Validity of a Steady-State Friction Model for Determining CO2 Emissions in Transient Driving Cycles
To be published on September 9, 2019 by SAE International in United States
Due to its high benefit-cost ratio, decreasing mechanical friction losses in internal combustion engines represents one of the most effective and widely applicable solutions for improved engine efficiency. Especially the piston group – consisting of piston, rings and pin – shows significant potential for friction reduction, which can be evaluated through extensive experimental parameter studies. For each investigated variant, the steady-state friction measurements are fitted to an empirical polynomial model. In order to calculate the associated fuel consumption and CO2 emissions in transient driving cycles, the steady-state friction model is used in a map-based vehicle simulation. If transient engine operation entails friction phenomena that are not included in the steady-state model, the simulation could yield erroneous fuel consumption and CO2 predictions. This issue is gaining in importance with the current regulatory driving cycles, which aim to better reflect real-world driving conditions and thus contain more frequent and steep transient events. Therefore, the purpose of this study is to assess the extent to which it is valid to use a steady-state friction model for the determination of CO2 emissions in transient driving cycles. To do so, fired friction measurements are taken using the indication method both (i) with artificial speed and load ramps at constant oil and coolant temperatures, and (ii) at a constant operating point with increasing oil and coolant temperatures. For an overall model validation, a detailed comparison between steady-state friction model and transient friction measurements is made in two regulatory driving cycles and a real-world driving profile. The results demonstrate that under specific operating conditions, a dynamic friction behavior due to a temperature lag of the engine components can lead to momentary friction differences compared to the steady-state values. The validation, however, indicates that these differences appear to be small in driving cycles. Additionally, it could be shown that dynamic friction phenomena only have a minor effect on cumulative fuel consumption in comparison to the influence of the inherent uncertainty of the steady-state friction model.