并联混合动力汽车能量控制策略研究

发布时间：2018-11-19 14:38

【摘要】：随着世界社会经济的高速发展，能源危机和环境污染问题日益凸显。并联式混合动力汽车因具有环保、节能和技术相对成熟等优点而备受关注。并联混合动力汽车的能量控制策略的优劣是影响汽车能耗的重要因素，因此在满足汽车动力性的前提下，优化并联式混合动力汽车能量控制策略对控制能源危机和实现环境的可持续发展具有非常重要的现实意义。论文以并联混合动力汽车能量控制系统为研究对象，对发动机转矩和电机转矩的分配问题进行了深入研究。在分析了并联混合动力汽车的驱动系统结构和结合方式的基础上，建立了发动机数学模型、电机数学模型、蓄电池数学模型、车轮数学模型和传动系动力学方程，并构建了能量控制策略的优化数学模型。因并联混合动力汽车能量控制系统具有动态非线性的特点，论文采用模糊神经网络算法对发动机转矩和电机转矩进行优化分配，分别设计了基于模糊逻辑算法的能量控制策略和基于模糊神经网络算法的能量控制策略，为仿真平台提供了理论基础。在模糊神经网络能量控制策略中，根据并联混合动力汽车控制策略的优化数学模型，采用补偿神经网络结构，构建由输入层、隐含层和输出层组成的前向神经网络，使网络的输入层与模糊逻辑算法的模糊化过程对应，网络的隐含层与模糊推理过程对应，网络的输出层与解模糊过程对应，为了解决神经网络节点与模糊逻辑输入输出接口统一的问题，通过量化公式对输入输出变量进行量化处理，进而利用神经网络自学习和自适应的能力，自动生成模糊规则和隶属函数，并不断优化神经网络输入输出隶属函数的中心和宽度。为了提高系统精度，加快收敛速度，神经网络采用动态调整步长补偿梯度下降的学习算法对能量控制策略进行优化。以丰田普锐斯轿车为例，在ADVISOR2002软件环境下，建立了整车的后向仿真模型，其中包括发动机、电机、蓄电池、传动系和汽车行驶动力学模型，为整车控制策略研究和开发提供了必要的仿真平台，并基于此仿真平台，在典型的NEDC循环工况下，对采用模糊逻辑能量控制策略和采用模糊神经网络能量控制策略的控制系统进行仿真，仿真结果验证了模糊神经网络能量控制策略的有效性。采用模糊神经网络能量控制策略，，能够同时保证发动机和电机工作在高效区域，从而提高了整车燃油经济性和排放性。论文对模糊神经网络能量控制策略的研究，对我国自主研发新型节能环保汽车，提高混合动力汽车能量控制系统设计水平，构建自主知识产权的汽车电子开发平台具有重要意义
[Abstract]:With the rapid development of world economy, energy crisis and environmental pollution have become increasingly prominent. Parallel hybrid vehicles (HEVs) have attracted much attention because of their advantages of environmental protection, energy saving and relatively mature technology. The energy control strategy of the parallel hybrid electric vehicle is an important factor that affects the energy consumption of the vehicle. Therefore, under the premise of satisfying the power performance of the vehicle, Optimizing the energy control strategy of parallel hybrid electric vehicle (HEV) is of great practical significance in controlling the energy crisis and realizing the sustainable development of the environment. In this paper, the energy control system of parallel hybrid electric vehicle is taken as the research object, and the distribution of engine torque and motor torque is studied deeply. On the basis of analyzing the driving system structure and combination mode of parallel hybrid electric vehicle, the mathematical model of engine, motor, battery, wheel and dynamic equation of transmission system are established. The optimal mathematical model of energy control strategy is constructed. Because the energy control system of parallel hybrid electric vehicle has the characteristics of dynamic nonlinearity, the fuzzy neural network algorithm is used to optimize the distribution of engine torque and motor torque. The energy control strategy based on fuzzy logic algorithm and the energy control strategy based on fuzzy neural network algorithm are designed respectively, which provides the theoretical basis for the simulation platform. In the energy control strategy of fuzzy neural network, according to the optimal mathematical model of the control strategy of parallel hybrid electric vehicle, a forward neural network composed of input layer, hidden layer and output layer is constructed by using compensatory neural network structure. The input layer of the network corresponds to the fuzzy process of the fuzzy logic algorithm, the hidden layer of the network corresponds to the fuzzy reasoning process, and the output layer of the network corresponds to the process of resolving the fuzzy logic. In order to solve the problem of the unity of the interface between the neural network node and the fuzzy logic input and output, the input and output variables are quantized by the quantization formula, and then the self-learning and adaptive ability of the neural network is utilized. The fuzzy rules and membership functions are generated automatically, and the center and width of the input and output membership functions of the neural network are optimized continuously. In order to improve the accuracy of the system and speed up the convergence, the neural network optimizes the energy control strategy by using the learning algorithm of dynamically adjusting the step size to compensate for the gradient descent. Taking the Toyota Prius car as an example, the backward simulation model of the whole vehicle is established under the environment of ADVISOR2002 software, which includes engine, motor, battery, transmission system and vehicle driving dynamics model. It provides a necessary simulation platform for the research and development of vehicle control strategy, and based on this simulation platform, under typical NEDC cycle conditions, The fuzzy logic energy control strategy and the fuzzy neural network energy control strategy are simulated. The simulation results verify the effectiveness of the fuzzy neural network energy control strategy. The fuzzy neural network (FNN) energy control strategy can ensure that the engine and motor can work in the high efficiency area at the same time, thus improving the fuel economy and emission performance of the whole vehicle. The research on the fuzzy neural network energy control strategy is of great significance to the independent research and development of new energy saving and environmental protection vehicles in China, to improve the design level of hybrid electric vehicle energy control system, and to construct the automobile electronic development platform with independent intellectual property rights.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：U469.7

【参考文献】