Adaptive ECMS Energy Management Strategy for Hybrid Vehicles

Current energy management strategies based on operating condition recognition do not account for driving conditions that do not fully conform to typical scenarios. As a result, optimization parameters derived from typical conditions may not adequately adapt to actual driving conditions. This paper proposes an Equivalent Consumption Minimization Strategy (ECMS) energy management method based on operating condition probability prediction. Firstly, the k-means clustering algorithm is employed to create a dataset of typical driving conditions from vehicle speed data. The Dynamic Programming (DP) algorithm is then used to obtain the optimal equivalent factors for these typical conditions. A feedforward neural network is utilized to establish a model for predicting the probability of operating condition categories. The probability predictions are used to determine whether to use the optimal equivalent factors derived from typical conditions. For conditions with low recognition probabilities, an adaptive equivalent factor update strategy that integrates typical conditions and speed information is proposed. Based on this, a new adaptive ECMS is designed. Simulation results show that this energy management strategy achieves good fuel economy, reducing fuel consumption by 3.92% compared to the traditional PI control strategy that minimizes equivalent fuel consumption.