【作者】 于蓉蓉；

【导师】 魏学业；

【作者基本信息】 北京交通大学 ， 交通信息工程及控制， 2011， 博士

【摘要】 光伏发电和电动汽车是新能源领域的两大重要发展方向。本文以光伏发电系统中的光伏电池和光伏逆变器,以及电动汽车充电站为研究对象,对光伏电池外特性、光伏逆变器控制算法和电动汽车充电站可靠性等问题进行了系统深入的研究。论文首先以单体光伏电池和光伏电池板的构成关系为依据,推导出了二者的输出特性方程,给出了仿真模型建立方法,并在此基础上,推导出了复杂光照情况下光伏阵列的输出特性方程。根据输出特性方程,重点对基于最优梯度法的最大功率点跟踪算法进行了深入研究,提出了一种基于近似梯度法的最大功率点跟踪算法,该算法结合了最优梯度法自动调整搜索步长、自动判断搜索方向,以及干扰观测法硬件电路实现简单的优点,提高了算法的实用性和有效性。针对已有光伏并网逆变器模糊控制算法过多依赖专家经验且需要对隶属度函数、模糊规则反复调试的缺点,在建立逆变器输出级数学模型的基础上,提出了一种基于T-S模糊模型的并网电流调节算法,该算法可根据电路离散模型精确调整功率开关管占空比,实现并网电流对参考电流的快速且精确跟踪。针对T-S模糊控制算法无法处理电路未知或不精确参数的缺点,对光伏并网逆变器预测电流控制算法进行了深入研究,进一步提出了一种改进型预测电流控制算法。该算法通过对控制目标进行优化设计,减少了预测过程中的假设近似环节,从而降低了电感值未知导致的电路模型参数偏差对系统稳定性的影响,改善了预测算法的可靠性和准确性,提高了系统稳定所允许的模型电感和实际电感的不匹配程度。针对预测电流控制算法忽略线阻且未考虑干扰信号的缺点,提出了一种适用于光伏并网逆变器的自适应控制算法,该算法基于李雅普诺夫直接法进行设计,在达到控制目标和保证系统稳定的前提下,实现了根据实时采样数据对电路未知参数进行自适应调节,并在此基础上进一步研究了干扰信号对该算法的影响,实现了干扰信号的实时预测,从而极大增强了系统对噪声的抑制能力。从可靠性分析的角度对电动汽车充电站典型拓扑结构进行了研究,提出了一种多负载系统可靠度计算方法,基于该方法计算出了典型拓扑结构下充电站的可靠度和平均无故障工作时间。通过对典型拓扑结构的对比研究,权衡可靠性和经济性两项指标,提出了冷备用冗余供电方式充电站和分布式充电站,最终将冷备用冗余供电方式引入到分布式充电站中,提出了一种N+1冗余供电的分布式充电站结构,通过对这三种充电站进行可靠性对比分析,验证了N+1冗余供电的分布式充电站是综合可靠性和经济性两方面优势的最优方案。更多还原

【Abstract】 Photovoltaic power generation and electric vehicle are two important directions in the field of renewable energy. This thesis focuses on photovoltaic module (array) and photovoltaic inverter in the photovoltaic power generation system, and electric vehicle charging station, which includes external characteristics of photovoltaic module (array), control strategies of photovoltaic inverter and reliability of electric vehicle charging station.Firstly, this thesis derives the characteristics equations of photovoltaic cell and photovoltaic module based on their relation and gives the method of establishing simulation model, on the basis of which the characteristics equation of photovoltaic array under complex irradiation is derived. According to the characteristics equation, maximum power point tracking algorithm using steepest ascent method is deeply studied. And a new maximum power point tracking algorithm using approximate gradient method is proposed to automatically judge the searching direction and adjust the step size as steepest ascent method does, and to adopt simple hardware as perturbation and observation method does. The proposed algorithm can improve practicability and validity of traditional algorithms.Secondly, to overcome the defect of the existing fuzzy control algorithms of photovoltaic grid-connected inverter, such as relying on experience of experts excessively, adjusting membership and fuzzy rules repeatedly, a grid current regulating algorithm using T-S fuzzy model is proposed on the basis of establishing mathematical model of inverter. Simulation result demonstrates that the algorithm can adjust the duty cycle accurately and make grid current track the reference current rapidly and accurately.Thirdly, for the purpose of overcoming the defect of being unable to deal with unknown or inaccurate parameters for T-S fuzzy control, predictive current control algorithm of photovoltaic grid-connected inverter is deeply studied and an improved predictive current control algorithm is further proposed. The proposed algorithm is optimized through adjusting the control objective, which can reduce the influence had on stability by parameter deviation as a result of unknown inductance, improve the reliability and accuracy of algorithm and increase the dismatch of model inductance and actual inductance on the premise of stability. Fourthly, aiming at overcoming the defect of neglecting resistance and disturbance signals of predictive current control algorithm, a self-adaptive control algorithm suitable for photovoltaic grid-connected inverter is proposed. The proposed algorithm is designed based on Lyapunov’s direct method, which can realize adaptive adjustment of the unknown parameters according to realtime data, on the premise of reaching control objective and stability. Then the influence had on the proposed algorithm by disturbance signals is further studied, which can realize realtime prediction of disturbance signals and improve greatly the interference rejection capability.Finally, typical topologies of electric vehicle charging station are studied with reliability analysis methods. A kind of reliability calculation method suitable for multi-stress system is proposed, based on which both reliability and mean time to failure of typical topologies are calculated. Through comparison of typical topologies, taking account of reliability and economics, two kinds of charging station, namely cold-standby redundancy power supplying mode charging station and distributed charging station are proposed. Then through introducing cold-standby redundancy power supplying mode to distributed charging station, a kind of distributed charging station with N+1 redundancy is proposed. Through comparison of reliability for the three proposed kinds of topologies, distributed charging station with N+1 redundancy is verified to be the optimum structure integrating both reliability and economics.更多还原