Energy 36 (2011) 3563e3571. Contents lists available at ScienceDirect. Energy. journal homepage:

Transcription

1 Energy 36 (2011) 3563e3571 Contents lists available at ScienceDirect Energy journal homepage: An experimental investigation of high performance natural gas engine with direct injection M.A. Kalam *, H.H. Masjuki Department of Mechanical Engineering, University of Malaya, Kuala Lumpur, Malaysia article info abstract Article history: Received 29 July 2010 Received in revised form 18 February 2011 Accepted 23 March 2011 Available online 22 April 2011 Keywords: DI Engine CNG-BI engine Power SFC Emissions This paper presents experimental results of a new compressed natural gas direct injection (CNG-DI) engine that has been developed from modification of a multi cylinder gasoline port injection (PI) engine. The original gasoline-pi engine was also modified to a CNG bi-fuel system. The test results obtained from CNG fuel using two different systems (i.e. bi-fuel and DI) have been investigated and compared with the original gasoline engine. The objective of this investigation is to compare the test results between CNG-DI, with CNG-BI and gasoline-pi engines with the same displacement volume. It was found that the CNG-DI engine produces similar brake power at 6000 rpm and wide open throttle (WOT) but produces higher brake power at part load condition as compared to the original gasoline. The CNG-BI engine produces 23% lower brake power than the CNG-DI engine. The average brake specific fuel consumption (BSFC) of the CNG-DI engine was 0.28% and 8% lower than gasoline-pi and CNG-BI engines respectively. The CNG-DI engine reduces 42% NOx emission as compared to the base engine. However, the CNG-DI engine produces higher HC and CO emissions as compared to the base engine. This paper discusses a review on the direct injection (DI) natural gas engine with new information along with other investigations. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction High performance compressed natural gas direct injection (CNG-DI) engine development has now become a challenging and innovative technology. In particular, automotive engine researchers have sought this technology to improve engine efficiency with natural gas fuel to meet stringent emission limits. This innovative development will reduce emissions to limit the negative impact of the greenhouse effect. This investigation is related to in-cylinder direct injection based on Otto cycle and with an accelerating effort to design and develop better efficient engines while researchers have devoted significant resources to developing a CNG-DI engine. It is believed that the CNG-DI engine has great potential to optimise fuel supply and combustion, which in turn can deliver better performance and lower fuel consumption. Many researchers have conducted works Abbreviations: CNG-DI, compressed natural gas direct injection; CNG-BI, compressed natural gas bi fuel; Gasoline-PI, gasoline engine with port fuel injection; AFR, air fuel ratio; CAS, combustion analysis system; ECU, electronic control unit; MBT, maximum best torque; TWC, three way catalytic converter; MPI, multiport injection; NG, natural gas; NOx, oxides of nitrogen; CO, carbon monoxide; HC, unburn hydrocarbon; WOT, wide open throttle. * Corresponding author. Tel.: þ ; fax: þ address: kalam@um.edu.my (M.A. Kalam). on a CNG-DI system for diesel engines as in Refs. [1e6]. It is hoping that the output of this investigation and developing capabilities for advanced CNG-DI engine using gasoline cycle with spark ignition system will be a realisation of engineering dreams Reason for CNG-DI engine In conventional fuel injection systems natural gas is injected into engine cylinders either by a mixer, single-point injection or multi-point injection with electric motors. With all these injection systems, the natural gas engine produces lower brake power as compared to gasoline fuel. Hence, the CNG-DI engine system is more suitable where the fuel is injected through a high pressure pipe line straight into the cylinder with the required amount to produce similar or higher brake power than a gasoline engine. With the recent increase of oil price and limited reservation, it becomes necessary to accelerate the use of natural gas (NG) especially for the automotive sector. Therefore, new technologies encompassing fuel systems, combustion chambers, control units, vehicle bodies, fuel storage and refuelling infrastructure also need to be investigated. The usage of NG in the transportation sector is increasing due to stringent emission regulations and limited fossil fuel reservation. Presently the manufacturer is involved in conversion of gasoline /$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi: /j.energy

2 3564 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e3571 Otto cycle Diesel cycle Pre-Mixed Direct Injection Mixer MPI CNG-DI Dual-Fuel Mono-fuel Spark plug ignition Diesel Oil Pilot Glow plug ignition Fig. 1. Various combustion systems for CNG fuel. engine into bi-fuel gasoline/cng fuels. This bi-fuel system is not a sustainable solution as it produces low brake power and higher exhaust emissions. The reasons of producing low brake power are slow burning velocity, poor lean burn ability that lead to incomplete combustion, high misfire ratio and large cycle by cycle variation at lean mixture combustion [7,8]. In addition, volumetric efficiency and less gas injections are also responsible for producing low brake power. To overcome these problems, a sustainable development for long term usage of NG, a new high performance direct injection (DI) engine is being proposed through this investigation Fuel injection system classification Many car companies have proposed and developed dedicated NG engines during the last fifteen years and most of them are multiport port injection (MPI) system, where the engine thermal efficiency is low and three way catalytic (TWC) is utilised to reduce emissions. However, some researchers like Westport Innovations Inc. and ISUZU car company (Japan) have proposed and developed a CNG-DI engine based on diesel cycle combustion system [1e3]. It was proposed [1e3] that NG, DI and shielded glow plug ignition with hot surface system mounted on cylinder head would improve engine efficiency. However, based on the ISUZU CNG-DI engine, an attempt was undertaken to produce a dedicated NG engine to replace diesel fuel. In this investigation, an attempt is taken to produce a dedicated NG engine to replace gasoline fuel. The following Fig. 1 shows various combustion systems for CNG fuel. From the figure, it can be explained that each combustion system has unique features to reflect specific strategies of mixture preparation, combustion control and emissions reduction. However, all systems have a common goal of achieving substantial fuel economy improvement while simultaneously achieving large reductions in engine output and tailpipe emissions. Table 1 shows the recent study on CNG-DI engines. It can be seen that most of the studies are based on diesel engines Objectives of this study The main objectives of this investigation are: 1. To experimentally investigate the performance and emissions characteristics of a newly developed CNG-DI engine under various test conditions. 2. To study on benchmarking between CNG-DI engine with gasoline port injection (gasoline-pi) and CNG bi-fuel (CNG-BI) engines when the displacement volume is the same for all the cases. 2. Procedure A schematic diagram of the experimental setup is shown in Fig. 2. A total of three engines were tested and their specifications have been shown in Table 2. The CNG-DI engine was developed Table 1 Recent investigation on DI natural gas engine. Ref. Original engine Modified engine Injector/injection/ignition Results [7,9] Single cylinder diesel engine [10] Rapid Compression Machine [11] Single cylinder diesel engine [12] Single cylinder diesel engine [13] Single cylinder diesel engine Direct injection spark ignition engine for natural gas-hydrogen fuels. Used as direct injection SI engine for natural gas fuel. Direct injection dual fuel diesel-natural gas engine. Natural gas direct injection engine (Pilot injection) Direct injection natural gas engine Injection start from 170 to 240 CA BTDC. Ignition angle from 28 to 42 CA BTDC. CR 12, Fuel injection pressure 80 bar. Constant ignition timing as 80 ms from start of compression and constant fuel injection pressure as 90 bar. Injector installed at intake manifold. CR 17.1, at 2000 rpm, SOI was at 39 deg BTDC for pilot fuel and 11 deg BTDC for main fuel as CNG. Westport Innovation Inc., dual-fuel concentric needle. Modified injector, constant ignition at 32 deg BTDC. Injection pressure 80 bar. Injection timing 180 deg BTDC. Increasing hydrogen fraction (max 20%) in blends increases brake effective thermal efficiency and NOx emissions, but decreases HC and CO emissions. Minimum equivalence ratio obtained as F ¼ Injectors and injectors location does not affect on combustion and emissions. Results depend on pilot fuel amount together with the rpm and load conditions. Increasing pilot fuel amount increases rate of combustion during early phase of combustion. Emissions reduced except black carbon. Fuel injection at 180 deg BTDC was better for low emissions such as HC and CO and higher thermal efficiency.

3 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e Silencer Calorimeter Emission Analyzers Air Intake Exhaust Gas Air Flow Meter Water in Water out Water valve Oil Valve Compressor Heat Exchanger Gasoline FMS FMS-Fuel measuring system. ECU-Electronic control unit. CNG-Compressed natural gas. CAS- Combustion analysis system. Shaft Encoder Pulse Counter Dynamometer CAS Switch Box Engine ECU CNG Gas Flow Meter System Gas main supply Data logger Fig. 2. Schematic diagram of the experimental set-up. through modification of a gasoline engine (gasoline-pi engine). The major modifications done were e (1) Increasing compression ratio from 10 to 14 through modifying piston and cylinder head, (2) new spark plugs with long edge were used to ignite the CNG fuel and (3) fuel injection system was modified from MPI to DI system. The CNG injection pressure was 20 bars at the common rail. The temperature of CNG at the common rail was a constant 16 C. The injector was designed to inject CNG fuel into the engine cylinder. The injector was initially set with a spring preload of 38 N. The spring preload was then adjusted with 1 N to trim the dynamic flow at 100 Hz with 2.0 ms pulse width. The average stroke length, dynamic flow rate, opening and closing times are mm, mg/shot, 1.50 ms and 0.93 ms respectively. An eddy current dynamometer with maximum absorption power of 150 kw was used to maintain the variation of loads at different engine speeds. The dynamometer could be started, loaded and monitored via remote operation of the control-instrumentation unit and data acquisition control system. The dynamometer was also equipped with a speed sensor, switches for low pressure and high temperature for cooling water, the drive shaft, water inlet valve and load cell torque measurement unit. Table 2 Test engines specifications. Item Gasoline-PI a CNG-BI CNG-DI Bore stroke (mm) Displacement (cc) Number of cylinder Compression ratio Combustion chamber Bowl Bowl Bowl IVO (BTDC) IVC (ABDC) EVO (BBDC) EVC (ATDC) Fuel system MPI Bi-fuel CNG-DI Rated power (kw/rpm) 82/ / /6000 Rated torque (Nm/rpm) 148/ / /4000 Fuel pressure (bar) Valve train and cylinder configuration DOHC 16V 4 cylinders in-line a The base engine for CNG-DI and CNG-BI engine. DOHC 16V 4 cylinders in-line DOHC 16V 4 cylinders in-line The air flow rate into the engine inlet manifold was measured by a hot-wire anemometer (accuracy 0.2%) which comes with the engine. A hot-wire anemometer keeps the temperature of a thin wire constant by adjusting the current flow through the wire. The current required to keep the temperature constant depends on the convective heat transfer, which depends on the mass airflow past the wire. This air mass flow meter data is transferred to an analogue input card through a signal cable of 0e5 volts. Finally, the actual airflow into the engine was analysed from the data logged (Cadet 12 engine controlled software) into the computer. A Coriolis micro motion mass flow meter was used to measure CNG flow rate into the engine. The water and lubricant temperatures were controlled at 80 C and 90 C respectively. An HORIBA exhaust gas analyser was used to measure emissions concentration for the CNG-DI engine. This analyser was interfaced with main engine controlled software (CADET12), so that all the emissions data and engine operating data can be logged at the same time for analysis. These analysers consist of individual modules of each emission parameter and have zero and span gas calibration facility. The measurement technique of the analyser is infrared for CO, CO 2 and HC while chemiluminescent for NOx emissions. Detailed working principles can be seen in the HORIBA website. A conventional and standard lambda meter was used to measure exhaust air fuel ratio to be tuned up by electronic control unit (ECU). It accurately determines the exhaust gas mixture Table 3 Natural gas compositions. Component Mole (%) Methane Ethane 2.29 Propane 0.03 Isobutane 0.23 Normal-butane 0.02 isopentane 0.01 Hexane 0.01 Carbon dioxide 0.57 Nitrogen 0.44 Others e

4 3566 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e3571 Table 4 Physicochemical properties of CNG and gasoline fuels. Properties CNG Gasoline Density (kg/m 3 ) 0.81 e Gross calorific value (MJ/kg) Molecular weight Specific gravity 0.64 (compared to air) (compared to water) strength over a wide range of engine operating conditions with a fast response time. The operating range of the used device is between 0.70 and lambda and the air fuel ratio range of a typical spark ignition engine is about 10 to 22 (which is within the measurable range of the lambda meter). Hence, for CNG-DI engine development, the used lambda meter was good enough to tune up the engine configuration to achieve maximum best torque (MBT). The used lambda meter accurately determines only one mixture strength to achieve best performance. The combustion analysis system (CAS) includes control software, encoder and pressure sensors [14]. Other sensors (a total of nine thermocouples and six pressure sensors) were installed into the engine test bed to measure temperature and pressure at various test points. The instrument used in this investigation was fully equipped in accordance with SAE standard J1349 JUN90 (ref. SAE Handbook 2002). All the engines were tested from 1500 rpm to 6000 rpm with wide open throttle (WOT) condition for comparisons purposes Fuels used in this investigation The composition of an NG fuel varies with location, climate and other factors. It is anticipated that such changes in fuel properties affect emission characteristics and performance of CNG fuel in engines as shown by [14,15]. The physicochemical properties of CNG and gasoline fuels used in this experiment are shown in Tables 3 and 4 respectively. The lube oil used was ordinary commercial lube oil (SAE 40). 3. Result and discussions The air temperature inside the laboratory room was constant at 25 C. The CNG-DI engine did not have any initial starting difficulties due to fuel ignited by a spark plug. In this investigation, a total of three engines have been tested such as (1) gasoline-pi gasoline fuel with port injection system engine, (2) CNG-BI compressed natural gas fuelled engine with bi-fuel injection system, and (3) CNG-DI compressed natural gas engine with direct injection system. These three engines have the same cylinder volume i.e. 1.6 l. The results shown in this paper are obtained from WOT with variable speed conditions. All the equipment were properly calibrated by local expert (Hakita Engineering co.ltd). The accuracy level of all the devices were listed in a table and presented in our recent published paper in Journal of Energy [16]. However, the errors associated with the equipment accuracy level were minimised with the measured data. The experimental test was conducted three times repeatedly and the variation of test data was calculated to measure the uncertainty level. A sample calculation has been provided in Appendix. The uncertainty level was below 4% and the polynomial trend line curve was well fitted for presentation and discussion Brake power at WOT Fig. 3(A) shows brake power versus engine speed from 1500 rpm to 6000 rpm for all the test engines such as gasoline-pi, CNG-BI and CNG-DI engines at WOT. The gasoline-pi and CNG-DI produce maximum brake power at 6000 rpm which are kw and kw respectively. However, the CNG-BI produces maximum brake power at 5500 rpm which is kw (23% lower than CNG-DI engine). The average brake power over the test cycle obtained was kw, kw and kw by the gasoline-pi, CNG-BI and CNG-DI engines respectively. The CNG-DI engine produces 2.83 kw (4%) higher brake power at 6000 rpm but on average all over the engine speed range it is 2.02 kw less brake power as compared to the base gasoline-pi engine. At 6000 rpm, the CNG-DI engine produces higher brake power which can be confirmed through Fig. 4 as torque curve. Further it can be confirmed through Fig. 3(B) which shows the brake power at part load condition; it is found that after 5000 rpm the CNG-DI engine exceeds the brake power of the gasoline-pi engine. From Fig. 3 it is well understood that at higher engine speeds as above 5000 rpm, the CNG-DI engine produces higher brake power due to efficient NG fuel conversion efficiency which was a major deficiency in the case of the CNG-BI engine. However, from 1500 rpm to 5000 rpm, the CNG-DI engine shows lower level of brake power as compared to the gasoline-pi engine. The reason of producing lower brake power from CNG-DI engine (from 1500 to 5000 rpm) is mainly due to producing lower brake torque which is strongly related to volumetric efficiency, gas inlet temperature, gas mixture distribution, air fuel ratio (AFR) as well as effective cylinder pressure. In comparison with the CNG-BI engine, on average all over the speed range, the CNG-DI engine produces 22.95% higher brake power than the CNG-BI engine. Fig. 3. A. Brake power versus engine speed at WOT. B. Brake power versus engine speed at 50% throttle.

5 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e Brake torque at WOT Fig. 4 shows brake torque versus engine speed from 1500 rpm to 6000 rpm for all the test engines such as gasoline-pi, CNG-BI and CNG-DI engines at WOT. It is found that gasoline-pi, CNG-BI and CNG-DI produced their maximum torque are Nm (at 4500 rpm), 100 Nm (at 4500 rpm) and Nm (at 5500 rpm) respectively. The average brake torque over the engine speed range for gasoline-pi, CNG-BI and CNG-DI engines obtained are Nm and Nm and Nm, respectively. The reason of producing lower brake torque by the CNG-DI engine is mainly due to lack of chemical energy conversion to mechanical energy which is strongly related to volumetric efficiency, fuel mixing, net heat release rate as well as cylinder pressure. Improper cylinder pressure such as too high or too low cylinder pressure causes lower brake torque. However, the CNG-BI shows the lowest level of brake torque production as compared to CNG-DI and gasoline-pi systems Brake specific fuel consumption at WOT Fig. 5 shows the variation of brake specific fuel consumption (BSFC) versus engine speed for all the test engines from 1500 rpm to 6000 rpm at WOT. It can be seen that the BSFC increases initially at 1500 rpm for all the engines due to an increase in magnitude of friction, pumping work and the increased relative importance of friction and heat transfer, which decreases the gross indicated fuel conversion efficiency [17]. It is found that the gasoline-pi engine reduces BSFC from 1500 rpm to 3500 rpm due to increasing fuel conversion efficiency and then started to increase SFC due to increasing frictional effect with increasing engine speed. However, the average BSFC of the CNG-DI engine is lower than the gasoline-pi as well as CNG-BI engines. The lowest BSFC ( g/kwh) comes from the CNG-DI engine at 3500 rpm followed by gasoline-pi ( g/kwh@3500 rpm) and CNG-BI ( g/kwh@3500 rpm) engines. The average BSFCs over the test cycle for CNG-DI, CNG-BI and gasoline-pi engines are g/kwh, g/kwh and 264 g/kwh respectively Exhaust emissions Unburned hydrocarbon at WOT Unburned hydrocarbon or partially oxidised hydrocarbon emissions increase if (a) the injection occurs too early, in which case the delay time increases with the result that more fuel goes to Fig. 5. Brake specific fuel consumption versus engine speed at WOT. contact at the relatively cool cylinder wall, or (b) injection too late in which case there may be insufficient time for completion of combustion. The latter case may be matched with the CNG-DI engine as the DI cooled gas enters the engine cylinder, which is the main reason for the increase of HC emission as compared to the gasoline-pi engine. It is found that however, the maximum level of HC is produced by the CNG-BI engine followed by CNG-DI and gasoline-pi engines (Fig. 6). The average HC emissions over the engine speed range were 0.35, 0.28 and 0.78 g/kw h by CNG-DI, gasoline-pi and CNG-BI respectively. The CNG-DI engine produces slightly higher HC (by 25%) than the base gasoline-pi engine and 56% lower than the CNG-BI engine Oxides of nitrogen at WOT The main cause for the increase of NOx is high combustion temperature [18]. The NOx concentration versus engine speed is illustrated in Fig. 7. It was found that the lowest NOx was produced by CNG-BI (average 4.85 g/kw h) followed by the CNG-DI (6.65 g/ kw h) and gasoline-pi (11.44 g/kw h) engines. It is very interesting that the CNG-DI reduces (42%) NOx emissions as compared to the base gasoline-pi engine. This is mainly due to cool gas entering into the engine cylinder, so that the overall combustion is completed at a low temperature. The CNG temperature at common rail is 16 C, and the intake temperature is about 35 C which gives lower combustion temperature, hence the NOx reduces. The maximum NOx at g/kw h was produced by the gasoline-pi engine at 6000 rpm. The CNG-DI engine produces maximum NOx emission (9.39 g/kw h) at 6000 rpm and the overall NOx emissions level is Fig. 4. Brake torque versus engine speed at WOT. Fig. 6. Unburned hydrocarbon versus engine speed at WOT.

6 3568 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e3571 Fig. 7. Oxides of nitrogen versus engine speed at WOT. Fig. 9. Carbon dioxide versus engine speed at WOT. lower than the gasoline-pi engine by 4.79 g/kw h. Hence, it is an important finding that the modification from gasoline/mpi system to CNG-DI system reduces NOx emissions Carbon monoxide at WOT Carbon monoxide (CO) is formed during the combustion process with rich fuel-air mixtures and when there is insufficient oxygen to fully burn all the carbon in the fuel to CO 2.AsCOis strongly related to rich fuel-air mixtures, hence spark ignition engines are the significant source for CO emission, because they use stoichiometric or close to stoichiometric air fuel ratio which may divide into a fuel rich zone and a fuel lean zone in the cylinder during combustion. The rich zone increases CO emission. Hence, increasing CO emission is referred to as incomplete combustion of fuel. It is found that both the gas engines such as CNG-DI and CNG-BI engines increase CO with increasing engine speed (Fig. 8), while the gasoline-pi decreases CO with increasing engine speed. For the CNG-DI engine, the low combustion temperature is the main reason to increase CO emission. On average over the engine speed range, it is found that the CO produced is 92.88, and g/kw h by CNG-DI, CNG-BI and gasoline-pi respectively Carbon dioxide at WOT Carbon dioxide (CO 2 ) is directly related to the combustion of fuel. Increasing CO 2 means better fuel combustion. However, there is a limitation as for gasoline fuel the maximum CO 2 should be below 12.62% (this value comes from stoichiometric calculation). If the CO 2 concentration increases more than this (12.62%) then it will affect directly on NOx concentration such that NOx will increase. Similarly for CNG fuel, the maximum CO 2 will be equal or below 9.60% (at complete combustion). The CO 2 emission in g/kw h is shown in Fig. 9. On average all over the engine speed range, the CO 2 produced are 6.67, and g/kw h by CNG-DI, CNG-BI and gasoline-pi respectively. The CNG-DI engine produces the highest CO and lowest CO 2 emissions Exhaust gas temperature Fig. 10 shows exhaust gas temperature versus engine speed range from 1500 rpm to 6000 rpm at WOT condition. It is found that increasing engine speed increases exhaust gas temperature mainly due to increasing utilisation of fuel per unit time. Normally exhaust temperatures above 800 C indicate the higher combustion temperature inside the engine cylinder which increases NOx emissions and it happened in the gasoline-pi engine. At this engine test condition (such as at WOT) the maximum average temperature was produced by gasoline-pi engine (817 C ) followed by CNG-BI (725 C) and CNG-DI (697 C) engines. Hence, lowering exhaust gas temperature gives an indication as combustion occurred with lower temperature as compared to the gasoline-pi engine Oxygen level at WOT Fig. 11 shows oxygen (O 2 ) concentration in exhaust emissions versus engine speed at WOT. The O 2 concentration in the exhaust emissions depends on combustion performance such as increasing Fig. 8. Carbon monoxide versus engine speed at WOT. Fig. 10. Exhaust gas temperature versus engine speed at WOT.

7 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e Fig. 11. Oxygen concentration versus engine speed at WOT. CO 2 decreases O 2 concentration. Normally it is stated that less O 2 means better combustion at stoichiometric condition. It can be seen that starting from 1500 rpm, increasing engine speed decreases the O 2 level due to better combustion. On average all over the engine speed range, the O 2 produced are , and g/kw h by CNG-DI, CNG-BI and gasoline-pi respectively. The CNG-DI engine shows a lower level of O 2 in exhaust emission AFR Fig. 12 shows AFR versus engine speed from 1500 rpm to 6000 rpm. It can be said that gasoline-pi and CNG-BI engines AFR are optimised by engine manufacturers, not controlled during this investigation. However, for the CNG-DI engine, the AFR has been controlled to improve two out of three parameters such as thermal efficiency, brake mean effective pressure and lowest NOx emissions. The AFR was controlled by a commercial editable ECU. It is found that the average AFRs all over the engine operating speed range are 25, 20 and 20 by CNG-BI, CNG-DI and gasoline-pi respectively. The AFR from 3500 rpm to 6000 rpm was almost similar at for all the engines Thermal efficiency Thermal efficiency is a measure of the efficiency and completeness of combustion of the fuel, or more specifically, the ratio of the output or work done by the working substance in the cylinder in a given time to the input or heat energy of the fuel Fig. 13. Thermal efficiency versus engine speed at WOT. supplied during the same time. The brake thermal efficiency is shown in Fig. 13. It is found that all over the speed range the average thermal efficiency of the gasoline-pi engine (31%) is higher than the CNG-BI (27%) and CNG-DI (29.21%) engines. It can be seen that the thermal efficiency of the CNG-DI engine is higher than the CNG-BI engine Exhaust gas flow rate Fig. 14 shows exhaust gas flow rate for all the engines. It is found that the exhaust gas flow rate for the CNG-DI and gasoline-pi engines was the same at kg/min kw. The exhaust gas flow rate for CNG-BI was kg/min kw Maximum cylinder pressure The change of in-cylinder pressure for all the engines have been determined and analysed. Factors affecting the peak pressure are the engine compression ratio, load, combustion duration, net heat release rate and AFR distribution inside the cylinder. Fig. 15 shows the rate of change of cylinder pressure for all the engines looks very similar although there is a difference in the magnitude of the peak pressures. It is observed that the CNG-DI engine produces higher maximum cylinder pressure (105 bar) followed by the gasoline-pi (91 bar) and CNG-BI engines (75 bar). Too much high or too much low cylinder peak pressure is not suitable for effective pressure acting on the cylinder during expansion stroke such as it reduces engine torque [17]. The reasons of producing higher cylinder peak pressure by CNG-DI engine is mainly due to increasing compression ratio (which is 14) and other effects of combustion characteristics such as improper burning rate, fuel Fig. 12. Air fuel ratio versus engine speed at WOT. Fig. 14. Exhaust gas flow rate at WOT.

8 3570 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e3571 Fig. 15. Cylinder pressure versus crank angle at 6000 rpm and WOT. mixing problem due to cool gas entering into the engine cylinder and higher compression ratio as compared to gasoline-pi and CNG-BI engines. 4. Conclusion The CNG-DI engine did not have any initial starting difficulties due to fuel ignited by a spark plug. The engine did not show any combustion noise at compression ratio of 14 (initial compression ratio was 10). The following conclusions may be drawn from the present investigation: I. On average over the speed range, the CNG-DI engine produces 2.02 kw (4.26% lower) less power as compared to gasoline-pi and 8.67 kw (19.10% higher) power than CNG-BI engine. II. The CNG-DI engine reduces fuel consumption as 0.28% and 7.87% in comparison to gasoline PI and CNG-BI respectively. III. The CNG-DI engine reduces 42% NOx emissions as compared to the original base gasoline engine such as gasoline-pi system. IV. The CNG-DI engine produces 25% higher and 56% lower HC as compared to gasoline-pi and CNG-BI engine respectively. V. The CNG-DI engine produces higher CO emission as compared to gasoline-pi and CNG-BI engines. VI. In general, it can be stated that CNG-DI engine performs similar to gasoline-pi engine and better than CNG-BI engine. Acknowledgements The authors would like to thank Mr. Sulaiman Bin Ariffin (Laboratory Assistant), Muhammad Redzuan bin Umar (Research Assistant) and Mohd Khair bin Hassan (Research assistant from UPM) for providing special technical assistance related to engine test bed, ECU calibration and data collections. Without their help, it was very difficult to complete the engines test. A special acknowledgement is also offered to the University of Malaya and the Ministry of Science, Technology and Innovation for the research grant of this project through Vote- IRPA No: & UMRG 040/9AET. The authors would like to thank UPM, UKM, UiTM, UTP and Proton Berhad which successfully to produce the new CNG-DI engine through collaborative research works. In addition, the authors would like to thank, References from [19e24] which were helpful to understand the differences among CNG, Dual fuel and hydrogen combustions characteristics. Appendix Uncertainty calculation for NOx emission of CNG-DI engine. Three tests Max-Min Value Analyser accuracy [16] Average (ppm) % Uncertainty Average in (g/kw h) Test 1 Test 2 Test3 Max Min ppm ppm ppm ppm ppm þ10 ppm 10 ppm ppm þ g/kw h L3.20 Uncertainty level of NOx emission for CNG-DI engine: 3.20%.

9 M.A. Kalam, H.H. Masjuki / Energy 36 (2011) 3563e References [1] Sandeep M, Patric O, James H, Costi N, Jeff T, Stewart W. Direct injection of natural gas in a heavy-duty diesel engine, SAE paper no: ; [2] Dale G, Mark D, Sandeep M, Edward LP, John Wright, Vinod Duggal, et al. Development of a compression ignition heavy duty pilot-ignited natural gas-fuelled engine for low NOx emissions, SAE paper no: ; [3] Michael RF, Edward LP, Mostafa MK, Scott Wayne, Ralph DN, Nigel NC, et al. An emission and performance comparison of the natural gas Cummins Westport Inc. C-Gas Plus Versus Diesel in Heavy- Duty Trucks, SAE paper no: ; [4] Hill PG. Analysis of combustion in diesel engines fueled by directly injected natural gas. ASME Journal of Engineering for Gas Turbines and Power January 2000;122:141e9. [5] Hodgins KB, Ouellette P, Hung P, Hill PG. Directly injected natural gas fueling of diesel engines, SAE Paper No ; [6] Mtui PL, Hill PG. Ignition delay and combustion duration with natural gas fueling of diesel engines, SAE Paper No ; [7] Wang Jinhua, Huang Zuohua, Zheng Jianjun, Miao Haiyan. Effect of partially premixed and hydrogen addition on natural gas direct-injection lean combustion. International Journal of Hydrogen Energy November 2009; 34(22):9239e47. [8] Schlapbach Louis. Hydrogen-fuelled vehicles. Nature 2009;460:809e11. [9] Huang Zuohua, Wang Jinhua, Liu Bing, Zeng Ke, Yu Jinrong, Jiang Deming. Combustion characteristics of a direct-injection engine fueled with natural gasehydrogen blends under different ignition timings. Fuel February 2007; 86(3):381e7. [10] Shiga S, Ozon S, Machacon HTC, Karasawa T, Nakamura H, Ueda T, et al. Study of the combustion and emission characteristics of compressed-natural-gas direct-injection stratified combustion using a rapid-compression-machine. Combustion and Flame 2002;129(1e2):1e10. [11] Carlucci AP, Risi AD, Laforgia D, Naccarato F. Experimental investigation and combustion analysis of a direct injection dual-fuel dieselenatural gas engine. Energy February 2008;33(2):256e63. [12] McTaggart-Cowan GP, Rogak SN, Munshi SR, Hill PG, Bushe WK. The influence of fuel composition on a heavy-duty, natural-gas direct-injection engine. Fuel March 2010;89(3):752e9. [13] Zeng Ke, Huang Zuohua, Liu Bing, Liu Liangxin, Jiang Deming, Ren Yi, et al. Combustion characteristics of a direct-injection natural gas engine under various fuel injection timings. Applied Thermal Engineering 2006;26:806e13. [14] Hassan MH, Kalam MA, Mahlia TMI, Aris I, Nizam MK, Abdullah S, et al. Experimental test of a new compressed natural gas direct injection engine. Energy & Fuel 2009;23:4981e7. [15] Byung HM, Chung JT, Kim HY, Simsoo P. Effects of gas composition on the performance and emissions of compressed natural gas engines. KSME International Journal, Korea 2002;16(2):219e26. [16] Kalam MA, Masjuki HH, Jayed MH, Liaquat AM. Emission and performance characteristics of an indirect ignition diesel engine fuelled with waste cooking oil. Energy 2011;36:397e402. [17] Heywood JB. Internal combustion engine fundamentals. Mcgraw-Hill International Editions; [18] Bittner RW, Aboujaoude FW. Catalytic control of NOx, CO and NMHC emissions from stationary diesel and dual fuel engines. Journal of Engineering for Gas Turbines and Power July 1992;114:597e601. [19] Wang Jinhua, Huang Zuohua, Fang Yu, Liu Bing, Zeng Ke, Miao Haiyan, et al. Combustion behaviors of a direct-injection engine operating on various fractions of natural gasehydrogen blends. International Journal of Hydrogen Energy October 2007;32(15):3555e64. [20] Alberto AB, Harry CW. The lean burn direct injection jet ignition gas engine. International Journal of Hydrogen Energy September 2009;34(18):7835e41. [21] Korakianitis T, Namasivayam AM, Crookes RJ. Natural-gas fueled sparkignition (SI) and compression-ignition (CI) engine performance and emissions. Progress in Energy and Combustion Science 2011;37(1):89e112. [22] Banapurmath NR, Tewari PG, Hosmath RS. Experimental investigations of a four-stroke single cylinder direct injection diesel engine operated on dual fuel mode with producer gas as inducted fuel and Honge oil and its methyl ester (HOME) as injected fuels. Renewable Energy September 2008;33(9): 2007e18. [23] Papagiannakis RG, Rakopoulos CD, Hountalas DT, Rakopoulos DC. Emission characteristics of high speed, dual fuel, compression ignition engine operating in a wide range of natural gas/diesel fuel proportions. Fuel July 2010;89(7): 1397e406. [24] Wang Jinhua, Huang Zuohua, Miao Haiyan, Wang Xibin, Jiang Deming. Characteristics of direct injection combustion fuelled by natural gasehydrogen mixtures using a constant volume vessel. International Journal of Hydrogen Energy 2008;33:9239e47.