Development and verification of a real-time energy management system for a dual-energy electric bus using particle swarm optimization

• A dual-energy storage system combining lead-acid and lithium batteries is integrated for electric bus applications. • A control-oriented third-order EV dynamics model is developed for performance evaluation under varying driving conditions. • A RBC strategy and an advanced EMS using ECMS and PSO are designed and compared. • A real-time HIL testing framework is implemented to validate control strategies in realistic scenarios. • Compared to RBC, ECMS and PSO achieve up to 10 % improvement in energy efficiency, demonstrating practical feasibility for next-generation EVs. For electric vehicles (EVs), traditional single-battery systems struggle to balance cost, lifespan, and performance under varying load demands and driving cycles. Integrating dual-energy storage systems has thus become an attractive approach to leverage the complementary advantages of each technology. In this study, the main focus is to coordinate power distribution between two different battery types, lead-acid battery and lithium battery. Firstly, a control-oriented third-order EV dynamics was constructed in advance for performance evaluation. For the baseline control, a rule-based control (RBC) with three modes considering demanded power and State-of-Charges (SOCs) of dual energy sources was developed. This study next proposed an advanced energy management system (EMS) adopting the equivalent consumption minimization strategy (ECMS) as the reference of 100 % optimization. Four for-loop layers (demanded power, dual SOCs and power ratio) with a cost function which was the combined power of dual sources influenced by weighting factors of SOCs were constructed. For the Particle Swarm Optimization (PSO) algorithm, the optimal torque distribution ratio was evaluated considering group size. Furthermore, to verify the EMS, a hardware-in-the-loop (HIL) testing framework was constructed to validate the practical feasibility of the proposed strategies under realistic driving conditions. The results demonstrate that, compared to RBC, the proposed ECMS and PSO achieved notable improvements in energy efficiency during the New European Driving Cycle (NEDC), with maximum improvement of 6.62 % and 6.56 %, separately, in pure simulation. For HIL experiments, 10.48 % and 10.41 % improvements were achieved. These findings highlight the practical potential of intelligent optimization algorithms for dual-energy storage systems employed in next-generation electric buses.