Drive-cycle descriptor-based comparative analysis and observed regime mapping of hybrid electric vehicle architectures

Conventional comparisons of hybrid electric vehicle (HEV) architectures commonly report aggregate energy-consumption metrics but do not explicitly relate architecture performance to quantitative characteristics of the driving cycle. Consequently, the reasons for architecture-dependent performance changes across different operating regimes remain insufficiently characterised. For a given architecture, fuel economy varies substantially across different drive cycles (e.g. parallel HEV: 31.7 km/L on US06 vs 20.8 km/L on New York City Cycle, NYCC). Many comparative studies rely on a limited set of standard cycles and traditional metrics such as average speed, which do not fully capture transient driving behaviour. This study addresses this gap by introducing a descriptor-based framework using two cycle-level parameters: stop frequency S (stops/km) and mean positive acceleration ā (m·s −2 ). The proposed descriptors S and ā are vehicle-parameter-free; however, the mapping between descriptors and vehicle performance is configuration-dependent. A simulation model was used to compare Series, Parallel and Series-Parallel HEV architectures across 12 drive cycles spanning a wide S–ā space. A parallel plug-in hybrid electric vehicle (PHEV) was additionally evaluated in charge-depleting and charge-sustaining modes. For the simulated vehicle configuration, the Parallel architecture retained the highest fuel economy across all 12 cycles; Series-Parallel provided intermediate performance, whereas Series showed higher electrical conversion losses despite reduced engine thermal losses. The descriptor-based mapping identifies distinct observed driving regimes with different energy-flow and battery-duty characteristics. These findings are synthesised into an observed performance regime map that supports configuration-aware comparison of HEV architectures. The study also presents a battery duty-cycle analysis, revealing that aggressive driving cycles impose significantly higher battery thermal stress than urban stop-and-go cycles despite fewer regenerative braking events. The proposed framework provides a configuration-aware comparative tool for assessing architecture performance according to drive-cycle characteristics while identifying battery-duty trends across diverse operating conditions and V2G relevance.