Energy Management Strategy Of Hybrid Electric Vehicle Based On The Recognition Of Driving Cycle And Driving Style

The power system of hybrid electric vehicle(HEV) composed by multiple sources of power, through the energy management strategy can distribute the required power between multiple sources reasonably and control each component of the power system coordinately, in order to improve the fuel economy of the vehicle on the premise that the HEV keep a good power property performance. The existing energy management strategies have failed to consider how to apply driving cycle recognition into the control unit of the actual car and driving style recognition algorithm influenced by different driving cycle's style. Therefore, the research of energy management strategy for HEV based on driving cycle recognition and driving style recognition has important theoretical significance and application value for further improve the vehicle energy consumption economy.This thesis takes the single motor ISG hybrid electric vehicle as the object of study. In order to improve fuel consumption, it is necessary to carry out research the energy management strategy based on driving cycle recognition and driving style recognition. The specific research content as follows:(1) Experimental data of a city driving cycle obtains by driving test vehicle, and then take it into filter function to decrease noise. Correlation coefficient between drifferent characteristic parameter and correlation coefficient between characteristic parameter and fuel consumption are comprehensively considerated to reduce dimensions for genetic optimization K-means clustering algorithm based on combining genetic optimization K-means clustering algorithm with “micro-trip”. After the clustering analysis, pick out the micro-trips on the principle of the closer between the micro-trip's driving feature and the center of cluster, it is more representative of the cluster, and develop the final city driving cycle. Compare characteristic parameter between different typical driving cycles. Through the comparison and analysis, draw a conclusion that the driving cycle constructed by this new approach have a good response of the city actual traffic conditions and the method have high practical value. The driving data and the classic driving cycle of the city provide the data for optimizing the parameter of HEV and formulating the energy management strategy.(2) Taking a single motor ISG hybrid electric vehicle as prototype vehicle,present a parameter optimized method of hybrid electric vehicle based on multiple driving cycles. Establishing the optimization model by Matlab/Similink and genetic algorithm toolbox, determining the constraints of power performance from the parameters of the prototype vehicle and relevant technical indicator, taking fuel consumption as target of the optimization, optimizing the power train parameters and energy management strategy parameters are concurrently performed respectively aim at single driving cycle and multiple driving cycles. The results show that the parameter optimized based on multiple driving cycles can apply to different driving cycles and improve fuel economy on the premise of meeting the design specifications of vehicle power performance. Experiments and theory models about components of the pivotal power transmission system's are carried out base on obtaining the parameters of the power transmission. Combining with the experiment results and theoretical calculations, the model of the engine, ISG motor, battery, transmission are established.(3) Correlation coefficient between drifferent characteristic parameters, correlation coefficient between characteristic parameter and fuel consumption, sensitiveness of characteristic parameter change which caused by driving cycle are comprehensively considerated to reduce dimensions for optimizing the characteristic parameter. The optimized characteristic parameters are applyed to the driving cycle recognition method based on genetic optimized K-means clustering algorithm and the driving cycle recognition method based on statistical analysis. Comparing the precision and effectiveness between two recognition methods, choose the better driving cycle recognition method as driving cycle recognition in the later research. Four types of typical driving cycles were selected, and then according to the equivalent fuel minimum energy management strategy(ECMS) obtain relation between equivalent fuel factor and specific fuel consumption of four different types of typical driving cycles. After analysis found each driving cycle had an equivalent fuel factor and optimal way of distribution the required power. Energy management strategy of HEV based on driving cycle recognition using genetic optimized K-means clustering algorithm is combine ECMS and driving cycle recognition method based on genetic optimized K-means clustering algorithm. The simulation results demonstrate that comparing with the ECMS without driving recognition, the formulated control strategy makes the fuel consumption reduce 6.84%.(4) Driving style divide into radical driving style, ordinary driving style and clam driving style. Take the influence bring by different typical driving cycles into account, analysis the jerk which caused by the vehicle traveling state for getting the coefficient to distinguish different driving styles under different driving cycles. Combine the coefficient and genetic optimized K-means clustering algorithm to formulate the method for driving style recognition. Acquired the optimal power distribution with combination of driving cycle identification, speed of different driving styles, ECMS to formulate the energy management strategy of HEV based on driving style recognition. The simulation results demonstrate that relative to the energy management strategy based on driving cycle recognition using genetic optimized K-means clustering algorithm the formulated control strategy has the approximate fuel economy, but they focuse on the different recognize direction. Relative to the control strategy without driver style recognition, the formulated control strategy make the fuel consumption reduce 8.47%.(5) Through the comparison and analysis by rotary drum test bench and exhaust-gas analyzer, draw a conclusion that the driving cycle constructed by the new approach have a good response of the city actual traffic conditions and the method have high practical value. The test bench for hybrid electric vehicles powertrain is constructed. The energy management strategy and the control software of test bench for HEV are develop based on Matlab/ Simulink and D2 P platform. Then the data acquisition and calibration control system are builded by ATI-VISION. The reliability of function by power system is verified after the bench test which is include motor drive mode test, engine drive mode test, engine starting while driving model, and comprehensive function test. The road test of energy management strategy based on driving cycle recognition using genetic optimized k-means clustering algorithm implement to verify the real-time performance and fuel economy performance.