Optimization Research On Energy Management Strategies For Parallel Hybrid Electric Vehicle

The low pollution hybrid electric vehicle (HEV) is a kind of hybrid fuel-electricvehicle. It is a main research scheme to transition to no pollution vehicle. Therefore, HEVis a realistic scheme that can be industrialized production during the transition period ofcompletely realizing no pollution vehicle. The researches of HEV key technology haveimportant implications.In order to achieve optimize distributive to the HEV power source we must design agood energy management control strategy (EMCS). At present most of research resultshave obtained some achievements in improving fuel economy, but it still can’t reach theglobal optimal results. In order to seek a global optimal control method, the paper takesHonda Insight parallel hybrid electric vehicle (PHEV) as the research subject, the originalrule-based control strategy of the parallel motor assist has studied and improved as well.Firstly mathematical modeling and analyzing to the Insight vehicle simulationmodel in the ADVISOR software, then modeled the simplified equivalent mathematicalmodel based on the standard static principle. Under the given condition of UrbanDynamometer Driving Schedule (UDDS), using the method of Dynamic Programming(DP) to realize optimizing control of the EMCS problem, obtain the DP energymanagement controller that is global optimum, and simulated calculation and experimentto the DP energy management controller in the MATLAB/Simulink environment. Theresults have shown that compared with the traditional rule-based control strategy, the fueleconomy of the put forward energy management strategy is improved to10.22%, theefficiency of the engine and the motor both have corresponding improved at the sametime.In order to enhance the practical applicability of the controller, according to thefeature of unknown driving cycle in practice driving, designed the SDP controller by usedthe stochastic dynamic programming (SDP) method. Firstly, modeling the driver demandpower for markov chain (Markov), and then calculate the state transition matrix, throughsolved the expect performance index obtain the optimal control strategy, it makes the SDPcontrol strategy has the ability to adapt the driving cycle. Designed SDP full dimensioncontroller and SDP local controller based on two kinds of solving methods of stochasticdynamic programming. Simulating results demonstrated that compared with therule-based controller, the fuel economy and the two power source efficiency bothimproved obviously.The simulating experiments of those proposed energy management controllers havebeen carried out in Matlab. The overall results of experiments show that under the energymanagement strategy based on the dynamic programming method, the operating points ofengine are mainly in the high efficiency region, the motor is much more comprehensive totake park in the system working, and the fuel economy gets effectively improved. To the drivability simulation shows that the vehicle engine activities are more sensitive to the fueleconomy than shifting activities. In general, for the given driving cycle, DP controllerworks best, but it can not adapt to the changes of driving cycle. While SDP control canadapted to the driving cycle change, it often could not reach the optimal solution in theory,but it is good in practicability. The control effect of the SDP full dimension controller isbetter than the local controller, but the calculated amount of the SDP full dimensioncontroller is great, it is easy to arise “curse of dimensionality”, and its real-time property isnot good. Although the real-time property of the SDP local dimension controller is better,its solution is suboptimal.