine was also operated on diesel fuel, and the results obtained are compared in the figure with the output measured for DME-HCCI combustion. It is seen that reducing the compression ratio increased the crankshaft output. That is thought to be attributable to the delayed ignition timing (i.e., combustion in the vicinity of TDC) and the higher equivalence ratio in the operable region of SETC 5 7

8 the engine. A lower compression ratio and a higher equivalence ratio made it possible to achieve crankshaft output with DME-HCCI combustion that approached the output seen for diesel operation at a low equivalence ratio Diesel HCCI Equivalence Ratio, Fig. 1 Output for DME-HCCI and Diesel Fuel 5. CONCLUSIONS The experimental results obtained in this study with dimethyl ether (DME) as the test fuel made the following points clear. (1) In experiments conducted under conditions of a low compression ratio and a high equivalence ratio, light emission at a wavelength corresponding to HCHO was observed during low-temperature oxidation reactions. () Low-temperature oxidation reactions began at nearly the same time depending on the compression ratio without being influenced by the equivalence ratio. (3) At lower compression ratios, the temperature of the in-cylinder gas rose more slowly due to compression by the piston, with the result that both the low-temperature and high-temperature oxidation reactions began at later crank angles. () The average heat release per degree from the low-temperature oxidation reactions increased with an increasing equivalence ratio. (5) A higher intake air temperature shifted the onset of both the low-temperature and high-temperature oxidation reactions to an earlier crank angle. () The low-temperature oxidation reactions proceeded more slowly due to a rise in the intake air temperature. (7) The onset time of the high-temperature oxidation reactions is presumed to be strongly influenced by the amount of heat released during the interval of the low-temperature oxidation reactions. () As a result of lowering the compression ratio and increasing the equivalence ratio, the crankshaft output obtained during DME-HCCI operation approached the output measured for test engine operation on diesel fuel at a low equivalence ratio. Experiments are under way to improve the power output of the HCCI combustion process further by using a blend fuel prepared by adding methane to DME and by applying EGR REFERENCES (1) M. Konno, S. Kajitani: Combustion Characteristics of a Compression Ignition Engine Fueled with Dimethyl Ether, Transaction of JSME (in Japanese),Vol. 7 No. B, pp. 3-9, 1. () Ohta. Y., Furutani. M., Komatu. K., Onset Behavior of Low-Temperature Autoignition Caused by Piston Compression, Proceedings of the 1th Internal Combustion Engine Symposium (in Japanese), pp. 5-7, 199. (3) Furutani. M., Ohta. Y., A Novel Method of Eliminating Piston-Compression Ignition, Transaction of JSME (in Japanese), Vol., No. 571B, pp , 199. () Konno. M, S. Kumagai, and Inuma K.,Thermodynamic and Experimental Determination of Knocking Intensity by Using a Spark Ignited Rapid Compression Machine, Combustion and Flame, Vol. 5, pp. 33-7, 193. (5) Green, R.M., Cernansky, N.P., Pitz, W.J. and Westbook, C.K.,The Role of Low Temperature Chemistry in Autoignition of N-Butane, SAE Paper 71, 197. () Gaydon, A. G., "The Spectroscopy of Flames", Chapman and Hall Ltd., London, 1957 (7) Ohara, H., Ogawa, J., Yoshida, K., Shoji, H., Saima, A., An Analysis of Light Emission Intensity Behavior Corresponding to Intermediate Products in Different Places of the Combustion Chamber, SETC1, Paper No () T. Koseki, S. Tamura, H. Kashiwagi, K. Yoshida, H. Shoji, Measurement of Radical Behavior in Homogeneous Charge Compression Ignition Combustion Using Dimethyl Ether, SAE Paper (9) Tosaka, Y., Shoji, H., and Saima, A., A study of the influence of intermediate combustion products on knocking, JSAE Review 9533, Vol. 19, No. 3, (1995). (1) Leppard, W. R., The Autoignition Chemistry of Isobutane: A motored Engine Study, SAE Paper 11, 19. (11) Leppard, W. R., The Autoignition Chemistries of Primary Reference Fuels, Other Mixture and Blending Octane Numbers, SAE Paper 935, 199. (1) N. Iida, T. Igarashi Study of Auto Ignition and Combustion in Homogeneous Charge Compression Ignition Engine, Proceeding of the 1th Internal Combustion Engine Symposium (in Japanese), pp. 77-, (13) Curran, H. J., W.J., Westbook, C.K., Dagaut, P., Boettner, J-C., Cathonnet, M., A Wide Range Modeling Study of Dimethyl Ether Oxidation, International Journal of Chemical Kinetics, Vol. 3, No. 3, pp. 9-1,199. (1) M.Oguma, S.Goto A Measurement of Combustion Radicals Fueled with Dimethyl Ether (DME), Transactions of JSME (in Japanese), Vol., No. -1, pp. 1-,. (15) K. Yoshida, T. Koseki, S. Tamura, K. Yoshida, H. Shoji, Combustion Analysis of the Effect of Compression Ratio on HCCI using Dimethyl Ether, Proceedings of the Forty-Second Symposium on Combustion (in Japanese), pp.95-9,. (1) E.Tomita, et al., The th AJTEC 3. Paper No. TED-AJ SETC 5