Research On Optimization Of Powertrain Control Strategy For Plug-in Hybrid Electric Vehicles

With the increasingly prominent contradictions between the limited energy reserves and environment pollution,the new energy vehicle has been become a positive research direction.Plug-in hybrid electric vehicle(PHEV)can be driven by the large capacity battery pack,which can be charged by grid.The problem of short driving range of pure electric vehicle can be solved by PHEV.Therefore,PHEV has been become an effective transitional products from the traditional vehicle to electric vehicle(EV).PHEVs are equipped with dual energy storage system for internal-combustion engine(ICE)and battery pack.The rational and effective energy split of the ICE and the battery pack is critical for the fuel economy and emission level of PHEVs.The energy management method is to realize the energy split of the ICE and battery pack.In the paper,the major research contents can be divided into four aspects.Firstly,PHEVs are classified,according to the structural forms and the energy flow.According to the detailed analysis,a parallel PHEV is defined as the research object.Based on the different method of operation,the energy management methods are classified and its merits and drawbacks are analyzed.The mainstream construction methods of vehicle driving cycle are analyzed.therefore,the main research contents of parallel PHEV energy management optimization method are determined in the paper.Secondly,there are 10 taxis in normal operating are chose as testing vehicles,which equipped with a high-precision data-acquisition-terminal with GPS/GPRS function,respectively.The high-precision data-acquisition-terminal is connected with vehicle via the CAN-? bus standard interface.The tested data of the vehicle is smoothed by Chebyshev-low-pass-filter and then divided into short-trips by the pre-set rules.The Principal-Components-Analysis(PCA)method is adopted to reduce the dimensionality of the characteristic parameters.The two-step K-means algorithm is applied to extract the clustering and abnormal value for the characteristic parameters to obtain four types of short-trips database.The four types of short-trips are the low-speed,moderate-speed,sub-high-speed and high-speed,respectively.Based on the short-trip partition method,the idle mode database and the running mode database are constructed,then the time distributions of the idle mode section and the running mode section are solved,respectively.Based on the minimum mean absolute deviation of distribution characteristics,the low-speed driving cycle,the moderate-speed driving cycle,the subhigh-speed driving cycle and the high-speed driving cycle of 1200 s are built in the paper.Then,modeling of the driveline system,working mode and original CD/CS strategy are deeply analyzed.Due to the multiple mechanical coupling and electrical coupling mechanisms,the parallel PHEV driveline system has a high non-linear relationship,which brings difficulties to the analysis and simplification of the energy management strategy optimization.With considering of shifting logic,motor efficiency Map,ICE optimal operating curve and fuel-rate Map,driver,automatic gearbox,motor,battery pack and ICE of parallel PHEV power system are modeled and analyzed.Meanwhile,according to the operating modes,the driveline components are simplified and simulated analysis in this paper.Finally,according to the principle analysis of dynamic programming(DP),a parallel PHEV energy management optimization method based on the DP algorithm is proposed,which the battery power commands can be solved and compared with the CD/CS method.In order to solve the high computation complexity,computational burden and engine ON/OFF frequently,an energy management optimization method based on the convex programming is developed.According to the detail analysis of driveline system,a number of quadratic convex functions are proposed to interpret the relationship between the battery power command and the ICE fuel rate,under different vehicle velocity and the driveline power demand.Based on the typical driving cycles and the real-road driving cycles,battery power command and engine on/off power threshold are optimized by the convex programming and the simulated annealing(SA)algorithm.Compared with the CD/CS method and the energy management optimization method based on the DP,the proposed method based on the convex programming is verified and constructed.In the paper,the proposed strategies are extended to considering the different initial SOC.