| It has been more than twenty years since the hybrid electric city bus became a trend again. And the hybrid electric vehicle (HEV) itself became better and better design, thanks to the mainstream vehicle company regarded the HEV as their important product. In the fields of HEV parts matching, HEV control strategy, and HEV regenerative control strategy, there are a lot of achievements made by the researchers before. For example, the Instantaneous optimization algorithm, and Global optimization algorithm have been utilized as the control strategies of the HEVs. However, the baseline control strategy is the most widely used strategy among the mature strategies.In the project, not only we got the opportunity to obtain further knowledge of HEV bus, but also collected the data of the real driving cycle of city bus. By analyzing the data, it was acknowledged that the proportion of which engine is working in the economic region is low, and the motor worked in the high efficiency region not so much which resulted of the bad economic performance, and the SOC (State of Charge) of the batteries cannot be maintained on a high level.According to the results of the works of the researchers before, the control strategy of Parallel HEV and the driving cycle are the key factors those effect the economy. Therefore, in this paper, I focus on the design of the control parameters. To increase the adaptability of the strategy, I gain a Neuron Network model in the model of control strategy.Main content of this paper is as followed:1. Firstly, set up an AVL CRUISE model, connect it with the MATLAB software, with which I set up the control strategy model. After modified four parameters, I got the economic performance changed. Design the control parameters using the algorithm in the MATLAB software Toolbox. 2. Classify the data which we collected during the project into three collections. Label them as Light Load, Middle Load, and the other, Heavy Load according to the number of the passengers. After the classifying, pick one typical cycle data from each collection as simulation test data.After simulating in a typical real driving cycle, get the4control parameters which related to the economic, the engine torque high envelope, low envelope1, low envelope2, the ratio of assistant torque for the motor.In each typical real driving cycle, design the control parameters by simulation optimization method. Because of the complexity and the data type of the controlling parameters, choose the Genetic Algorithm to optimize the control parameters. In every typical driving cycle, set the variables as the data in a certain parameter, for instance, a number of torque value in a baseline curve. Afterwards, set the options of the variables. Every time after a simulation, the fitness function is calculated with the results of the simulation. Fuel consumption is always objective, and, regard certain functions of the SOC and the missed trace of the vehicle as the punishment functions.When the algorithm meets the stopping criterion, the calculation stops. The value of the variables is the results of the algorithm for the control parameters. With the results, can we get a better result in the certain driving cycle.3. To gain the adaptability of the simulation model, set an adaptive model up inside the vehicle control strategy model. Every100seconds, the model calculates the mean vehicle speed and the time ratio of idle, and then translates them into the Neuron Network model for calculate. The calculation results stand for the driving patterns numbers, so that the control parameters are set to the certain optimized value in the particular driving pattern.Put the vehicle model and the driving pattern recognition model together; get an adaptive simulation model of the HEV city bus. Run the simulation with the original model, the model without adaptability, and the adaptive model. Analyze the working points of the engine and the motor. To test the model, run the model in a test cycle, and get a better result than the others did. |
PDF Full Text Request