Real-World Driving CO₂ Emissions and Engine Thermal Efficiency Evaluation of Hybrid Vehicles: A Regression-Based Parameter Identification Approach

Real-world driving tests reflect vehicle performance in everyday use, but extracting energy conversion data from extensive measurements remains challenging. The study compared a hybrid electric vehicle and a plug-in hybrid electric vehicle, with the latter operated in charge-sustaining mode, against three conventional vehicles in real-world driving tests. The moving average window method transformed second-by-second measurements into tens of thousands of quasi-trip samples representing diverse driving scenarios. Using the linear correlation between slope CO₂ emissions and slope height in the data window samples, a linear regression-based parameter identification method was proposed to infer the total braking energy recovery ratios of the hybrid vehicles, exploring a way to assess vehicle energy conversion characteristics on the roads. The tested hybrid vehicles exhibited lower urban CO₂ emissions and, under mountainous and aggressive driving conditions, greater advantages primarily attributable to braking energy recovery. However, under the tested conditions, the hybrid vehicles showed no significant advantage in engine thermal efficiency over the conventional counterpart equipped with a downsized turbocharged engine. Engine downsizing may contribute to efficient engine operation, offering some of the efficiency benefits comparable to those achieved through optimized engine–motor torque distribution in hybrid vehicles.