An Energy Management Strategy For Fuel Cell Bus Considering Composite Brake And Hydrogen Consumption

With the gradual popularity of fuel cell buses and the continuous improvement of driver comfort,it has become particularly important to improve the energy efficiency of fuel cell vehicles and reduce braking impacts.In order to achieve this goal,this paper will focus on reducing equivalent hydrogen consumption and braking impact,while ensuring improved economy and braking experience.The specific research contents are as follows:The structural selection and key component parameter determination of the fuel cell vehicle power system were carried out through an analysis of the characteristics of the fuel cell vehicle power system.Corresponding mathematical models were established in matlab/Simulink accordingly.Subsequently,the fuel cell vehicle energy consumption model and electro-hydraulic composite brake model were constructed.The equivalent hydrogen consumption rate was used as an evaluation index for energy consumption,and the braking impact was introduced as an evaluation index for braking experience.This study employs the K-means algorithm to classify drivers into three types of driving styles-smooth,normal,and aggressive-based on nine characteristic parameters derived from real-time vehicle speed information.The effectiveness of three prediction methods,namely Long Short-Term Memory(LSTM),Convolutional Neural Networks(CNN),and Bi-directional Long Short-Term Memory(BiLSTM),is compared under identical training conditions.The results demonstrate that the BiLSTM prediction method achieves better accuracy.Based on these findings,and with the incorporation of information on driving style and historical vehicle speed,a speed correction module is designed to further enhance prediction accuracy.Ultimately,a K-BiLSTM speed prediction module is developed.The threshold values for the hybrid power system based on fuel cell vehicles were determined to design a rule-based energy management strategy.By varying simulation parameters and determining fitting equations,an energy management strategy based on a Linear Quadratic Regulator(LQR)was devised.Additionally,an energy management strategy based on Dynamic Programming(DP)was designed,using equivalent hydrogen consumption rate and brake shock as system state variables,and the three strategies were used as comparative baselines.Then,the K-BiLSTM prediction module was utilized to predict short-term vehicle speeds.Finally,DP was used to balance the composite brake allocation scheme and energy consumption level within the prediction range,completing the design of the K-BiLSTM-DP prediction-based energy management strategy.This study verified the effectiveness and feasibility of the K-BiLSTM-DP prediction-based energy management strategy through MATLAB/Simulink simulation experiments.The results showed that the proposed strategy effectively reduced the impact of brake shock while improving the energy efficiency of fuel cell vehicles,ensuring driver comfort.In comparison to other strategies,this strategy demonstrated superior performance.