Multi-objective optimization of hybrid semi-trailer powertrains based on data-driven matching and backward simulation

Heavy-duty commercial vehicles, despite accounting for less than 10% of the total vehicle parc, contribute nearly one-third of the total carbon emissions in the automotive industry. The electrification of heavy-duty vehicles represents a critical pathway to achieving sustainable and green transportation. Hybrid semi-trailer tractors combine the range of diesel vehicles with the advantages of electrification; their powertrain serves as the physical foundation of the entire vehicle, and its parameter matching plays a fundamental role in optimizing the vehicle’s overall energy consumption. Conventional forward simulation approaches based on standardized driving cycles and fixed parameters inherently suffer from information loss and also incur high computational cost. To overcome these limitations, a data-driven framework integrating real operational data analysis and backward simulation is proposed. Real-world operational data are analyzed using a two-layer clustering method to directly extract representative wheel-side demand characteristics that reflect actual user behavior. A decoupled backward simulation model is then developed to efficiently evaluate powertrain configurations integrated with energy management strategies. Based on this framework, a multi-objective optimization model is formulated with the minimization of purchase cost and fuel consumption as the primary objectives, subject to dynamic performance constraints including acceleration, gradeability, and battery state-of-charge sustainability. The optimization problem is solved using the NSGA-II algorithm. Validation through simulation and on-road experiments demonstrates that the proposed method improves simulation accuracy from 94.40% to 98.10%. In road tests, the optimized vehicle reduces equivalent fuel consumption by 2.8% compared with the baseline while satisfying all dynamic performance requirements. The results confirm that the proposed framework provides an efficient and robust methodology for powertrain parameter optimization of hybrid commercial vehicles under real-world operating conditions.