【作者】 卫国爱；

【导师】 全书海；

【作者基本信息】 武汉理工大学 ， 动力机械及工程， 2010， 博士

【摘要】 质子交换膜燃料电池(PEMFC)是燃料电池电动汽车的主要动力源,而空气供给系统是PEMFC的主要组成部分之一,其空气流量和空气压力不仅影响燃料电池堆化学反应速度和质子交换膜性能,而且影响燃料电池堆发电效率和负载能力。由于空气供给系统存在较强的非线性、参数强耦合性,对其建模与控制非常困难。为了提高空气供给系统的响应速度,保证燃料电池运行在最佳工作状态,本文开展车用PEMFC空气供给系统建模及控制策略研究,主要研究内容及成果如下：根据Elman动态神经网络对非线性模型的自适应辨识能力,提出了基于Elman神经网络的空气流量、空气压力目标值与燃料电池输出功率之间的预测模型,并通过仿真验证了预测模型的有效性。通过对空气流量、空气压力给定值的预测,可有效提高车用PEMFC空气供给系统的动态响应速度,达到较好的控制效果。针对空气流量和空气压力之间的非线性和强耦合性,设计了空气流量、空气压力的解耦矩阵,并采用递推辨识算法,实时辨识空气流量、空气压力控制通道及其控制变量耦合通道的模型参数,通过实时调节解耦矩阵的参数,使空气流量、空气压力的控制相对独立,彼此不受或少受另外一个控制变量的影响。车用PEMFC空气流量随燃料电池堆输出功率的变化而改变,为提高空气流量的响应速度,提出了基于空气流量机理模型的Fuzzy-PID复合控制策略,通过设定控制阀值和控制参数的整定,使Fuzzy-PID复合控制既具有Fuzzy控制的快速性,又具有PID控制的精确性,改善了空气流量的控制性能。仿真结果表明,空气流量采用Fuzzy-PID复合控制时,其响应时间缩短为采用传统PID控制时的二分之一。车用PEMFC空气压力不仅与燃料电池堆输出功率的变化有关,而且与空气流量直接耦合,因此控制难度较高。提出了空气压力的神经PID控制策略,利用神经网络对非线性系统的辨识能力,在线辨识空气压力控制回路模型,并采用神经PID控制,通过神经网络的自学习和加权系数的自调整,自动调整控制参数kp、ki、kd,使空气压力控制回路的稳定状态对应于最佳PID控制参数,仿真结果表明,空气压力采用神经PID控制,对燃料电池堆输出功率和空气流量的变化具有自适应性。综上所述,本文针对车用PEMFC空气供给系统,建立了空气流量、空气压力给定值预测模型,并采用对角矩阵解耦法,设计了空气流量、空气压力解耦矩阵,根据空气流量、空气压力控制回路特点,分别采用机理建模和基于实验数据的实验建模方法,建立了空气流量、空气压力控制回路的控制模型,根据空气供给系统的建模方法不同,采用不同的控制策略来研究,使得空气流量对燃料电池堆输出功率的变化具有较快的响应速度,空气压力对燃料电池堆输出功率和空气流量的变化具有较强的适应性,仿真结果表明了所采用控制策略的有效性。此空气供给控制系统可以满足车用PEMFC的实际需求。更多还原

【Abstract】 Proton exchange membrane fuel cell (PEMFC) is a main powr source of electric vehicles in fuel cells. And air supply system is one of the mainly constitute of PEMFC, its air flow and air pressure affected fuel cells’ stack not only electrochemistry reactivity rate and proton exchange membrane’s performance, but also generate electricity efficiency and load ability. It’s very difficult to modeling and control for air supply system as its stronger nonlinear and coupling. In order to improve air supply rate to insure fuel cells running in best state, the paper studied on modeling and control strategy of air supply system in fuel cells. The main studying content and production was as follows.It put forward prediction model between target value of air flow and air pressure with fuel cells’output power basing on Elman dynamic ANN, according its self-adapting distinguish for non-linear model. Simulation approved that its prediction model was validity. It advanced dynamic responding speed of air supply system of PEMFC in electric vehicle, and had better control result.It designed decoupling matrix of air flow and air pressure aim at its non-linear and strong coupling. It adopted recursion identification arithmetic to real time identify model parameters of air flow, air pressure control channel, and its control variable coupling channel. By adjusting parameters of decoupling matrix, it made air flow, and air pressure control be independent, the one didn’t affect by the other.Air flow of PEMFC in electric vehicles changed following output power of fuel cells. For advancing its responding speed, it brought forward Fuzzy-PID compound control strategy of air flow basing on its mechanism-model. By setting control limit value, and adjusting control parameters, Fuzzy-PID compound control improved contol performance of air flow for it had speediness of fuzzy control and accuracy of PID. Simulation result showed that air flow responding time adopting Fuzzy-PID compound control could shorten half than adopting traditional PID.Air pressure of PEMFC in electric vehicles changed not only following with fuel cells’output power, but also coupling with air flow, so it’s difficult to control. It advanced neural-PID control strategy of air pressure so as to identify model of air pressure control loop online using ANN identifying non-linear system. And it adopted neural-PID control to adjust control parameters kp, ki and kd, through self-studying of ANN and self-regulating of authority coefficient, so that stable states of air pressure control loop were better corresponding to PID control parameters. Simulation result showed that air pressure adopting neural-PID control have adaptability for fuel cells’ output power and air flow.On all accounts, the paper founded model of given value prediction of air flow and air pressure in PEMFC air supply system of electric vehicles. And it designed decoupling matrix of air flow and air pressure using method of on the cross matrix. It set up different control model basing on mechanism mode and experiment model according characteristics of air flow and air pressure control loops. It adopted different control strategies according different model to make air flow quickly responding output power changes, and make air pressure having stronger adaptability for output power and air flow changes. Simulation result showed that the control strategies were validity. The air supply control system could satisfy practice need of PEMFC in electric vehicles.更多还原