ECMS Energy Management Strategy for Plug-in Hybrid Electric Vehicles Optimized by BP Neural Network

This dataset contains the simulation model, training data, and source code associated with the manuscript entitled "ECMS Energy Management Strategy for Plug-in Hybrid Electric Vehicles Optimized by BP Neural Network". The research proposes a feedforward adaptive prediction method for the equivalence factor in ECMS based on a BP neural network. The dataset includes: (1) Vehicle model: MATLAB/Simulink model of a parallel P2-configuration plug-in hybrid electric vehicle, including engine, motor, battery, transmission, and driver sub-models. (2) Training dataset: Optimal equivalence factor trajectories obtained via offline dynamic programming (DP) under six standard driving cycles (WLTC, CLTC-P, US06, NEDC, HWFET, and SC03), along with corresponding vehicle state variables (speed, SOC, demanded torque, and acceleration). (3) BP neural network code: MATLAB scripts for network training, validation, and testing, including the dual-hidden-layer architecture and hyperparameter settings. Simulation results: Comparison data of fuel consumption, SOC trajectories, and emission performance (CO, THC, NOx) under UDDS cycle for ECMS, PI-ECMS, Fuzzy-ECMS, QL-ECMS, and BP-ECMS. All data and code are provided to ensure the reproducibility of the experimental results. Readers are encouraged to use this dataset for further research on adaptive energy management strategies for hybrid electric vehicles.