Research On Online Energy Management Strategy Of Plug-in Hybrid Electric Vehicle Based On Driving Condition Prediction And Offline Optimal Trajectories

Reasonable energy management strategy can significantly reduce fuel consumption and emissions of the plug-in hybrid vehicle;however,the traditional energy management strategy does not consider the impact of variability of actual driving conditions on fuel consumption.Although the offline global optimal energy management strategy can obtain the theoretical minimum fuel consumption value,it can only be applied to the offline state.Therefore,this paper presents an on-line energy management strategy based on conditions prediction and off-line optimal trajectories,which applies off-line optimal trajectories to online situation and at the same time improves the fuel economy of the vehicle and the adaptability of the strategy to different working conditions.The details of the work are as follows.Based on the Matlab/Simulink environment,the simulation models of the whole vehicle dynamics and transmission system are built with the experimental modeling as the main method and the theoretical modeling as the supplement,which provide a simulation platform for further researchAn off-line global optimal energy management strategy is established by using the dynamic programming algorithm.The algorithm was improved by limiting the range of SOC values to 0.6~0.8 at every moment,to reduce the calculation time and improve the versatility of the algorithm,at the same time to ensure the battery SOC within a reasonable range of fluctuations.By limiting the last-minute SOC value to 0.7,the SOC balance under the driving condition is strictly guaranteed in the reverse search for the feasible SOC domain,which improves the credibility of the results.The simulation results show that the proposed DP program has good universality and can be applied to different driving conditions.The established offline global optimal strategy obviously improves the fuel economy compared with the motor assisted strategy.The characteristic parameters of the condition are dimensionally reduced by the correlation analysis,and 12 characteristic parameters are used for the driving condition prediction;on-line prediction of driving conditions is realized by Euclidean approach degree method.Considering that the driving mode represents the driver’s response to the road environment in the short term,which is an important state of the car at present,Therefore,the driving mode is defined as 5 levels.The basic principle of neural network is studied,and the online prediction of driving mode is realized by using BP neural network.Neural network is used to learn and train off-line optimal trajectories and corresponding vehicle condition under standard driving conditions,and online strategy based on neural network and off-line optimal trajectories is established and verified by simulation.The results show that the strategy performs well,and the online application of off-line optimal trajectories and the rational distribution of energy are realized.On the basis of the above strategy,combined with the prediction method of driving condition and driving mode,a comprehensive online energy management strategy based on driving condition and driving mode prediction and off-line optimal trajectories is designed and verified by simulation.The results show that under the designed strategy,the vehicle SOC changes smoothly,and the vehicle has a good capacity to maintain electricity.The strategy improves the fuel economy of the car and has good adaptability to driving conditions at the same time...