Fuzzy Neural Network Control Strategy for Powertrain of Parallel Hybrid Electric Vehicles

摘要

This paper proposes a Fuzzy Neural Network (FNN) control strategy for the powertrain of the parallel Hybrid Electric Vehicles (PHEV). This control strategy properly controls the operation of the engine and the motor, makes them operate efficiently and reasonably distributes the torque of powertrain between the engine and the motor. The proposed FNN control strategy should reduce the fuel consumption and emissions of the PHEV, when it makes the PHEV satisfy the dynamic demand and maintain the balanced State Of Charge (SOC) of the battery. The proposed Fuzzy Neural Network Controller (FNNC) applies the five-layered neural network to realize the function of Fuzzy Logic Controller (FLC), such as fuzzification, fuzzy inference and defuzzification. The FNNC is trained by using the genetic algorithm (GA) so that its parameters, such as weights and thresholds, are optimized. The FNNC architecture is optimized by applying Self-organization Competition Neural Network (SCNN). This FNNC is adaptive, self-learning, robust and error endurable, so it is a kind of intelligent controller. The FNNC is trained with some test data of the PHEV so that it can learn and memorize control rules of the PHEV. When the FNNC is integrated into the PHEV model, it is evaluated via computational simulations under some kinds of driving cycle and is improved in terms of simulation results. The FNNC should be effective within the entire operating rang of the PHEV and achieve the expectant control objective.