| With the advent of mass-producing of hybrid electric vehicles (HEV), studying the control strategy of HEV is significance. Simulation, prototype development, hardware-in-the-loop simulation and system bench test are emploied to address the HEV control problem. In this paper, a platform framework was presented for the development of HEV control strategy by abstracting the essential factors of the HEV system. The system was modularized, the states of the modules and their relationship for different developing stage was described. To explore a large design space for HEV, a universal interface definition for Meta-Model was designed, and a neural network based scaling modeling method for engine and electric machine was proposed. Bench test indicated that the scaling model can satisfy the requirement for simulation.The functional modules of the HEV test system are designed with modularization and hierarchical method. Virtual parts or physical parts are optional to make up various types of test systems, so the system is flexibility to satisfy different of control strategy development purpose. The designed test system can implement the regeneration braking test of HEV. Proper interfaces were designed to integrate the dSPACE system, dynamometer system and hybrid powertrain. Some tests were implemented on the designed test system. The functions of duty cycle simulation, regenerative brake test, HEV mode switch control test and gear shift control test were validated. The test system meet the expected functions, this made a foundation for the study of the transient process control in this paper.For either energy management or transient process control, it is essential to obtain the torque of the engine. Based on automatic test data, a look-up-table model and neural network model were constructed. This paper comes up with a new type of neural networks based mean value engine model. Experiments results show that these three types of engine models all have certain accuracy for engine torque estimation, but the repeatability of the former two types of model is not steady. The neural network based mean value engine model observe the engine torque by measuring intake pressure and fuel injector flow rate, it has more better robustness and can feedback the transient engine torque if the change rate of throttle is not too high.For traditional HEV as well as Plug-in HEV, it is very important to have good mode switch and gear shift process. This paper come up with a mode switch control strategy of model-predictive based motor speed regulation, it dose not depend on the accurate engine torque estimation and has better robustness than the only engine torque based torque compensation strategy. The simulation and test results testified that the proposed control strategy can effectively suppress the oscillation of output torque of the parallel hybrid powertrain, and realized a smooth mode switch.In order to address the gear shifting control problem, a planetary gearing automatic mechanical transmission (PGAMT) was proposed. The torque converter of a hydraulic automatic transmission was removed to form a PGAMT, which transmission efficiency was improved without sacrificing the convenience of the gear shifting. This paper analyzed the operation conditions required for PGAMT in the parallel hybrid powertrain, modeled the dynamic of the PGAMT, proposed a gear shifting control strategy that use the electric motor as the actuator to actively synchronize clutch disks to realize clutch-to-clutch gear shifting, simulated the process of up shift and down shift. The simulation results indicate that the active synchronizing by motor shift control strategy can effectively reduce the oscillation of the output torque of the transmission during the course of gear shifting.
|
PDF Full Text Request