This work describes:

1) The investigation of the transient response and the fuel scheduling during throttle transients of an air-cooled 4-stroke single-cylinder engine. Which was done by measurement and analysis of the exhaust air fuel ratio with a fast response wide range oxygen sensor (UEGO). The engine was equipped with an experimental control system that allowed precisely controlled throttle step transients with the stock carburetor at various initial air/fuel ratios. The air fuel ratio response of the engine to the transients was analyzed quantitatively and qualitatively.

2) The development of a methodology to assess the characteristics of the fuel transfer in the intake manifold. In order to eliminate the fuel metering uncertainty associated with carburetors, the stock carburetor was modified with a fuel injection system replacing the main jet of the carburetor providing complete control over the fuel delivery. The fuel injector was installed in two different configurations in the throttle body - with the fuel spray exiting directly into the intake air stream (direct mount injector) and with the fuel spray entering the air stream after passing through a small cavity (siphon tube injector). Pictures of the fuel spray showed significant differences in the fuel spray formation and mixture preparation for the two configurations. Fueling step transients at constant throttle position between various air/fuel ratios both rich and lean of stoichiometric were run. The engine exhaust air-fuel ratio response to the fuel perturbations was analyzed with a two-parameter fuel flow model for the intake manifold. The deposit fraction parameter indicates how many percent of the metered fuel are deposited in the wall film and the wall film depletion time constant parameter shows how fast the wall film is depleted through evaporation and film transport into the cylinder. An optimization technique was used to determine the model parameters yielding the best agreement between the engine response and the simulation. The parameters were determined as a function of the throttle position for both injector configurations. As part of the fuel perturbation transients, the response characteristics of the UEGO sensor were determined with propane induced air fuel ratio step transients at constant throttle position.

The throttle transients for the carbureted engine setup revealed complex response characteristics. The opening and closing throttle transients involved both lean and rich excursions, which were shown to be repeatable. The excursions were more severe for leaner air fuel ratios. For opening throttle transients, lean excursions were follows by short rich dips in some cases. For closing throttle transients, short rich excursions over 1 … 2 engine cycles were followed by relatively long lean excursions lasting for 20 … 40 engine cycles. Among others, the throttle transients between 25 and 50% throttle position showed a response behavior that cannot be explained only with the intake manifold fuel flow effects. This indicated that the transient air-fuel ratio response is influenced by the fuel metering and response characteristics of the carburetor itself. Despite large engine speed changes during the transient, no apparent correlation between engine speed and air-fuel ratio could be found. The carbureted engine setup showed the highest cyclic variability of the air fuel ratio of all investigated engine setups. This is attributed to fuel metering instabilities.

The fuel perturbation transients revealed a very similar wall film deposition behavior for both injector versions at smaller throttle openings, with deposit fractions in the order of magnitude of 10 to 20 percent. Whereas the deposit fraction for the siphon tube injector changed only slightly at larger throttle positions, the fraction for the direct mount injector increased significantly with throttle position to values in the order of magnitude of 35 percent. Overall, the direct mount injector version showed a faster depletion of the fuel wall film through evaporation and film transport into the cylinder. This is attributed to the larger wall film surface area caused by the direct mount injector.

The calculated intake manifold wall film mass would be sufficient to buffer a range of 3 to 7 engine cycles, depending on the injector version and throttle position. The buffering capacity is similar for both versions at 50 and 75% throttle position. The data indicate that at 25% throttle position, the buffering capacity of the siphon tube injector is larger whereas at 100% throttle position, the buffering capacity of the direct mount injector is larger by a factor of almost two.

Despite the higher wall film mass for the direct mount injector at wide-open throttle, the cycle-to-cycle variability of the air-fuel ratio was less than half of the variability for the siphon tube injector. This indicates that the air-fuel ratio variability was more strongly influenced by fuel metering variabilities than wall film effects for this set of experiments.

The fuel perturbation method showed sensitivity of the calculated fuel deposit fraction and wall film depletion time constant to the UEGO sensor response characteristics. However, the calculated wall film mass was shown to be insensitive to changes in the UEGO sensor transport delay and first order time constant of ±10%.