Engine Downsizing Through Two-Stroke Operation in a Four-Valve GDI Engine

2015-01-XXXX

Engine downsizing through two-stroke operation in a four-valve GDI engine

Author, co-author (Do NOT enter this information. It will be pulled from participant tab in MyTechZone)

Affiliation (Do NOT enter this information. It will be pulled from participant tab in MyTechZone)

Copyright © 2015 SAE International

Page 1 of 13

Abstract

With the introduction of CO2 emissions legislation in Europe and many countries, there has been extensive research on developing high efficiency gasoline engines by means of the downsizing technology. Under this approach the engine operation is shifted towards higher load regions where pumping and friction losses have a reduced effect, so improved efficiency is achieved with smaller displacement engines. However, to ensure the same full load performance of larger engines the charge density needs to be increased, which raises concerns about abnormal combustion and excessive in-cylinder pressure. In order to overcome these drawbacks a four-valve direct injection gasoline engine was modified to operate in the two-stroke cycle. Hence, the same torque achieved in an equivalent four-stroke engine could be obtained with one half of the mean effective pressure. A wet sump was employed to avoid the inherent lubrication and durability issues of conventional two-stroke engines, and the scavenging process was ensured via external boosting. The adoption of direct fuel injection removed the problem of fuel short-circuiting present in mixture scavenged engines. Several loads were tested at 800 rpm and 1600 rpm and the overall engine performance was presented. Gaseous and smoke emissions were measured and examined, as well as an analysis of the spark ignition combustion process. The results demonstrated that very high torque at low engine speeds could be obtained at relatively low in-cylinder pressures and reasonable fuel consumption results.

Introduction

In order to meet the stringent carbon dioxide (CO2) emission target of 95 g/km for passenger cars in the European Union by 2020 [1], intensive efforts are being spent in the research and development of higher efficiency spark ignition (SI) engines. Controlled auto-ignition (CAI) combustion, also known as homogeneous charge compression ignition (HCCI),has been quoted for many years as apromising technology to reduce both fuel consumption and oxides of nitrogen (NOx) emissions in gasoline engines [2]. More recently, attention has also been given to gasoline partially premixed combustion (PPC), where diesel-like efficiencies and low NOx and soot emissions can be obtained [3].However, while these concepts are not fully developed and their transient characteristics improved, SI flame propagation combustion remains as the majorheat release process among gasoline engines.

In the context of SI combustion, engine downsizing has been accepted as an effectivemethod to reduce fuel consumption at part load operation in four-stroke engines. Under this approach the engine displacement and the number of cylinders are reduced so the mid-low operating points are shifted towards regions where pumping and friction losses are minimized. Considering the nature of driving cycles such the New European Drive Cycle (NEDC) and the Worldwide Harmonized Light Duty Test Cycle (WLTC), such improvements in the mid-low load range have a large impact over the vehicle’s total CO2 emission. Performance results from a two-cylinder 0.85 dm3 gasoline direct injection (GDI) downsized engine [4] have shown 24% improvement in fuel consumption compared to a four cylinder 1.6 dm3 naturally aspirated engine over the NEDC.Similarly, a 60% downsized engine,able to reach 3.3 MPa brake mean effective pressure (BMEP), has demonstrated up to 38% better fuel economy in certain regions of the same driving cycle [5].

The reduced swept volume of downsized engineshas a negative impact on the low-end torque and transient response. Results demonstrate that at least 50% in boost buildup is needed to meet the transient response of a 30% downsized engine [6]. Thus, the adoption of two-stage turbocharging and variable valve actuation has become essential [7]. The increased requirement for torque at mid-low engine speeds has raised concerns about knocking combustion, as the charge temperature and pressure become higher at increased boost levels.In this case, direct fuel injection and higherRON gasolines [8], as well as alcohol fuels [9], turned out to be a requirement in heavily downsized engines.However, the longer engine operation at higher loadsinduces thermal and structural stresses, particularly withengine downsizing beyond 50%. Greater in-cylinder pressures and temperatures result in not only more frequent knocking combustion but also low speed pre-ignition (LSPI), also known as super-knock or mega-knock.LSPIis among the main issues compromising engine operation and its durability [10], particularly when engine downspeeding is also sought.

Compared to four-stroke engines, the two-stroke cycle operation can provide similar values of torque with lower in-cylinder pressure and hence less structural and thermal stresses. Its doubled firing frequency provides a greater power density and power-to-weight ratio particularly in the mid-low speed range. These advantages, mostly attributed to ported two-stroke engines, are sometimes offset by poor fuel consumption, excessive unburned hydrocarbon (UHC) emissions and cranktrain lubrication and durability issues [11]. The first two mentioned problems are mainly associated to crankcase scavenged engines, where the air-fuel mixture is always prone to short-circuit to the exhaust port and hence increasing fuel consumption and UHC emissions. In this regard, direct fuel injection has basically left the scavenging of burned gases to be performed by solely air, as the start of fuel injection (SOI) can take place after the piston covered the ports in the liner. Several direct fuel injection systems have been proposed in this framework, with special attention to low cost air assisted fuel injection concepts [12, 13]. Concerning the cranktrain lubrication and durabilityin crankcase scavenged engines, the main issue remains in the use of a dry sump so lubricant oil needs to be added to the air stream. Besides, the presence ofcold intake ports and hot exhaust ports in the cylinder liner results in uneven thermal stresses and bore distortion [14]. To overcome theseissuesthe so called poppet valve two-stroke cycle spark ignition engine was proposed in the early 1990’s. It shared the same design of four-stroke engines with overhead valves and a wet sump [15, 16].

In the poppet valve two-stroke spark ignition enginethe scavenging process is performed during a positive valve overlap period around bottom dead center (BDC), when both intake and exhaust valves remain opened for typically 120° of crank angle (CA). By employing a wet sump the engine durability was found similar to that obtained in four-stroke engines, despite of the valvetrain which operates at the same speed of the cranktrain. To scavenge the burned gases a roots type supercharger isusually employed [17], although arrangements with turbochargers and electric compressors have been also evaluated [18]. The adoption of high pressure direct fuel injection has improved fuel economy and UHC emissions by eliminating the charge short-circuiting and enhancing the mixture preparation. Moreover, the possibility of using the poppet valve two-stroke enginewith low temperature combustion has demonstrated outstanding fuel economy results with CAI combustion [19] and gasoline PPC [20].

The requirement for smaller engines with higher power densities for passenger cars, as well as its application as range extenders in hybrid powertrains [21], has renewed the interest in the two-stroke cycle. Recently, a two-stroke two-cylinder poppet valve Diesel engine, with similar unitary displacement to the engine used in this research, was proposed to be used in small vehicles [22]. In case of gasoline engines, the development of more efficient superchargers and electrically assisted compressors, variable valve actuation systems and higher pressure direct injection components has enabled the achievement of improved fuel economy results. For these reasons, the present research focuses on investigating performance, combustion and emissions of a spark ignition GDI poppet valve two-stroke engine. Two engine speeds i.e.800 rpm and 1600 rpm were tested at five different load conditions ranging from 0.2 MPa to 1.0 MPa indicated mean effective pressure (IMEP). Therefore, the low speed operating range could be investigated in the framework of engine downsizing and downspeeding.

Experimental setup

The experiments were performed in a single cylinder research engine (Table 1) equipped with a fully variable valvetrain unit capable of independent control over the fourvalves. The engine was coupled to a transient dynamometer and instrumented according to Figure 1. The boosted air necessary to scavenge the burned gases in the two-stroke cycle was supplied by an AVL 515 compressor with closed loop control over the air pressure. The intake air mass flow rate was measured by a Hasting HFM-200 laminar flow meter, while the air temperature was kept at 300±5 K. The fuel flow rate was measured by an Endress+Hauser Promass 83A Coriolis meter and maintained at 300±5 K and 14.5±0.5 MPa. The in-cylinder pressure was measured by a Kistler 6061B piezo electric sensor,while the intake and exhaust pressures were measured by two Kistler piezo-resistive sensors 4007BA20F and 4007BA5F, respectively. The acquired pressure data was related to the crankshaft position through a LeineLinde 720 pulses per revolution encoder. K-type thermocouples were employed to collect averaged temperatures at intake, exhaust, oil and coolant galleries, fuel rail and valvetrain oil supply. Carbon monoxide (CO), CO2, UHC and NOx emissions were recorded by a Horiba MEXA 7170DEGR, while smoke emissions were measured by an AVL 415. Engine coolant and oil temperatures were kept at 353±3 K during all tests. Spark and injection timings, as well as valve parameters, were controlled through a Ricardo rCube engine control unit (ECU). All data was received by a National Instruments NI USB-6353 data acquisition (DAQ) card and analyzed on real-time via in-house software in a host computer, which was also used to setup the ECU and dynamometer parameters.

Table 1– Engine specifications.

Engine model / Ricardo Hydra Camless Two/four-stroke
Displaced volume / 0.35 dm3
Bore / 81.6mm
Stroke / 66.9 mm
Compression ratio / 11.8:1
Combustion chamber / Four valves pent-roof with central spark plug
Fuel / UK commercial gasoline RON 95
Fuel injector / Side mounted solenoidtype Magneti Marelli six-holes
Exhaust valve opening / 120° CA ATDC
Intake valve opening / 130° CA ATDC
Exhaust valve closing / 230° CA ATDC
Intake valve closing / 240° CA ATDC
Valve lift / 8.0 mm

Figure 1 – Schematic representation of the engine test cell.

By correlating the in-cylinder pressure to the crankshaft position it was possible to estimate the net heat release () based on the first law of thermodynamics [23]. In this case the combustion chamber contents were considered as a single zone, so the pressure and volume changes were correlated to the energy released during the combustion as presented in Equation (1). By integrating this equation over the crank angle of interest, the mass fraction burned (MFB) was obtained.

/ (1)

Where: is the crank angle, is the ratio of specific heats (considered constant and equal to 1.33 [23]), is the in-cylinder pressure and is the in-cylinder volume.

Engine-out emissions were converted from volumetric basis (ppm) to indicated specific values basedon the UN Regulation number 49 [24], as presented in the appendix. The combustion efficiency () was calculated based on the species not fully oxidized during the combustion:

/ (2)

Where: is the mass flow rate of CO, is the lower heating value (LHV) of CO (10.1 MJ/kg), is the mass flow rate of UHC, is the LHV of UHC (44 MJ/kg) and is the LHV of the fuel (44 MJ/kg).

The air trapping efficiency () is defined as the ratio of in-cylindertrapped air mass ()at IVC or EVC (whichever later) to the intake air mass () supplied per cycle. It was calculated through the exhaust oxygen content analysis under fuel-rich operation [25]. It is based on the presumption that any remaining oxygen in the exhaust derives from scavenging inefficiencies, such as mixing-scavenging and air short-circuiting. This method was chosen to be used with the knowledge that inaccuracies may have taken place due to charge stratification resulted from direct injection.

/ (3)

Where: , , and are the exhaust gas concentrations of carbon monoxide, carbon dioxide, oxides of nitrogen and oxygen, respectively, is hydrogen to carbon ratio of the fuel (considered 1.87), is the water-gas reaction equilibrium constant (considered 3.5). All exhaust gas concentrations in ppm (volumetric basis).

The scavenge ratio is defined as the ratio between the intake air mass supplied to the in-cylinder reference mass under intake conditions. The reference volume used to calculate the reference mass was the sum of clearance volume () and instantaneous in-cylinder volume at EVC or IVC, whichever later.

/ (4)

The charging efficiency was employed to quantify how efficiently the cylinder was filled with air. This parameter expressed the ratio between the in-cylinder trapped air mass at IVC or EVC (whichever later) and the in-cylinder reference mass at intake conditions (intake air density,). By definition, it resulted from the product between scavenge ratio and trapping efficiency as seen in Equation(5).

/ (5)

Under idealized flow conditions i.e. the scavenging occurring at isobaric and isothermal conditions, the in-cylinder charge and burnt gases may assume identical densities. Thus, the internal exhaust gas recycled (iEGR), or residual gas trapped,could be estimated from the difference between the charging efficiency and the unit [26].

Differently from crankcase-scavenged two-stroke engines where the piston works as an air/mixture pump, in the poppet valve engine an external compressor is used to scavenge the burned gases. Hence, inreal world conditions, part of the engine’s output power would be delivered to the compressor. If a supercharger is to beused, its power requirement ()can be estimated based on the first and second laws of thermodynamics[23]:

/ (6)

Where: is the specific heat of air at constant pressure (1.004 kJ/kg.K), is the ambient temperature, is the engine’s intake pressure(after the compressor), is the atmospheric pressure(upstream the compressor), γ is the ratio of specific heats of air (1.4) and is the compressor efficiency (considered 0.65).

Test conditions

In the four-valve GDI engine the two-stroke cycle was realized by operating the intake and exhaust valves around BDC every crankshaft revolution as presented in Table 1 and Figure 2. The exhaust valves opened before the intake valves so the exhaust blowdown phase could help expelling the burned gases. The intake valves closed after the exhaust valves so an extra charging could be realized as the intake pressure was permanently greater than the exhaust pressure. The choice for the intake and exhaust valve events was based on previous valve timing optimization studies [27], where the intake and exhaust valve timings were adjusted for different engine loads and speeds. In comparison, conventional ported engines have fixed symmetric arrangement of intake and exhaust ports, wherein the exhaust portstypically close after the intake onesdue to the need for those to open first during the expansion phase. With such port timings some of the fresh charge spills out the cylinder until the exhaust ports closes, which increases fuel consumption and UHC emissions. In this regard, exhaust sliding valves have been usedto improve the trapped charge in conventional ported two-stroke engines [28].

The long valve overlap of 90° CA allowed the boosted fresh air, supplied by the external compressor, to scavenge the burned gases by means of a reverse tumble flow motion as presented in Figure 3. To avoid the fresh charge from going straight into the exhaust port, two masks were added around the intake valve seats so the majority of the incoming air could be deflected downwards [29]. As the effective valve curtain area was reduced, upright intake ports were employed so the port discharge coefficient could be improved. The exhaust port shared the same features of four-stroke engines with a side exit.

The SOI was varied according to the engine load and speed from 220° to 270° CA after top dead center (ATDC). At higher engine speeds there was less time available for the charge preparation,so the SOI was advanced to improve its homogeneity. Similarly, at higher engine loads the amount of fuel injected increased and so was the time necessary for vaporization and mixing, which required advanced SOIs. At lower engine speeds and loads the SOI was delayed to create a stratified mixture around the spark plug, so injection timings as late as 90° CA before top dead center (BTDC) were realized in these circumstances.