Reserch On Predictive Energy Management Of Connected Plug-in Hybrid Electric Vehicle Considering Vehicle To Vehicle Communication Interrupton

With the global automobile industry moving towards the trend of "electric,connected and intelligent",intelligent connected new energy vehicles have become an important measure to achieve the "double carbon" goal.Connected Plug-in Hybrid Electric Vehicle(PHEV)has great potential to improve energy efficiency and reduce emissions.However,the short-term absence of network information has an important impact on the speed prediction and energy management optimization of PHEV.In order to improve the fuel economy and real-time performance of connected PHEV,considering the impact of vehicle-to-vehicle(V2V)communication interruption,the stochastic vehicle speed prediction and energy management strategy based on model predictive control for connected PHEV are analyzed and verified by experiments.(1)Model of plug-in hybrid electric vehicle is established.The working mode of plug-in hybrid electric vehicle is briefly analyzed,and the main components such as engine and motor as well as the vehicle longitudinal dynamics model are established.The rule-based energy management strategy is designed and used to verify the effectiveness of the model by comparing it with business model in Autonomie,thus providing a model support for the formulation of energy management strategy.(2)The influence mechanism of V2V communication interruption on the prediction accuracy of stochastic vehicle speed is revealed.The optimal Conditional Linear Gaussian(CLG)prediction model is determined by comparing different input variables without V2V communication interruption,and the sensitivity of its main parameters is analyzed to determine the reasonable range of historical data is 1～3.Aiming at the problem of V2V communication interruption,a method of interrupt data processing based on piecewise cubic Hermite spline interpolation is proposed.In two different test conditions,the vehicle speed prediction results of nine kinds of V2V interruption situations are compared to explore the main factors affecting the accuracy of vehicle speed accuracy of vehicle speed prediction.It is clear that the stochastic vehicle speed prediction based on the best CLG prediction model is effective within 5 seconds of V2V communication interruption.Compared with the prediction methods based on Back Propagation(BP)neural network and Long Short-Term Memory(LSTM)neural network,the validity of the best CLG prediction model is verified.(3)Based on the best CLG prediction model and considering V2V communication interruption,a stochastic vehicle speed prediction method combining Bayesian network and BP neural network is proposed.Firstly,BP neural network is used to establish the mapping relationship between prediction velocity and error based on the best CLG prediction model,and then the prediction velocity error is compensated.Then,aiming at the local fluctuation of the prediction velocity of the best CLG prediction model,a prediction velocity smoothing module is designed to smooth the compensated prediction velocity and reduce the local fluctuation of the prediction velocity.Finally,four evaluation indexes,root mean square error,standardized residual,goodness of fit and online computing time,are introduced to verify the effectiveness and generalization ability of the proposed fusion prediction method through simulation analysis in two different test conditions.Compared with the best CLG prediction model,the prediction accuracy of the fusion prediction method in the two test conditions is increased by 24.56%,16.99%,5.91%,and 27.28%,17.23%,6.43%,respectively,indicating that the method can effectively improve the prediction accuracy of random speed of connected vehicles considering V2V communication interruption.(4)Considering the impact of V2V communication interruption,two energy management strategies for connected PHEV based on different cost functions and prediction models are proposed.The energy management strategy based on nonlinear model predictive control adopts PHEV nonlinear dynamic model as the prediction model,and solves the nonlinear optimization problem in finite time domain through dynamic programming algorithm.Then,the linearized linear time-varying model is used as the prediction model,and the quadratic programming algorithm is used to solve the linear optimization problem,and then the energy management strategy based on the linear time-varying model predictive control is proposed.The effectiveness of the two proposed energy management strategies is verified by simulation,and the impact of V2V communication interruption on the performance of energy management strategies is discussed.Compared with other energy management strategies,the fuel economy and real-time performance of different energy management strategies considering the impact of V2V communication interruption are evaluated.The results show that the energy management strategy based on linear time-varying model predictive control has similar computational efficiency as the Charge Depleting and Charge Sustaining(CD-CS)control strategy.It can satisfy the real-time requirement,achieve good fuel economy,and has good application potential.(5)The energy management strategy of connecte PHEV is verified by driver-in-loop test.The driver-in-loop test test system is designed based on dSPACE to verify the performance of energy management strategy based on linear time-varying model predictive control.The experimental results show that the fuel economy of the energy management strategy is good.The fuel consumption and equivalent fuel consumption of the energy management strategy under the test condition are 9.59%and 4.72%different from the fuel consumption results optimized by the Dynamic Programming(DP)control strategy,and can meet the real-time requirements.