Experimental Study on HCCI Combustion in a Small Engine with Various Fuels and EGR

Because Asian countries have a large number of motorcycles, motorcycle engine exhaust emission poses a major problem in the region. Homogeneous charge compression ignition (HCCI) is a promising combustion technology with high efficiency and low nitrogen oxide emission. This study investigated the combustion characteristics of HCCI in a 150 cc motorcycle engine with three types of dual fuel. The main fuels consisted of dimethyl ether (DME), kerosene, and nheptane, and gasoline was the additive fuel. External exhaust gas recirculation (EGR) was used to expand the operating range of the engine. All test points were executed under stable HCCI operation with a coefficient of variation < 5%. A total of 107 data sets were obtained by adjusting the amount of main fuel, additive fuel, and EGR. Two-stage ignition was observed because of the diesel-like main fuels. The results revealed that the temperature at the start of combustion was approximately 500–700 K, and the temperature of the first-stage maximum heat release rate (MHRR1) was approximately 600–800 K. The relationship between the crank angle (CA) at which the mass fraction burnt is 50% (CA50) and the CAs of the maximum cylinder pressure, maximum cylinder gas temperature, and MHRR were linear. The R square for the curve fitting of each relationship was > 0.9. The findings suggest that adjusting the CA50 to a value greater than 5° after top dead center, or limiting the maximum cylinder pressure to < 50 bar is crucial to preventing excessive pressure rising. Overall, kerosene was deemed unsuitable for engines because of its high sulfur content, whereas DME is considered an excellent option.


INTRODUCTION
The exhaust emissions of internal combustion engines are major sources of air pollution (Cheng, 2013;Yao and Tsai, 2013), which deteriorates human health and causes the global warming effect.Therefore, numerous technologies regarding engine combustion, after treatments, and fuels have been studied.Previous studies have demonstrated that using a three-way catalytic converter in spark ignition (SI) engines can reduce most exhaust pollutants including carbon monoxide (CO), hydrocarbon (HC), and nitrogen oxides (NO x ) (Lou et al., 2003).However, catalytic converters do not reduce carbon dioxide (CO 2 ), which is a main contributor of the global warming effect.Several fuel blends and additives that improve engine efficiency and reduce emissions in diesel engines have been developed, such as microalgae biodiesel (Mwangi et al., 2015a, b), biodiesel (Lu et al., 2013;Shukla et al., 2014), H 2 /O 2 mixture (Wang et al., 2012(Wang et al., , 2013)), and plasma (Lin et al., 2013).However, they cannot be used in motorcycle engines.Motorcycle exhaust emissions are a major concern in Asia because motorcycles are a commonly used mode of transport in most countries (Lin et al., 2014) and generate more air pollution than other vehicles (Chuang et al., 2010).Thus, exhaust emission regulations have become increasingly stringent for motorcycles.
A homogeneous charge compression ignition (HCCI) engine has a combination of conventional SI engine and diesel engine characteristics.Its fuel-air mixture is premixed, as in a conventional SI engines, whereas its combustion is initiated by autoignition, as in diesel engines.Therefore, HCCI has been widely investigated in recent years as a promising alternative combustion technology with high efficiency and very low emissions of NO x and soot (Sjöberg and Dec, 2007;Wang et al., 2009;Agarwal et al., 2015;Nishi et al., 2015).Nevertheless, HCCI also has unresolved challenges, including combustion phase control, a narrow operating range, cold starts, high noise levels, and homogeneous charge preparation (Hasan and Rahman, 2016).
Methods that facilitate HCCI operation can be classified according to two factors: fuel type and combustion environment.The fuel types include diesel-like fuel (high cetane number) and gasoline-like fuel (high octane number).The combustion environment category considers factors including the compression ratio, intake temperature, air fuel ratio, engine speed, residual exhaust gas, and exhaust gas recirculation (EGR).
Diesel-like fuels, such as n-heptane, kerosene, dimethyl ether (DME), and biodiesel, have two-stage ignition.By contrast, gasoline-like fuels, such as iso-octane, ethanol, liquefied petroleum gas, and butanol, have single-stage ignition.For diesel-like fuels, the first stage of ignition is the result of cool combustion and a negative temperature coefficient (Kelly-Zion and Dec, 2000).Diesel-like fuels can autoignite easily and do not require high compression ratios compared with gasoline-like fuels, which require high compression ratios, intake heating, or high levels of residual gas retained in the cylinder.Kelly-Zion and Dec's (2000) simulation demonstrated that at a given engine speed and air fuel ratio, and an intake temperature of 340 K, the compression ratio that produced ignition near top dead center was 13 and 27 for n-heptane and iso-octane, respectively.
Several techniques for improving the combustion environment of HCCI have been studied, such as intake heating (Martinez-Frias et al., 2000), varying the compression ratio (Haraldsson et al., 2002), and negative valve overlap (Zhong et al., 2006).Nevertheless, these methods are too complex to employ in a motorcycle engine.
A simple approach to conducting HCCI in a conventional SI engine is to use dual fuel with EGR (Wu et al., 2015).A diesel-like fuel that autoignites easily can be used as the main fuel and combined with an ignition suppressor.Gasolinelike fuel and EGR can be used as ignition suppressors.Ogawa et al. (2003) reported that ignition suppressors, such as water, methanol, ethanol, hydrogen, and methane, can be employed to control ignition timing and expand the operating range in an HCCI engine with DME as the main fuel.Numerous studies have demonstrated that the addition of EGR slows the autoignition process (Dec, 2009) and timing (Lu et al., 2005;Sjöberg and Dec, 2011), and thus enables controlling combustion phasing (Jang, 2013;Nishi et al., 2016).
This study investigated the combustion characteristics of HCCI in a motorcycle engine by using several fuels.Kerosene, n-heptane, and DME were selected as main fuels and gasoline and EGR were used as ignition suppressors.

HCCI engine
A 150 cc single-cylinder, air-cooled SI engine was chosen as the target engine.It is a motorcycle engine with an electronic fuel injection system.Detailed engine specifications are listed in Table 1.
The target engine was retrofitted for HCCI operation without considerable changes for easy application in motorcycles.To conduct HCCI in the target engine, the compression ratio was increased from 10.5 to 12.4 by replacing the cylinder head with a smaller clearance volume.The increased compression ratio produced a higher compression temperature, which facilitated compression ignition.
Because HCCI combustion occurs through autoignition, the autoignition property of fuels was a critical factor of HCCI operation.DME, kerosene, and n-heptane were selected as the main fuels for their low autoignition temperatures.Gasoline was used as an additional fuel to adjust the ignitibility of the dual-fuel mixture.The dual fuel and EGR were incorporated to expand the engine operating range.Fuel properties are listed in cetane number of DME and n-heptane is 60, which is high compared with the normal cetane range of diesel fuel for motor vehicles (approximately 40-60).The cetane number of kerosene is lower than that of diesel fuel.However, all three fuels can autoignite in the target engine.By contrast, the autoignition temperature of gasoline is relatively high, preventing it from auto igniting.
The configuration of the proposed HCCI engine is shown in Fig. 1.A dual-fuel supplying system was built into the target engine.The original fuel and ignition systems of the SI engine were retained for starting the engine.The original gasoline injector was also used to supply gasoline in HCCI.The additional HCCI main-fuel supplying systems for DME, kerosene, and n-heptane were attached to the target engine separately for each HCCI fuel test.After starting the engine using gasoline in SI mode and achieving a stable cylinder head temperature, the controller switched the engine operation to HCCI mode with dual fuel.To measure the gasoline and HCCI main fuel flow rates, the gasoline flow meter, gasoline defoaming tank, and HCCI main fuel flow meter were attached to the dual-fuel supplying system.Both fuels were used after starting the engine.The fuel supply was adjusted for stable HCCI operation; therefore, the percentages of these two fuels were not fixed.
The external EGR system was established on the target engine with a small EGR pump and an EGR control valve, which adjusted the EGR ratio and thus controlled HCCI combustion.

Experimental setup
For the engine test, an eddy-current engine dynamometer (Borghi & Saveri S.R.L., FE150-S) was employed to measure the engine brake torque and speed.The gasoline flow rate was measured using a mass burette flow detector (Ono Sokki, FX-1110).A thermal mass flow controller (Tokyo Keiso, NM-2100) was utilized to measure the DME flow rate.The exhaust emissions of CO, HC, nitric oxide (NO), CO 2 , oxygen and the air-fuel ratio were measured using an emission analyzer (Horiba, MEXA-584L).Additionally, another emission analyzer (Horiba, MEXA-584L) was installed in the intake system to measure the CO 2 percentage for EGR ratio calculations.Several K-type thermocouples were installed on the engine for measuring the temperature of the ambient air, intake gas, exhaust gas, cylinder head, oil, and EGR, as shown in Fig. 1.
The cylinder pressure was measured using a Kistler 6051B air-cooled piezoelectric pressure transducer, as shown in Fig. 2. A shaft encoder (BEI H25) was used to detect the crank angle (CA).The output signal of the pressure transducer was amplified using a charge amplifier.This amplified signal was transmitted to a data acquisition system, an AVL IndiCom 619 combustion analyzer.Additionally, the cylinder pressure was recorded for every 1° of the CA for 100 cycles.The pressure data were then used for calculating the heat release rate (HRR).
Because the target engine was retrofitted for HCCI operation, the spark timing (for SI mode), gasoline injector, kerosene or n-heptane injector, DME flow control valve, EGR pump, and EGR control valve were manually controlled by a control system (Woodward, MotoHawk ECU 555-80).The MotoHawk enables the user to automatically generate machine codes from Simulink blocks and operate control hardware in real time.
Temperature is the most critical factor in HCCI combustion (Mack et al., 2005); thus, the engine cylinder head and oil temperatures were monitored using negative temperature coefficient thermistors.After these thermistors were connected to the input interface, Motohawk ECU for thermistor, voltage signals could be generated and read for engine-control algorithms.In this study, cylinder head and oil temperatures were held above 120°C and 65°C, respectively, for stable HCCI operation (Wu et al., 2010).

Testing
In most port-injection SI engines, fuel injection ends before the intake valve opens to prevent liquid fuel from entering the cylinder directly.A fuel injection timing of 90° bTDC (before top dead center) of the compression stroke was chosen not only to prevent the liquid fuel from entering the cylinder directly but also to maintain the homogeneity of the air-fuel mixture (Wu et al., 2010).
The engine speed, brake torque, HCCI main-fuel flow rate, gasoline flow rate, intake air mass flow rate, exhaust emissions, and EGR ratio were measured and recorded during engine tests.The EGR ratio was calculated using the CO 2 percentages measured in the intake and exhaust systems, as in Eq. (1) (Yao et al., 2005).
HCCI with DME fuel was tested at 2000 rpm five times in the target engine.The uncertainty of the retrieved experimental data was subsequently calculated using the Kline and McClintock method (Holman, 1989).The results are listed in Table 3.
After the fuels, engine temperature, and fuel injection timing were determined, HCCI combustion was executed on the target engine.First, the target engine was started using SI with its original ignition and fuel systems.When the cylinder head and oil temperatures reached 120°C and 65°C, respectively, the engine was switched to HCCI mode by interrupting the ignition system and turning the throttle to wide open, while adjusting the fuel supply for Several test points were executed under stable HCCI engine operation, and the combustion properties were then analyzed using the HRR.The test points with a coefficient of variation (COV) > 5% were deleted.The data sets of the main fuels comprised 17 sets of DME, 70 sets of kerosene, and 20 sets of n-heptane data.Thus, the test data comprised 107 sets.

Heat Release Rate Calculation
Cylinder pressure can be used to analyze engine combustion parameters such as the in-cylinder gas temperature, COV, and HRR.
The in-cylinder gas temperature can be obtained using the state equation of ideal gas.The COV is a reliable indicator of engine combustion variation and is defined as the standard deviation of indicated mean effective pressure (IMEP) in 100 cycles (IMEP std ) divided by the average IMEP in 100 cycles (IMEP avg ), as expressed in Eq. ( 2).
The HRR can be obtained using the experimental cylinder pressure.The equation derived from the first law of thermodynamics is where dQ hr /dθ is the HRR, dQ ht /dθ is the heat transfer rate between cylinder gas and the wall, θ is the CA, γ is the specific heat ratio of cylinder gas, and P and V are the cylinder pressure and volume, respectively.A heat transfer model for motorcycles was proposed by Wu et al. (2006).The specific heat ratio γ is a function of temperature.In the combustion and exhaust strokes, it can be expressed as Eq. ( 4) (Ceviz and Kaymaz, 2005).
where T is the cylinder gas temperature in Kelvin.
where Q HV is the low heating value of fuel, and m f is the fuel mass supplied per cycle.

RESULTS AND DISCUSSION
DME, kerosene, and n-heptane are diesel-like fuels that can autoignite in the target engine.However, the operating range of each fuel is markedly narrow.Gasoline and EGR additives were used to expand their operating ranges.All experiments were performed by adjusting the rates of the main fuel, gasoline, and EGR to achieve stable operation under the constraint of COV < 5%.
The combustion characteristics of the HCCI engine were investigated by analyzing the cylinder pressure and HRR.Several combustion parameters are defined in Fig. 3: MHRR1 and MHRR2 are the maximum heat release rate (MHRR) in the respective first and second stages of ignition, the start of combustion (SOC) is the CA at which the HRR achieves 20% of the MHRR1, and CA50 is the CA at which the MFB is 50%.

Ignition Properties
In this study, two-stage ignition was observed in all experimental results.As defined in Fig. 3, the first stage of ignition can be considered the SOC.The SOC was approximately 0°-30° bTDC, and most data were gathered at 10°-30° bTDC, as shown in Fig. 4. The SOC temperature was 400-800 K, and most data were gathered at 500-700 K.The SOC of kerosene was earlier than that of the other fuels.Generally, the timing of the SOC does not affect the temperature of the SOC.
The CA of MHRR1 indicates that the low-temperature reaction achieved MHRR and then slowed because of the temperature increase in the cylinder.The first stage of ignition was the result of cool combustion, which is the negative temperature coefficient of reaction.The temperatures of MHRR1 were 600-800 K, as shown in Fig. 5. Kelly-Zion and Dec (2000) reported that this two-stage ignition is caused by a competition between chain-branching and chainpropagating reactions.Chain-branching reactions begin at approximately 800 K.
The period between MHHR1 and MHHR2 indicates the transition from low-temperature reaction to main combustion.Table 4 displays the CA between MHRR1 and MHRR2 for   each fuel and the total data.For the total data, the average CA of all data is 18.2°, representing 19.2°, 18.1°, and 16.9°, for DME, kerosene, and n-heptane, respectively.The standard deviations of these data are ≤ 2.4° CA.

Correlations between Combustion Parameters
CA50 is widely used for analyzing engine combustion.Therefore, understanding the relationship between CA50 and other combustion parameters is beneficial for investigating HCCI engines.The study of these relationships was based on all 107 data sets.
Fig. 6 illustrates that the relationship between CA50 and the CA of maximum cylinder pressure is linear.The R square (R 2 ) of the curve fitting for each fuel is > 0.9.The relationship between CA50 and the CA of maximum cylinder temperature (Fig. 7) and that between CA50 and the CA of MHRR (Fig. 8) are also markedly linear.Here, the MHRR is in second stage of ignition.The CA of MHRR increased linearly with respect to the CA of cylinder pressure, as shown in Fig. 9.
Furthermore, the maximum rate of pressure rise (MRPR) was inversely, yet nonlinearly, related to maximum cylinder pressure (as shown in Fig. 10).
Table 5 displays the equation of curve fitting and the R 2 for each pair of combustion parameters.
The correlation of the CA of the MHHR and CA50 was 0.966 for all data sets and 0.994, 0.966, and 0.994 for DME, kerosene, and n-heptane, respectively.The MHRR occurred at a CA 0.66° after CA50, according to the average of all data sets.Thus, CA50 and the CA of the MHRR are close in value.The correlation of the CAs of maximum temperature and maximum pressure was 0.961 for all data sets and 0.961, 0.979, and 0.968 for DME, kerosene, and n-heptane, respectively.The maximum cylinder gas temperature arose at a CA 3.2° greater than that of the maximum cylinder pressure, according to the average of all data sets.
The curve fitting, R 2 , and correlation were determined using the functions of Microsoft Excel.For the two arrays y 1 and y 2 , the correlation and R 2 can be expressed as Eqs.( 6) and (7), respectively.
where y ̅ is the mean of y.
Here, y 2 can be the curve fitting value of y 1 regarding an independent number x. R 2 is 0-1.If R 2 is close to 1, the curve fitting result is close to the experimental value, and vice versa.

Engine Operating Properties
The engine throttle was turned to wide open for all operations of HCCI; thus, adding fuel was an approach to increasing the engine load.Herein, the engine load is indicated by the IMEP.However, an excessively pressure rise rate limits the high load operating range of HCCI engines (Mack et al., 2005;Sjöberg and Dec, 2007;Nishi et al., 2016).For all test data under the constraint of COV < 5%, the IMEP was not affected by the CA50 nor the maximum cylinder pressure, as shown in Figs.11 and 12.Under stable HCCI, the operating range was within an IMEP of 3-7 bar.
The MRPR and COV are generally used as indices of combustion quality.A high MRPR causes combustion noise (Sun et al., 2004) and engine damage (Mack et al., 2005).Previous researchers have set a limit of 6 bar deg -1 for the MRPR (Sun et al., 2004).
High maximum cylinder pressure caused a high MRPR (Fig. 10).When the cylinder pressure was < 50 bar, the MRPR was < 6 bar deg -1 .After further decline of the cylinder pressure to < 40 bar, the MRPR was considerably low and was unaffected by the cylinder pressure.Another factor Fig. 6.Relationship between CA50 and CA of maximum cylinder pressure.Table 5. Relationship between the combustion parameters.The exhaust emissions of this HCCI engine were discussed in previous research (Wu et al., 2015).The NO emission of HCCI is very small which is close to zero and the CO emission is also small as compared with the SI engine.However, the HC emission of HCCI engine is higher than that of SI engine.The overall evaluation of the dual-fuel systems revealed   that kerosene is unsuitable for engines because of its high sulfur content (1% as reported by the Taiwan Environmental Protection Administration).During testing, the sulfur content caused substantial carbon deposit and damaged the engine several times.DME is a type of biofuel that is environment friendly.Its low boiling point and high ignition ability are suitable for HCCI operation.Moreover, DME (CH 3 OCH 3 ) combustion is soot free, because it lacks double C-C bonds and contains oxygen atoms that oxidize intermediate soot product (Jang et al., 2013).An HCCI engine using a DME-gasoline dual fuel with EGR was demonstrated in a range-extended electric motorcycle (Wu et al., 2015).Hence, DME-gasoline dual fuel is an excellent choice for HCCI in small engines.

CONCLUSIONS
HCCI combustion was performed in a 150 cc air-cooled, four-stroke motorcycle engine with three types of dual fuel.DME, kerosene, and n-heptane were designated main fuels, whereas gasoline was an additive fuel.Combustion characteristics were analyzed using HRR calculations based on 107 data sets and led to the following conclusions: (1) The SOC temperature was approximately 500-700 K, and the MHRR1 temperature was approximately 600-800 K.The CA between MHRR1 and MHRR2 was approximately 18.2°.
(2) The relationships between CA50 and the CAs of the maximum cylinder pressure, maximum cylinder gas temperature, and MHRR, are linear.The R 2 for the curve fitting of these relationships was > 0.9.(3) The CA50 was very close to the CA of the MHRR.(4) Under stable HCCI, the IMEP had an operating range of 3-7 bar and was not affected by the CA50 nor the maximum cylinder pressure.(5) To maintain a MRPR < 6 bar deg -1 , a CA50 < 5° aTDC should be avoided or the maximum cylinder pressure should be limited to < 50 bar.(6) Kerosene is unsuitable for engines because of its high sulfur content, whereas DME is an excellent choice because of its low boiling point and effective ignition ability.

Fig. 3 .
Fig. 3. Illustration of the combustion parameters on HRR and MFB curves.

Fig. 4 .Fig. 5 .
Fig. 4. Effect of the timing of start of combustion on cylinder gas temperature.

Fig. 9 .
Fig. 9. Relationship between CA of MHRR and CA of maximum cylinder pressure.
affects the MRPR is CA50.The MRPR decreased as CA50 decreased, as shown in Fig.13.To maintain an MRPR < 6 bar deg -1 , adjusting the CA50 to a value greater than 5° after top dead center (aTDC), or limiting the maximum pressure to < 50 bar would be effective.
b Auto-ignition temperatures are obtained from Material Safety Data Sheet (MSDS).

Table 3 .
Uncertainty in experimental results of the target engine.Experiments were then conducted at various HCCI main-fuel flow rates, engine speeds, EGR ratios, and added gasoline amounts to investigate the operating characteristics of HCCI combustion.

Table 4 .
Crank-angle between the points of MHRR1 and MHRR2.