Experimental Investigation of Switching Oxygen Sensor Behavior Due to Exhaust Gas Effects

Abstract

Switching type exhaust gas oxygen sensors are critical to the performance of air-to-fuel ratio control in stoichiometric SI engines. Controlling the air-to-fuel ratio around stoichiometry is necessary for adequate three-way catalyst performance to meet government emissions regulations. However, the feedback signal from the sensor does not always truly depict the actual chemical mixture present in the exhaust gasses, which intrinsically affects the catalyst performance. A sensor may not provide correct air-to-fuel ratio feedback due to certain species in the exhaust gas which affect the equivalence ratio that the sensor switches from the high to low voltage or vice versa. This work attempts to characterize the impact of gas on fresh and aged sensors and builds upon earlier work in the field by using real engine exhaust rather simulated exhaust gas. In these experiments, the air-to-fuel ratio of a stoichiometric gasoline engine is incrementally increased from a lean to rich mixture to elicit the full switching response of the oxygen sensor. Additional sensor output curves are generated in the presence of external additive gases such as hydrogen, carbon monoxide, propane, and gasoline vapor. An automotive emissions analyzer and a hydrogen analyzer detect the concentrations of the exhaust gases and the chemical equivalence ratio is calculated using a carbon balance. This equivalence ratio creates a reference to examine the accuracy of the switch point of the sensor to actual stoichiometry. Using these data sets, it is possible to determine observe the effect of various gas species on the air to fuel ratio at which the sensor switches. The sensitivity of the different sensors to gas concentrations are quantified and presented, which form an elementary model to predict the sensor switch point in the presence of these gas species. Primary findings indicate that the impact of species on the sensor switch point is linearly related to the concentration of the species; sensor type and sensor age have an effect on a sensor's sensitivity to species; and different hydrocarbon species affect sensors differently. The findings support the simulated exhaust gas results reported in the literature in that the degree of interference of a species is related to the diffusion rate of the species with respect to oxygen through the sensor. The results also point toward the importance of the species of hydrocarbons in the engine exhaust, which are uncontrolled and can vary with engine operating conditions. This feature is critical to the application of this knowledge to automotive control.