Hu Yanqing

A Supervisory Control Strategy for Plug-In Hybrid Electric Vehicles Based on Energy Demand Prediction and Route Preview

This paper presents a supervisory control strategy for plug-in hybrid electric vehicles based on energy demand prediction and route preview. The aim is to minimize the fuel consumption in real-time operation. This strategy is realized through three successive steps. First, a neural network model is established to predict the energy demand of the vehicle. It reduces the complete traffic data to several statistical parameters, which contributes to ease the prediction process. Second, a mathematical model is proposed to translate the predicted energy demand into a state of charge (SOC) reference of the battery, which significantly simplifies the SOC-programming method. Finally, the adaptive equivalent consumption minimization strategy (ECMS) is used to track the SOC reference and determine the powertrain state. The proposed strategy can optimally distribute the energy between the engine and the motor on a global range and achieve an optimal torque split on a local range. Simulations are carried out on a power-split plug-in hybrid electric bus, and the proposed strategy shows substantial improvements in fuel economy and other indexes compared with the rule-based strategy and the ECMS.

A Neural Network Model to Calculate the Energy Demand of the Vehicle Based on Traffic Features

Hybrid electric vehicles (HEVs) and plug-in hybrid electric vehicles (PHEVs) can achieve high fuel economy and low emissions. And the optimization-based energy management strategies can fully exploits the potential of HEVs to reduce the fuel consumption. As a premise, necessary information about the driving cycles must be known prior. This paper proposes a model to obtain the energy demand of the vehicle, which is pretty useful in the energy management of the HEVs. It uses a radial basis function (RBF) neural network (NN) to process the characteristic parameters of a driving cycle and then outputs the predicted energy demand of the vehicle. The intrinsic parameters of the established NN are optimized using a genetic algorithm (GA). Through tests of real-world driving cycles and standard cycles, the accuracy of the model is verified.

Online Regenerative Braking Predictive Control for Hybrid Electric Bus

Regenerative braking is an effective approach for electric vehicles to reduce fuel consumption and emission. In this paper, we propose a novel online regenerative braking predictive control strategy for hybrid electric bus. Together with the real-time model estimated vehicle mass and road load force, and model recognized distribution information of the bus stations and traffic lights, the strategy can predict the coming deceleration and control regenerative braking appropriately before the driver starts friction brake. The simulation results show that the approach is effective to improve the energy recovery and help to smooth the vehicle decelerating process.