Experimental investigation on the effect of intake air temperature and air–fuel ratio on cycle-to-cycle variations of HCCI combustion and performance parameters

Applied Energy 88 (2011) 1153–1163

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Experimental investigation on the effect of intake air temperature and air–fuel ratio on cycle-to-cycle variations of HCCI combustion and performance parameters Rakesh Kumar Maurya, Avinash Kumar Agarwal ⇑ Engine Research Laboratory, Department of Mechanical Engineering, Indian Institute of Technology Kanpur, Kanpur 208016, India

a r t i c l e

i n f o

Article history: Received 12 May 2010 Received in revised form 8 September 2010 Accepted 29 September 2010 Available online 23 November 2010 Keywords: Homogeneous charge compression ignition (HCCI) Coefficient of variation (COV) Ethanol Combustion stability Cycle-to-cycle variation

a b s t r a c t Combustion in HCCI engines is a controlled auto ignition of well-mixed fuel, air and residual gas. Since onset of HCCI combustion depends on the auto ignition of fuel/air mixture, there is no direct control on the start of combustion process. Therefore, HCCI combustion becomes unstable rather easily, especially at lower and higher engine loads. In this study, cycle-to-cycle variations of a HCCI combustion engine fuelled with ethanol were investigated on a modified two-cylinder engine. Port injection technique is used for preparing homogeneous charge for HCCI combustion. The experiments were conducted at varying intake air temperatures and air–fuel ratios at constant engine speed of 1500 rpm and P-h diagram of 100 consecutive combustion cycles for each test conditions at steady state operation were recorded. Consequently, cycle-to-cycle variations of the main combustion parameters and performance parameters were analyzed. To evaluate the cycle-to-cycle variations of HCCI combustion parameters, coefficient of variation (COV) of every parameter were calculated for every engine operating condition. The critical optimum parameters that can be used to define HCCI operating ranges are ‘maximum rate of pressure rise’ and ‘COV of indicated mean effective pressure (IMEP)’. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Considering continuously stringent emission regulations, as well as increasing shortage of primary energy resources, the development of new highly efficient and environment friendly combustion systems, associated with alternative fuels has become increasingly important and hence research need to be carried out in this domain. Homogeneous charge compression ignition (HCCI) combustion concept received significant focus in recent years because of the advantages it offers. HCCI was identified as a distinct combustion phenomenon about 30 years ago [1,2]. The Homogeneous charge compression ignition engine concept is has potential to overcome the current fundamental NOx and particulate emission trade-off limitation of conventional diesel engines. However, HCCI mode engines generate higher amount of unburned hydrocarbons (HC) compared to conventional engines and operate at significantly lower indicated mean effective pressure (IMEP) [3–11]. Also, there are difficulties associated with control of combustion initiation and rate of combustion over the required speed and load range of the engines [12,13]. These factors presently restrict the commercial exploitation of the HCCI combustion concept in the engine applications. ⇑ Corresponding author. Tel.: +91 512 259 7982; fax: +91 512 259 7408. E-mail address: [email protected] (A.K. Agarwal). 0306-2619/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2010.09.027

The main problem with the HCCI is that the ignition is completely controlled by chemical kinetics, and is directly affected by the fuel composition, equivalence ratio, and thermodynamic state of the fuel–air mixture [14–16]. There is no external control of initiation of combustion such as the fuel injection or spark timing that are used in traditional CI or SI engines. Achieving the required level of control during transient engine operation is even more challenging since charge temperature has to be correctly matched to the operating condition during rapid transients with a high repeatability when the speed and load are changing. The ignition timing and combustion rates are dominated by physical and chemical properties of fuel/air/residual gas mixtures, boundary conditions including environmental temperature, pressure, and humidity and engine operating conditions such as load, speed, etc. Because of these reasons, wide cycle-to-cycle variations are observed in HCCI combustion engines. Even small changes in ignition timing and combustion rate bring large variation in engine performance and emissions [17]. As a result, wide cycle-to-cycle variations, combustion instability under lean combustion condition and knock combustion constrain operating range of HCCI combustion engines. Cycle-to-cycle variations in combustion process are essential as they play important role in combustion stability and operating limit decision for the HCCI engine operating range. Many researchers reported coefficient of variation of IMEP (COVIMEP) and maximum rate of pressure rise as HCCI operating region criteria [18,19].

1154

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

Nomenclature CAD CI COV COVIMEP EVO dP/dh HCCI IMEP ISFC IVC P

crank angle degree compression ignition coefficient of variation coefficient of variation of indicated mean effective pressure exhaust valve opening rate of pressure rise homogeneous charge compression ignition indicated mean effective pressure indicated specific fuel consumption intake valve closing cylinder gas pressure

Mechanism and control of cycle-to-cycle variation in SI engines are systematically investigated by several researchers [20–25]. However, – little work has been reported on cycle-to-cycle variation and combustion stability of HCCI combustion. Xingcai et al. investigated the combustion stabilities and cycle-to-cycle variations of HCCI combustion using n-heptane, primary reference fuels 20 (PRF20), PRF40, PRF50 and PRF60 [17]. Lu et al. also investigated cycle-to-cycle variations under the lean burn limits and rich burn limits in HCCI Combustion using n-heptane [26]. Persson performed preliminary study on the cylinder-to-cylinder and cycleto-cycle variations of controlled auto ignition (CAI) combustion with trapped residual gas [27]. Koopmans et al. investigated the cycle-to-cycle variations in a camless gasoline fuelled compression ignition engine [28]. Shi et al. investigated combustion stability of diesel fuelled HCCI and effects of engine load, speed and valve overlap [29]. Shahbakhti and Koch performed investigations of cyclic variation of ignition timing using primary reference fuels [30]. These cyclic variations and combustion instabilities lead to necessity of closed loop control of combustion phasing. This motivates the researchers to investigate the cycle-to-cycle variations in HCCI combustion engines. To gain an improved understanding of HCCI combustion, a systematic study of cycle-to-cycle variation of HCCI combustion is essential. The objective of this study is to investigate the effect of intake air temperature and air–fuel ratio on cycle-to-cycle variations in an ethanol fueled port injection HCCI engine operating at constant engine speed (1500 rpm).

2. Experimental conditions A two cylinder, four stroke, air cooled, naturally aspirated, bowl shaped combustion chamber design; direct injection diesel engine was modified for these experiments. The engine specifications are given in Table 1. One of the two cylinders of the engine is modified to operate in HCCI mode, while the other cylinder operated like a conventional diesel engine at low load, thus motoring the first cylinder for achieving HCCI combustion in this cylinder. Table 1 Detailed Engine Specifications. Engine characteristics

Specification

Make/model Injection type Number of cylinders Bore/stroke Power per cylinder Compression ratio Total displacement Fuel injection timing Fuel injection pressure

Indec/PH2 Direct injection Two 87.3/110 mm 4.85 kW @ 1500 rpm 16.5 1318 cc 24° before TDC 210 kg/cm2 @ 1500 rpm

Q ROHR SI TDC T V

heat release rate of heat release spark ignition top dead centre mean gas temperature volume of the cylinder

Greek letters k relative air-to-fuel ratio h crank angle c ratio of specific heats r standard deviation

The fuel (ethanol) was injected into the intake manifold using an electronically controlled fuel injector. The fuel delivery system consists of a solenoid based fuel injector and an injection timing and injection duration control circuit. Fresh air entering the engine is heated by an electric air pre-heater positioned upstream of the intake manifold. The intake air heater is operated by a closed loop controller, which maintains constant intake air temperature at the entry to the intake manifold. The schematic diagram of the experimental setup is shown in Fig. 1. A thermocouple in conjunction with a digital temperature indicator was used to measure the intake and exhaust gas temperatures. An orifice meter and a U-tube manometer were used to measure the air consumption of the engine. A surge tank fixed on the inlet side of the engine maintains a constant airflow through the orifice meter and dampens cyclic fluctuations. The in-cylinder pressure was measured using a water-cooled piezo-electric pressure transducer (Make: Kistler, Switzerland; Model: 6061B) which is mounted flush in the cylinder head. The pressure transducer minimizes thermal shock error by using a double walled diaphragm and integral water cooling system. In-cylinder pressure-crank angle history of 100 consecutive cycles was recorded for each test conditions using a high-speed data acquisition system. To measure the crank angle position, a precision shaft encoder (Make: Encoders India, Model: ENC58/6-720ABZ/5-24 V) is coupled with the engine crank-shaft using a flexible helical coupling. The in-cylinder pressure – crank angle history data acquisition and combustion analysis is performed using a program based on LabVIEW, which is developed at Engine Research Laboratory, IIT Kanpur for this purpose. Experiments were conducted at constant engine speed of 1500 rpm and varying intake air temperatures ranging from 120, 140, and 160 °C at different air–fuel ratios for each intake air temperature. 3. Definitions of combustion parameters To study the cycle-to-cycle variations of typical HCCI combustion and performance characteristics at different engine test conditions, following parameters are analyzed. Pmax: Maximum gas pressure in the cylinder. hP max : Crank angle corresponding to Pmax. ðdP=dhÞmax : Maximum rate of pressure rise. hðdP=dhÞmax : Crank angle corresponding to maximum rate of pressure rise.  Rate of heat release (ROHR) rate: Calculated from the acquired data using the zero dimensional heat release model [31]. Consequently, the main combustion parameters were extracted from the heat release and in-cylinder pressure curves.

   

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

1155

Fig. 1. Schematic of the experimental setup.

ROHR was calculated as







 dQ ðhÞ 1 dPðhÞ c dVðhÞ ¼ þ VðhÞ PðhÞ c1 dh dh dh c1

The following assumptions were made in this calculation. I. The cylinder charge was considered to behave as an ideal gas. II. Distributions of thermodynamic properties inside the combustion chamber were considered to be uniform. III. Dissociation of combustion products was neglected. IV. No variation in the cylinder mass due to blow-by was considered. V. Heat transfer from the cylinder is neglected in this model.  ROHRmax: Maximum rate of heat release in a cycle.  hROHRmax : Crank angle corresponding to ROHRmax  Mean gas temperature: Calculated by assuming uniform temperature within the engine cylinder using ideal gas law [31]. The results are valid between intake valve closing (IVC) and exhaust valve opening (EVO).

TðhÞ ¼

PðhÞVðhÞnðhÞ T IVC PIVC V IVC nIVC

In this calculation, molar ratio is assumed to be unity. 4. Results and discussion To evaluate the cycle-to-cycle variations of HCCI combustion at various test conditions, coefficient of variation (COV) of combustion parameters was found. COV of parameters was calculated using following equation.

COVðxÞ ¼

r

where x ¼

x

 100%

Pn

i¼1 xi =n

and standard deviation (r)

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u n uX r ¼ t ðxi  xÞ2 =ðn  1Þ: ½17 i¼1

Engine operating parameters like intake air temperature, air– fuel ratio, engine speed, properties of test fuel, etc. play an important role in combustion stability and cycle-to-cycle variation in HCCI combustion. In the following section, cycle-to-cycle variation of Pmax, hP max , ðdP=dhÞmax , hðdP=dhÞmax , ROHRmax, hROHRmax and maximum mean gas temperature, under different engine operating conditions are analysed and discussed. 4.1. Cycle-to-cycle variation of maximum pressure at different intake air temperatures In-cylinder pressure measurement is considered a very valuable source of information during the development and calibration stages of the engine. Moreover, many applications of in-cylinder pressure for control and diagnostics of engines can be found in literature. The in-cylinder pressure was measured using a high-precision, water-cooled piezo-electric pressure transducer for all engine operating conditions. The cylinder pressure was recorded for consecutive 100 cycles, with a resolution of 0.5 crank angle degrees. Number of cycles recorded for the investigation of cyclic variation in this study is based on published literature [17,26]. To support the number of cycles chosen, an adaptive method based on the variation of the standard deviation of the pressure signal at each crank angle is used. In-cylinder pressure signal is recorded

1156

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

for 1000 cycles of a typical operating condition and standard deviation of pressure signal is calculated at each crank angle using different number of cycle’s envelope. The difference between the maximum standard deviation (rmax) envelope and minimum standard deviation (rmin) envelope is calculated. The maximum value of difference between standard deviation envelopes is plotted with number of cycles chosen in Fig. 2a. Similar plot is also included for the rate of pressure rise and rate of heat release signals in Fig. 2b and c. It is noticed that after 80 engine cycles, increasing the number of cycles does not diminishes the variation of the standard deviation and maximum difference curve remains almost constant. It suggests that beyond this point, having data for additional cycles does not improve the precision of the average and COV values. However this observation is made at only one engine operating condition. The observations can be slightly different at different engine operating conditions, especially at idling. Therefore the number of cycles for making the observations and analysis of the data has been chosen as 100 cycles of pressure-crank angle history. Similar results can also be observed from rate of pressure rise and rate of heat release signals. After 100 cycles, variation in maximum difference between the maximum and minimum envelope of the standard deviation is not significant as compared to variation before 100 cycles (Fig. 2b and c). Since maximum in-cylinder pressure is a critical mechanical constraint in the engine design, it is important to analyze its variation with different engine operating conditions. Fig. 3 shows the cycle-to-cycle variations of the maximum gas pressure (Pmax) in the cylinder for 100 cycles for each test point (Pmax denotes the mean of 100 cycles). For every plot, highest average maximum pressure corresponds to the operating conditions with richest air–fuel mixture and lowest average maximum pressure corresponds to leanest air–fuel mixture at all intake air temperatures due to advanced ignition timing of rich fuel/air mixture. At 120 °C, the engine could be operated in HCCI mode with air–fuel mixture as rich as k = 2 however mixture leaner than k = 4 could not be used for successful HCCI combustion due to high cyclic variation in IMEP values, which indicates misfiring in the engine. The temperature of the intake air was then increased to 140 °C and the operating window of air–fuel mixtures for HCCI operation was 2.5–5.0. When the intake air temperature was raised to 160 °C, the HCCI mode combustion window was found to be 3.0–5.5. Thus it can be observed that increasing the temperature of intake air, it becomes possible to ignite leaner mixtures in HCCI mode whereas relatively richer mixture tends to knock. It can also be clearly observed from the Fig. 3 that the richer air– fuel mixtures lead to relatively higher COV and as mixture becomes leaner, the COV decreases. This is because richer mixture have tendency to knock, which lead to higher pressure oscillations. Similar trend in variation is found for n-heptane fuel by Lu et al. [26]. At any particular air–fuel ratio (k), the COV of Pmax increases with the increase in inlet air temperature as increasing the intake air temperature increases the tendency to knock [6]. It can also be observed from these figures that for all the test points, the variation in maximum in-cylinder pressure is rather small (COV < 3%). Apart from cycle-to-cycle variations of maximum in-cylinder pressure, it is also important to note the variations in crank angle position, at which, maximum pressure is obtained. Since in HCCI combustion, there is no direct control of start of combustion, ignition is completely controlled by chemical kinetics and is therefore affected by fuel composition, equivalence ratio, and thermodynamic state of the fuel–air mixture [16]. So, it has possibility of cycle-to-cycle variations in position where combustion starts, which in turn affects the position of maximum gas pressure in the cycle.

(a) Observation for pressure V/s crank angle signal

(b) Observation for rate of pressure rise V/s crank angle signal

(c) Observation for rate of heat release V/s crank angle signal Fig. 2. Evolution of the maximum difference between the maximum and minimum envelope of the standard deviation with varying number of cycles for k = 3.0 at intake air temperature 120 °C.

Fig. 3a. Cycle-to-cycle variation of Pmax at intake air temperature 120 °C.

1157

Frequency

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

Fig. 3b. Cycle-to-cycle variation of Pmax at intake air temperature 140 °C.

50 40 30 20 10 0 50 40 30 20 10 0 50 40 30 20 10 0 50 40 30 20 10 0 50 40 30 20 10 0

⎯θ= 5.1

0

CA, λ = 4

⎯θ= 7.98

0

⎯θ= 6.14

0

CA, λ =3.5

CA, λ =3

0

⎯θ = 7.22 CA, λ =2.5

-2

-1

0

1

2

3

⎯θ = 0.5

0

4

6

θ Pmax

5

CA, λ =2 7

8

9

10

Fig. 4a. Frequency distribution of crank angle corresponding to Pmax at intake air temperature 120 °C.

60 40 ⎯θ = 3.71

20

0

CA, λ =5

0 60 40 ⎯θ = 3.49

20

0

CA, λ =4.5

0 60

Fig. 3c. Cycle-to-cycle variation of Pmax at intake air temperature 160 °C.

Frequency

40 20

⎯θ = 4.03

0

⎯θ = 3.85

0

CA, λ =4

0 60 40

CA, λ =3.5

20

Fig. 4 shows the frequency distribution of crank angle corresponding to Pmax. It is essential to have the CAD corresponding to maximum gas cylinder pressure hP max close to top dead center (TDC) of piston for optimum efficiency of the engine. The power generated in the engine cycle is hampered in the both cases either because of too much delay or advance in hP max . The study of variation of this parameter is important to relate variation in power generated in the cycle. It can be seen from these figures that crank angle distribution is concentrated more near average hP max and scattered around average value for the richer mixtures as for richer mixtures in engine tend to knock. As mixture becomes leaner, the average value of the hP max moves to after TDC. It can also be observed from the Fig. 4 that for all test conditions, the maximum repeatability of hP max is less than 60%. The crank angle of maximum pressure during each cycle can be used as an estimation of combustion phasing, based on assumption that the crank angle where the pressure is maximum is in the vicinity of the middle of the combustion (since the combustion rates are fast in HCCI combustion). This is a rough estimate however it has the benefit of not requiring numerous calculations. Using the crank angle at Pmax as an estimation of combustion phasing fulfills our requirement of faster calculations. However, this approach works only when there is a global maxi-

0 60 40

0

⎯θ = 2.94 CA, λ =3

20 0 60 40

⎯θ = -2.40

20

0

CA, λ =2.5

0 -5

-4

-3

-2

-1

0

θ Pmax

1

2

3

4

5

Fig. 4b. Frequency distribution of crank angle corresponding to Pmax at intake air temperature 140 °C.

mum pressure due to combustion. It is also noted from Fig. 4 that cyclic variation of CAD at Pmax for richer mixtures is higher for each intake air temperature. Due to all these properties, CAD at Pmax is less effective combustion phasing feedback alternative in order to control HCCI combustion for richer mixtures.

1158

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

60 40 ⎯θ = 2.13

20

0

CA, λ =5.5

0 60 40 20

0

⎯θ = 1.90

CA, λ =5

Frequency

0 60 40 20

0

⎯θ = 1.66

CA, λ =4.5

0 60 40 ⎯θ = 1.69

20

0

CA, λ =4

0 60 Fig. 5a. Cycle-to-cycle variation of ðdP=dhÞmax at intake air temperature 120 °C.

40 0

⎯θ = 1.35 CA, λ =3.5

20 0 60 40

⎯θ = -0.11

20

0

CA, λ =3

0 -3

-2

-1

0

θ Pmax

1

2

3

4

Fig. 4c. Frequency distribution of crank angle corresponding to Pmax at intake air temperature 160 °C.

4.2. Cycle-to-cycle variation of peak rate of pressure rise at different intake air temperatures Rate of pressure rise is related to knock intensity and combustion noise generated in the engine. When the fuelling rates are increased (i.e. lower k), the HCCI combustion rates increase and intensify, and gradually cause unacceptable noise and may potentially cause engine damage, and also eventually lead to unacceptably high levels of NOx emissions. Therefore knocking combustion is often used to define the upper limit of HCCI combustion. Figs. 5 and 6 show cycle-to-cycle variations of maximum rate of pressure rise in the cycle and statistical analysis of crank angle degree corresponding to maximum rate of pressure rise for 100 consecutive combustion cycles at all engine test points. It can be seen from these figures that average rate of pressure rise is very high for richer fuel–air mixtures and is rather lower for leaner fuel–air mixtures due to advanced ignition timing for richer mixtures. For any air–fuel ratio, average rate of pressure rise increases with increase in intake air temperature of the engine. Figures shows that COV of ðdP=dhÞmax is larger corresponding to the richer fuel–air mixtures at constant intake air temperature due to higher pressure oscillations as engine starts knocking. Similar trend in results is found for n-heptane HCCI combustion in a single cylinder engine [26]. It is noticed from the Fig. 5 that variation in the value of ðdP=dhÞmax is less than 10% for all test conditions. But the average value of maximum pressure rise rate is very high for each intake air temperature in case of richer mixtures. Andreae et al. investigated the HCCI engine knock and their results suggest that the threshold value of rate of pressure rise for knock limited operating range is 5 MPa/ms [32]. This value corresponds to 5.5 bar/CAD for engine operating at 1500 rpm. It can be noticed from the Fig. 5 for the average value of maximum rate of pressure rise of 5.5 bar/CAD the COV values less than 6% for all operating conditions. Hence it can be concluded that for the determination

Fig. 5b. Cycle-to-cycle variation of ðdP=dhÞmax at intake air temperature 140 °C.

Fig. 5c. Cycle-to-cycle variation of ðdP=dhÞmax at intake air temperature 160 °C.

of upper boundary of HCCI region, the average value of ðdP=dhÞmax is a better parameter compared to COV of ðdP=dhÞmax .

1159

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

60

60 40

40

0

⎯θ = -1.605

20

CA, λ =4.5

CA, λ =5.5

0 60

0 60

40

40 ⎯θ = 2.185

20

0

⎯θ = -7.31

20

CA, λ =3.5

40

⎯θ = 2.35

0

Frequency

40 20

0

CA, λ =5

0 60

0 60

Frequency

0

⎯θ = --7.23

20

CA, λ =3

0 60

⎯θ = -7.24

20

0

CA, λ =4.5

0 60 ⎯θ = -5.95

40

0

CA, λ =4

20

40

⎯θ = 2.77

0

0 60

CA, λ =2.5

20

⎯θ = -6.14

40

0 60

0

CA, λ =3.5

20

40 ⎯θ = -2.15

20

0

0 60

CA, λ =2

40

0 -4

-3

-2

-1

0

θ (dP/dθ)max

1

2

3

4

⎯θ = -7.85

20

CA, λ =3

0 -9

Fig. 6a. Frequency distribution of hðdP=dhÞmax at intake air temperature 120 °C.

0

-8

-7

-6

θ (dP/d θ )max

-5

-4

Fig. 6c. Frequency distribution of hðdP=dhÞmax at intake air temperature 160 °C. 60 0

⎯θ = -3.29

40

test points at intake air temperatures 140 °C and 160 °C. It can also be observed that maximum repeatability of hðdP=dhÞmax is less than 60% for all engine operating conditions. Since it is calculated from derivative of pressure signals, it is susceptible to noise in the signals hence it is not a robust feedback candidate for closed loop control.

CA, λ =5

20 0 60 40

⎯θ = -4.14

20

0

CA, λ =4.5

4.3. Cycle-to-cycle variation of maximum rate of heat release at different intake air temperatures

0 60

Frequency

40 20

⎯θ = -3.02

0

CA, λ =4

0 60 40 20

⎯θ = -1.215

0

⎯θ = -1.89

0

CA, λ =3.5

0 60 40 20

CA, λ =3

0 60 40 ⎯θ = -6.06

20

0

CA, λ =2.5

0 -7

-6

-5

-4

-3

θ (dP/dθ )max

-2

-1

0

Fig. 6b. Frequency distribution of hðdP=dhÞmax at intake air temperature 140 °C.

It can be seen from Fig. 6, that CAD for maximum rate of pressure rise is concentrated near the average hðdP=dhÞmax . Average CAD for maximum pressure rise advances with the increase in intake air temperature due to advanced ignition timing. It can be observed from the Fig. 6 that average hðdP=dhÞmax is before TDC for all

Figs. 7 and 8 show the cycle-to-cycle variation of peak value of ROHR and corresponding crank angle for maximum ROHR at all engine test conditions. ROHR is measure of how fast chemical energy of fuel is converted to the thermal energy by process of combustion. This directly affects rate of pressure rise and accordingly the power generated in a cycle. It is necessary that cycle-to-cycle variation of ROHR is within optimum limit for smooth engine operating conditions. For all plots, average value of ROHRmax was highest corresponds to engine operation with the richest mixture, and the lowest corresponding to the leanest mixture at any given intake air temperature. Similar trends are also observed by Lu et al. [26]. It can also be noticed from the figures that average value of ROHRmax increases with increase in intake air temperature for any given air– fuel ratio. Fig. 7 shows that COV of ROHRmax is lowest for the richest fuel–air mixture and increases with increase in value of relative air fuel ratio (k). It is observed that cycle-to-cycle variation of ROHRmax is less than 8% for all operating conditions. This value is significantly higher than the maximum value of COV for the maximum pressure Pmax (Fig. 3) because calculation of rate of heat release depends on the rate of pressure rise and the cylinder pressure. The rate of pressure rise has higher variation in COV (Fig. 5) hence the values of ROHRmax demonstrates even higher COV. Fig. 8 shows statistical analysis of CAD of ROHRmax. It can be observed from these figures that average value of CAD for ROHRmax advances with increase in intake air temperature.

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

Frequency

1160

Fig. 7a. Cycle-to-cycle variation of ROHRmax at intake air temperature 120 °C.

70 60 50 40 30 20 10 0 70 60 50 40 30 20 10 0 70 60 50 40 30 20 10 0 70 60 50 40 30 20 10 0 70 60 50 40 30 20 10 0

⎯θ = 2.39, λ = 4

⎯θ = 4.21, λ = 3.5

⎯θ = 4, λ = 3

⎯θ = 3.55, λ = 2.5

⎯θ = -1.67, λ = 2 -3

-2

-1

0 1 2 ⎯θ ROHRmax

3

4

5

Fig. 8a. Frequency distribution of CAD ROHRmax at intake air temperature 120 °C.

60 40

⎯θ = -1.87, λ = 5

20 0 60 40

⎯θ = -2.78, λ = 4.5

20 0 60

Fig. 7b. Cycle-to-cycle variation of ROHRmax at intake air temperature 140 °C.

Frequency

40 20

⎯θ = -1.13, λ = 4

0 60 40 20

⎯θ = -1.31, λ = 3.5

0 60 40 20

⎯θ = -1.54, λ = 3

0 60 40 20

⎯θ = -5.76, λ = 2.5

0 -7

-6

-5

-4

-3

-2

-1

0

⎯θ ROHRmax Fig. 8b. Frequency distribution of CAD ROHRmax at intake air temperature 140 °C.

Fig. 7c. Cycle-to-cycle variation of ROHRmax at intake air temperature 160 °C.

It can be noticed from Fig. 8 that average value of CAD ROHRmax is before TDC for all test conditions at intake air temperature of 140

and 160 °C. Figures also show that the crank angle location of ROHRmax is concentrated in a very close range to the peak value of CAD ROHRmax and its spread is very limited. It is observed that repeatability of same CAD of maximum ROHR is up to 70%. The CAD for ROHRmax can also be used for an estimation of combustion phasing but it has the drawback of being more noise sensitive. The benefit is that it provides a fast estimate because it only requires pressure measurements until the maximal slope has been reached. As can be seen from the Fig. 8, the maximum rate of heat release

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

1161

60 40 20

⎯θ = -6.29, λ = 5.5

0 60 40

⎯θ = -6.39, λ = 5

20 0 60

Frequency

40 20

⎯θ = -6.0, λ = 4.5

0 60 40 20

⎯θ = -4.98, λ = 4

0 60 40 20

⎯θ = -6.1, λ = 3.5

Fig. 9a. Cycle-to-cycle variation of IMEP at intake air temperature 120 °C.

0 60 40 20

⎯θ = -7.53, λ = 3

0 -9

-8

-7 -6 ⎯θ ROHR max

-5

-4

Fig. 8c. Frequency distribution of CAD ROHRmax at intake air temperature 160 °C.

occurs at the same crank angle up to 60% observation points (Fig. 8). Considering these facts, it can be used as feedback for combustion phasing in control of HCCI engine.

4.4. Cycle-to-cycle variation of IMEP at different intake air temperatures It is important to investigate the cycle-to-cycle variation of IMEP because it directly affects the engine drivability. In the literature [33], it is reported that drivability problems in automobiles normally arise when COVIMEP exceeds 10%, hence this parameter can be used for investigation of the lower boundary for HCCI combustion. One major limitation of HCCI combustion is the requirement of a highly diluted mixture in order to slow down the speed of the chemical reactions sufficiently so that engine is not damaged. With lean operation, this will significantly reduce the output for a given air flow through the engine. The rich side limit for IMEP is limited by the rate of combustion and hence that of rate of pressure rise. Fig. 9 illustrate the cycle-to-cycle variations of IMEP at all test point for 100 consecutive combustion cycles. The maximum average value of IMEP encountered in this investigation is 4.3 bar. It is observed that the average value of IMEP is lower for leaner mixtures. It can be noticed from Fig. 9 that COV of IMEP increases with increase in k value (leaner mixture). The COV is lowest for richest fuel–air mixture at any intake air temperature. At any air–fuel ratio, the coefficient of variation decrease with increase in intake air temperature. These figures show that COV of IMEP exceeds 10% for some of the test conditions. According to published literature [33], the limit of COV of IMEP for drivability constraints is less than 10% for SI engine. HCCI is known for its low COV of IMEP compared to SI engines however with a late combustion timing, the COV of IMEP increases rapidly. At higher engine speeds, the time window for proper combustion phasing gets reduced, and therefore it is more difficult to retard the combustion. Researchers found that there is less risk for misfire when the COV is kept under 3.5% [34]. Hence

Fig. 9b. Cycle-to-cycle variation of IMEP at intake air temperature 140 °C.

Fig. 9c. Cycle-to-cycle variation of IMEP at intake air temperature 160 °C.

1162

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

4.5. Cycle-to-cycle variation of maximum in-cylinder temperature at different intake air temperatures

Fig. 10a. Cycle-to-cycle variation of Tmax at intake air temperature 120 °C.

It is worth examining the cycle-to-cycle variation of mean gas temperature inside the cylinder since it is directly related to emissions from the engine in case of HCCI combustion. With homogeneous combustion of a premixed charge, the temperature is expected to be same throughout the combustion chamber, except near the walls. This, in combination with very lean fuel–air mixtures gives low maximum temperature during the cycle. NOx formation is very sensitive to the peak temperature and duration of high temperature combustion in an engine cycle. At temperatures above 1800 K, NOx formation rate increases rapidly. Large variations in mean gas temperature gives rise to variation of NOx emissions as well. The cyclic variation in mean gas temperature also affects the variation in heat transfer from cylinder liner. Hence, it is important to investigate the variation in maximum mean gas temperature in the cylinder. Fig. 10 shows cycle-to-cycle variation of the mean gas temperature for 100 consecutive combustion cycles. It can be noticed that the average value of Tmax decreases with increasing value of k (leaner mixtures) at any given intake air temperature. Average Tmax in combustion cycles increases with increase in intake air temperature at any constant equivalence ratio. It can be noticed that COV of Tmax was higher for the richer mixture. As mixture becomes richer, the engine tends to knock and cylinder pressure oscillations increase, which results in higher heat transfer from the liner. The variation in the temperature inside the cylinder leads to higher COV of maximum temperature for the richer mixtures. The value of COV at all test points is less than 5%, hence the statistical variation is rather small in HCCI combustion.

5. Conclusions

Fig. 10b. Cycle-to-cycle variation of Tmax at intake air temperature 140 °C.

Fig. 10c. Cycle-to-cycle variation of Tmax at intake air temperature 160 °C.

the COV of IMEP should be used as criteria for lower boundary of HCCI operating range.

The cycle-to-cycle variations in combustion and performance parameters of HCCI combustion were investigated on a modified two cylinder direct injection diesel engine. The inlet air was supplied at 120, 140 and 160 °C temperature and the engine was operated at a constant engine speed of 1500 rpm, with a port fuel injection of ethanol in HCCI mode. It was found that at lower intake air temperature it is possible to ignite the richer mixture (up to k = 2) in HCCI combustion mode. As intake air temperature increase, engine running on richer mixture tend to knock with very high rate of pressure rise. But at higher intake air temperature it is possible to ignite the leaner mixture (up to k = 5.5) in HCCI combustion mode. Coefficient of variation (COV) of Pmax increases with increasingly richer mixture and COV increases with increase in intake air temperature. For all test points, the variation in maximum incylinder pressure is small (COV < 3%) and repeatability of CAD of Pmax is less than 60%. Average value of pressure rise rate is an important parameter for HCCI operating range criteria in comparison to COV of maximum pressure rise rate. The coefficient of variation in the maximum pressure rise rate is less than 10% for all the test conditions and increases with the increase in inlet air temperature. The average value of maximum rate of pressure rise is more than 10 bar/ CAD for some of the operating conditions. Maximum repeatability of CAD of maximum rate of pressure rise is up to 60% for all engine operating conditions. The COV of ROHRmax is lowest for the richest fuel–air mixture and cycle-to-cycle variation of ROHRmax is less than 8% for all engine test conditions. It is observed that repeatability of same CAD of maximum ROHR is up to 70%.

R.K. Maurya, A.K. Agarwal / Applied Energy 88 (2011) 1153–1163

The COV of IMEP increases with the increasing of k (leaner mixture) and decreases with the increasing intake air temperature. It is a critical parameter for the HCCI combustion engine. The COV of IMEP exceeds 10% for some test conditions. It is a critical parameter for the HCCI combustion engine. The COV of maximum mean gas temperature is less 5% and it is not a critical parameter for HCCI combustion. Hence, the critical optimum that can be used to define the HCCI operating range is the ‘maximum pressure rise rate’ and ‘COV of IMEP’. References [1] Onishi S, Jo SH, Shoda K. Actived thermo-atmosphere combustion (ATAC) – a new combustion process for internal combustion engines, SAE paper 1979-010501; 1979. [2] Noguchi M, Tanaka T, Takeuchi Y. A study on gasoline engine combustion by observation of intermediate reactive products during combustion, SAE paper 1979-01-0840; 1979. [3] Suzuki H, Koike N, Ishii H, Odaka M. Exhaust purification of diesel engines by homogeneous charge with compression ignition part 1: experimental investigation of combustion behavior under premixed homogeneous ignition method, SAE paper 970313; 1997. [4] Ishibashi Y, Asai M. Improving the exhaust emission of two-stroke engines by applying the activated radical combustion, SAE paper 960742; 1996. [5] Dickey DW, Ryan III TW, Matheaus AC. NOx control in heavy-duty diesel engine what is the limit?, SAE paper 980174; 1998. [6] Christensen M, Johansson B, Einewall P. Homogeneous charge compression ignition (HCCI) using isooctane, ethanol and natural gas – a comparison with spark ignition operation, SAE paper 972874, 1997. [7] Suzuki H, Koike N, Odaka M. Combustion control method of homogeneous charge diesel engines, SAE paper 980509; 1998. [8] Zheng J, Yang W, Miller DL, Cernansky NP. A skeletal chemical kinetic model for the hcci combustion process, SAE paper 2002-01-0423; 2002. [9] Machrafi H, Cavadias S, Amouroux J. A parametric study on the emissions from an HCCI alternative combustion engine resulting from the auto-ignition of primary reference fuels. Appl Energy 2008;85:755–64. [10] Komninos NP. Modeling HCCI combustion: modification of a multi-zone model and comparison to experimental results at varying boost pressure. Appl Energy 2009;86:2141–51. [11] Komninos NP. Investigating the importance of mass transfer on the formation of HCCI engine emissions using a multi-zone model. Appl Energy 2009;86:1335–43. [12] Ryan III TW, Callahan TJ. Homogeneous charge compression ignition of diesel fuel, SAE paper 961160; 1996. [13] Hou J, Qiao X, Wang Z, Liu W, Huang Z. Characterization of knocking combustion in HCCI DME engine using wavelet packet transform. Appl Energy 2010;87:1239–46. [14] Kelly-Zion L Peter, Dec JE. A computational study of the effect of the fuel-type on the ignition time in HCCI engines. Proc Combust Inst 2000;28:1187–94. [15] Sjöberg M, Dec JE. Combined effects of fuel-type and engine speed on intake temperature requirements and completeness of bulk-gas reactions for HCCI combustion, SAE paper 2003-01-3173; 2003. [16] Dec JE, Sjöberg M. Isolating the effects of fuel chemistry on combustion phasing in an HCCI engine and the potential of fuel stratification for ignition control, SAE paper 2004-01-0557; 2004.

1163

[17] Xingcai L, Libin J, Junjun M, Zhen H. Experimental study on the cycle-by-cycle variations of homogeneous charge compression ignition combustion using primary reference fuels and their mixtures. Proc Inst Mech Eng Part D: J Automob Eng 2007;221:859–66. [18] Iida M, Hayashi M, Foster DE, Martin JK. Characteristics of homogeneous charge compression ignition (HCCI) engine operation for variations in compression ratio, speed, and intake temperature while using n-butane as a fuel. J Eng Gas Turb Power 2003;125:472–8. [19] Iida M, Aroonsrisopon T, Hayashi M, Foster D, Martin J. The effect of intake air temperature, compression ratio and coolant temperature on the start of heat release in an HCCI (homogeneous charge compression ignition) engine, SAE paper 2001-01-1880; 2001. [20] Huang Z, Shiga S, Ueda T, Nakamura H, Ishima T, Obokata T, et al. Study of cycle-by-cycle variations of natural gas direct injection combustion using a rapid compression machine. Proc Inst Mech Eng Part D: J Automob Eng 2003;217:53–61. [21] Ball JK, Raine RR, Stone CR. Combustion analysis and cycle-by-cycle variations in spark ignition engine combustion – part 1: an evaluation of combustion analysis routines by reference to model data. Proc Inst Mech Eng Part D: J Automob Eng 1998;212:381–99. [22] Ball JK, Raine RR, Stone CR. Combustion analysis and cycle-by-cycle variations in spark ignition engine combustion – part 2: a new parameter for completeness of combustion and its use in modeling cycle-by-cycle variations in combustion. Proc Inst Mech Eng Part D: J Automob Eng 1998;212:381–99. [23] Brown AG, Stone CR, Beckwith P. Cycle-by-cycle variations in spark ignition engine combustion – part I: flame speed and combustion measurements and a simplified turbulent combustion model, SAE paper 960612; 1996. [24] Stone CR, Brown AG, Beckwith P. Cycle-by-cycle variations in spark ignition engine and combustion – part II: modeling of flame kernel heat release analysis on combustion and displacements as a cause of cycle-by-cycle variations, SAE paper 960613; 1996. [25] Sen AK, Litak G, Finney CEA, Daw CS, Wagner RM. Analysis of heat release dynamics in an internal combustion engine using multifractals and wavelets. Appl Energy 2010;87:1736–43. [26] Lu X, Ji L, Ma J, Huang Z. Combustion stabilities and cycle-by-cycle variations of n-heptane homogeneous charge compression ignition combustion. Energy Fuels 2007;21:1468–73. [27] Persson H, Pfeiffer R, Hultqvist A, Johansson B, Ström H. Cylinder-to-cylinder and cycle-to cycle variations at HCCI operation with trapped residuals, SAE paper 2005-01-0130; 2005. [28] Koopmans L, Backlund O, Denbratt I. Cycle to cycle variations: their influence on cycle resolved gas temperature and unburned hydrocarbons from a camless gasoline compression ignition engine, SAE paper 2002-01-0110; 2002. [29] Shi L, Deng K, Cui Y. Combustion stability of diesel-fueled HCCI. Int J Automot Technol 2007;8:395–402. [30] Shahbakhti M, Koch CR. Characterizing the cyclic variability of ignition timing in a homogeneous charge compression ignition engine fuelled with n-heptane/ iso-octane blend fuels. Int J Engine Res 2008;9:361–97. [31] Hi-techniques revelation operator reference manual, Madison, USA: HiTechniques, Inc.; 2004. [32] Andreae MA, Cheng WK, Kenney T, Yang J. On HCCI engine knock, SAE paper 2007-01-1858; 2007. [33] Heywood JB. Internal combustion engine fundamentals. New York: McGrawHill Book Company; 1988. [34] Johansson T, Borgqvist P, Johansson B, Tunestål P. HCCI heat release data for combustion simulation, based on results from a turbocharged multi cylinder engine, SAE paper 2010-01-1490; 2010.