Research on energy management strategy for incremental range electric vehicles integrating speed prediction

The main challenge facing current energy management strategies for extended-range electric vehicles is effectively balancing power demand and energy utilization to enhance fuel economy under complex and variable driving conditions. Therefore, to optimize the distribution between the two energy sources of extended-range electric vehicles and improve their fuel economy, this paper proposes an energy management strategy incorporating speed prediction. Firstly, the long short-term memory neural network speed prediction scheme is investigated, and its effectiveness under different cyclic conditions is verified. Secondly, the four hyperparameters of the long short-term memory neural network structure were optimized using the sparrow algorithm (SA) to further enhance the prediction accuracy of the long short-term memory speed prediction algorithm. After optimization, the mean square deviation and mean absolute error are reduced by 46.46% and 54.46%, respectively, compared with the pre-optimization period. Finally, an energy management strategy based on speed prediction was designed using the sparrow algorithm-long short-term memory model. The results show that the speed prediction-based energy management strategy reduces fuel consumption by 6.05% and 3.50% under the New European Driving Cycle and World Light Vehicle Test Cycle conditions, respectively, compared to the rule-based hybrid control strategy.