Search Results

The NOx emissions during the crank-start and cold fast-idle phases of a GDI engine are analyzed in detail. The NOx emissions of the first 3 firing cycles are studied under a wide set of parameters including the mass of fuel injected, start of injection, and ignition timing. The results show a strong dependence of the NOx emissions with injection timing; they are significantly reduced as the mixture is stratified. The impact of different valve timings on crank-start NOx emissions was analyzed. Late intake and early exhaust timings show similar potential for NOx reduction; 26-30% lower than the baseline. The combined strategy, resulting in a large symmetric negative valve overlap, shows the greatest reduction; 59% lower than the baseline. The cold fast-idle NOx emissions were studied under different equivalence ratios, injection strategies, combustion phasing, and valve timings. Slightly lean air-fuel mixtures result in a significant reduction of NOx.

The first 3 cycles in the cold crank-start process at 20°C are studied in a GDI engine. The focus is on the dependence of the HC and PM/PN emissions of each cycle on the injection strategy and combustion phasing of the current and previous cycles. The PM/PN emissions per cycle decrease by more than an order of magnitude as the crank-start progresses from the 1st to the 3rd cycle, while the HC emissions stay relatively constant. The wall heat transfer, as controlled by the combustion phasing, during the previous cycles has a more significant influence on the mixture formation process for the current cycle than the amount of residual fuel. The results show that the rise in HC emissions caused by the injection spray interacting with the intake valves and piston crown is reduced as the cranking process progresses. Combustion phasing retard significantly reduces the PM emission. The HC emissions, however, are relatively not sensitive to combustion phasing in the range of interest.

Particulate emissions from a production gasoline direct injection spark ignition engine were studied under a typical cold-fast-idle condition (1200 rpm, 2 bar NIMEP). The particle number (PN) density in the 22 to 365 nm range was measured as a function of the injection timing with single pulse injection and with split injection. Very low PN emissions were observed when injection took place in the mid intake stroke because of the fast fuel evaporation and mixing processes which were facilitated by the high turbulent kinetic energy created by the intake charge motion. Under these conditions, substantial liquid fuel film formation on the combustion chamber surfaces was avoided. PN emissions increased when injection took place in the compression stroke, and increased substantially when the fuel spray hit the piston.

The loss mechanisms associated with engine downsizing, boosting and compression ratio change are assessed. Of interest are the extents of friction loss, pumping loss, and crevice loss. The latter does not scale proportionally with engine size. These losses are deconstructed via a cycle simulation model which encompasses a friction model and a crevice loss model for engine displacement of 300 to 500 cc per cylinder. Boost pressure is adjusted to yield constant torque. The compression ratio is varied from 8 to 20. Under part load, moderate speed condition (1600 rpm; 13.4 Nm/cylinder brake torque), the pumping work reduces significantly with downsizing while the work loss associated with the crevice volume increases. At full load (1600 rpm; 43.6 Nm/cylinder brake torque), the pumping work is less significant. The crevice loss (normalized to the fuel energy) is essentially the same as in the part load case. The sensitivities of the respective loss terms to downsizing are reported.

The effects of the directed port flow produced by a Charge-Motion-Control-Valve (CMCV) on mixture preparation in a Port-Fuel-Injection engine were assessed under conditions typical of fast idle in a cold start process. The port fuel was found to comprise two components: a “valve” puddle (at the vicinity of the valve) that built up quickly, and that was mainly responsible for the delivery of the fuel to the cylinder charge; a “port” puddle located significantly upstream. The latter was mainly created by the reverse back flow process and built up slowly. Although the fuel amounts in these two components were roughly the same, the latter did not significantly interact with the fuel transport to the cylinder charge. The CMCV only weakly affected the purging or filling time of the valve puddle, hence the dynamics of the fuel delivery process was not materially affected.

Knock in a HCCI engine was examined by comparing subjective evaluation, recorded sound radiation from the engine, and cylinder pressure. Because HCCI combustion involved simultaneous heat release in a spatially large region, substantial oscillations were often found in the pressure signal. The time development of the audible signal within a knock cycle was different from that of the pressure trace. Thus the audible signal was not the attenuated transmission of the cylinder pressure oscillation but the sound radiation from the engine structure vibration excited by the initial few cycles of pressure oscillation. A practical knock limited maximum load point for the specific 2.3 L I4 engine under test (and arguably for engines of similar size and geometry) was defined at when the maximum rate of cycle-averaged pressure rise reached 5 MPa/ms.

The fuel property effects on first cycle mixture preparation were assessed by measuring the in-cylinder fuel equivalence ratio (Φ) with a Fast Flame Ionization Detector (FFID) using four different fuels. The Engine Coolant Temperature (ECT) was varied between -6°C and 80°C. The Φ values increased with both ECT and amount of injected fuel mass. The delivery fraction (fraction of the injected fuel that went into the combustible charge), however, increased with ECT but decreased with increase in injected fuel. The minimum required injected mass to produce a combustible mixture increased sharply with decrease in ECT below 20°C. There was, however, no single fuel parameter that would correlate with the measurements over the entire temperature range. Instead, the minimum required injected mass correlated to different distillation points on the ASTM distillation curve; e.g. at ECT of -6°C, it correlated to T20; at 40°C, it correlated to T50.

A strategy to facilitate the mixture preparation process in PFI engines is to delay the Intake Valve Opening (IVO) by shifting the cam phasing so that the cylinder pressure is sub-atmospheric when the valve opens. The physics of the effect are discussed in terms of the pressure differential between the manifold and the cylinder, and the resulting flow and charge temperature history. The effect was evaluated by measuring the equivalence ratio of the trapped charge and the exhaust HC emissions in the first cycle of cranking in a 2.4L engine. When the IVO timing was changed from 18° BTDC to 21° ATDC, the in-cylinder fuel equivalence ratio increased by approximately 10%. This increase was attributed mainly to the enrichment of the charge by displacing the leaner mixture at the top of the cylinder in the period between BDC and IVC. The exhaust HC, however, increased by 40%. No conclusive explanation was established for this increase in HC emissions.

Engine Hydrocarbon (HC) emissions behaviors in the shut down and re-start processes were examined in a production 4-cylinder 2.4 L engine. Depending on when the power to the ECU was cut off relative to the engine events, there could be two or three mis-fired cylinders (i.e. cylinders with fuel injected but no ignition). The total HC pumped out by the engine into the catalyst in the stopping process was ∼ 4 mg (approximately equaled to the amount of one injection at idle condition). Because the size of the catalyst was larger than the total exhaust volume in the stopping process, this HC was not observed at the catalyst exit. The catalyst temperature was also not affected. When the engine was purged after shut down (by cranking the engine with the injectors and ignition disconnected), the total exit HC was 33 mg. In a restart 90 minutes after shut down, the integrated amount of HC emissions due to residual fuel from the stopping process was 16 mg.

The mixture preparation and hydrocarbon (HC) emissions behaviors for a single-cylinder port-fuel-injection SI engine were examined in an engine/dynamometer set up that simulated the first cycle of cranking. The engine was motored continuously at a fixed low speed with the ignition on, and fuel was injected every 8 cycles. Unlike the real engine cranking process, the set up provided a well controlled and repeatable environment to study the cranking process. The parameters were the Engine Coolant Temperature (ECT), speed, and the fuel injection pulse width. The in-cylinder and exhaust HC were measured simultaneously with two Fast-response Flame Ionization Detectors. A large amount of injected fuel (an order of magnitude larger than the normal amount that would produce a stoichiometric mixture in a warm-up engine) was required to form a combustible mixture at low temperatures.

A cycle by cycle analysis of engine behavior during the first few cycles of cranking and start-up was performed on a production four-cylinder engine. Experiments were performed to elucidate the effects of initial engine position (rest position after last engine shut-down), first and second cycle fueling, engine temperature, and spark timing on fuel delivery to the cylinder, engine-out Hydrocarbon (EOHC) emissions, and Gross Indicated Mean Effective Pressure (IMEPg). The most important effect of the piston starting position is on the first firing cycle engine rpm, which influences the IMEPg through combustion phasing. Because of the low rpm values for the first cycle, combustion is usually too advanced with typical production engine ignition timing. For both the hot start and the ambient start, the threshold for firing is at an in-cylinder air equivalence ratio (λ) of 1.1.

Throttle ramp rate effects on the in-cylinder fuel/air (F/A) excursion was studied in a production engine. The fuel delivered to the cylinder per cycle was measured in-cylinder by a Fast Response Flame Ionization detector. Intake pressure was ramped from 0.4 to 0.9 bar. Under slow ramp rates (∼1 s ramp time), the Engine Electronic Control (EEC) unit provided the correct compensation for delivering a stoichiometric mixture to the cylinder throughout the transient. At fast ramp rates (a fraction of a second ramps), a lean spike followed by a rich one were observed. Based on the actual fuel injected in each cycle during the transient, a x-τ model using a single set of x and τ values reproduced the cycle-to-cycle in-cylinder F/A response for all the throttle ramp rates.

An adaptive air/fuel ratio controller for SI engine throttle transient was developed. The scheme is based on an event- based, single- parameter fuel dynamics model. A least- square- error algorithm with an active forgetting factor was used for parameter identifications. A one- step- look- ahead controller was designed to maintain the desired air/fuel ratio by canceling the fuel dynamics with the controller setting updated adaptively according to the identified parameters. When implemented on a Ford Ztech engine and tested under a set of throttle- transient operations, the adaptive controller learned quickly and performed well.

This paper describes the development of a spark ignition engine operating in a stratified-EGR mode at part load. The concept is to reduce the pumping loss with high levels of EGR while maintaining stable combustion via charge stratification. Since the engine operates stoichiometrically, the ability to control NOx emissions by the three-way catalyst is retained. The configuration of introducing the stoichiometric fresh mixture to the center portion of the combustion chamber with the EGR gas on the two sides is visualized in a transparent engine using planar laser-induced fluorescence (PLIF) and Mie scattering. Visualization results showed that the stratification between air/fuel mixture and EGR gas was relatively well established during the intake stroke. There was, however, significant mixing in the late part of the compression stroke.

A research program designed to measure the contribution from fuel absorption in the thin layer of oil, lubricating the cylinder liner, to the total and speciated HC emissions from a spark ignition engine has been performed. The logic of the experiment design was to test the oil layer mechanism via variations in the oil layer thickness (through the lubricant formulations), solubility of the fuel components in the lubricants, and variations in the crankcase gas phase HC concentration (through crankcase purging). A set of preliminary experiments were carried out to determine the solubility and diffusivity of the fuel components in the individual lubricants. Engine tests showed similar HC emissions among the tested lubricants. No consistent increase was observed with oil viscosity (oil film thickness), contrary to what would be expected if fuel-oil absorption was contributing significantly to engine-out HC. Similarly, no effect of crankcase purging could be observed.

The fuel effects on throttle transients in PFI spark ignition engines were assessed through experiments with simultaneous step change of the throttle position from part load to WOT and increment of the injected fuel amount. The test matrix consisted of various gasoline/methanol blends from pure gasoline to pure methanol, coolant temperatures at 40C (for cold engine condition) and 80C (for warm engine), and different levels of fuel enrichment at the WOT condition. The x-τ model was used to interpret the engine GIMEP response in the transient. Using the model, a procedure was developed to calculate the parameters of the transient from the data. These parameters were systematically regressed against the fuel distillation points, the increment in injected fuel mass in the transient, and the enthalpy required to evaporate the fuel increment as the explanatory variables.

A reciprocating test apparatus was constructed in which the friction of a single piston ring against a liner segment was measured. The lubrication oil film thickness was also measured simultaneously at the mid stroke of the ring travel using a laser fluorescence technique. The apparatus development and operation are described. Results are presented from a test matrix consisting of five different lubrication oils of viscosity (at 30°C) ranging from 49 to 357 cP; at three mean piston speeds of 0.45, 0.89 and 1.34 m/s; and at three ring normal loading of 1.4, 2.9 and 5.7 MPa. At mid stroke, the oil film thickness under the ring was ∼0.5 to 4 μm; the frictional coefficient was ∼0.02 to 0.1. The frictional coefficient for all the lubricants tested increased with normal load, and decreased with piston velocity. Both mixed and hydrodynamic lubrication regimes were observed. The friction behaviors were consistent with the Stribeck diagram.

The effect of engine operating parameters (speed, spark timing, and fuel-air equivalence ratio [Φ]) on hydrocarbon (HC) oxidation within the cylinder and exhaust system is examined using propane or isooctane fuel. Quench gas (CO2) is introduced at two locations in the exhaust system (exhaust valve or port exit) to stop the oxidation process. Increasing the speed from 1500 to 2500 RPM at MBT spark timing decreases the total, cylinder-exit HC emissions by ∼50% while oxidation in the exhaust system remains at 40% for both fuels. For propane fuel at 1500 rpm, increasing Φ from 0.9 (fuel lean) to 1.1 (fuel rich) reduces oxidation in the exhaust system from 42% to 26%; at 2500 RPM, exhaust system oxidation decreases from 40% to approximately 0% for Φ = 0.9 and 1.1, respectively. Retarded spark increases oxidation in the cylinder and exhaust system for both fuels. Decreases in total HC emissions are accompanied by increased olefinic content and atmospheric reactivity.

A method for determining the flame contour based on the flame arrival time at the fiber optic (FO) spark plug and at the head gasket ionization probe (IP) locations has been developed. The experimental data were generated in a single-cylinder Ricardo Hydra spark-ignition engine. The head gasket IP, constructed from a double-sided copper-clad circuit board, detects the flame arrival time at eight equally spaced locations at the top of the cylinder liner. Three other IP's were also installed in the cylinder head to provide additional intermediate data on flame location and arrival time. The FO spark plug consists of a standard spark plug with eight symmetrically spaced optical fibers located in the ground casing of the plug. The cylinder pressure was recorded simultaneously with the eleven IP signals and the eight optical signals using a high-speed PC-based data acquisition system.

The liquid fuel entry into the cylinder and its subsequent behavior through the combustion cycle were observed by a high speed CCD camera in a transparent engine. The videos were taken with the engine firing under cold conditions in a simulated start-up process, at 1,000 RPM and intake manifold pressure of 0.5 bar. The variables examined were the injector geometry, injector type (normal and air-assisted), injection timing (open- and closed-valve injection), and injected air-to-fuel ratios. The visualization results show several important and unexpected features of the in-cylinder fuel behavior: 1) strip-atomization of the fuel film by the intake flow; 2) squeezing of fuel film between the intake valve and valve seat at valve closing to form large droplets; 3)deposition of liquid fuel as films distributed on the intake valve and head region. Some of the liquid fuel survives combustion into the next cycle.