【作者】 孙尧；

【导师】 粟梅；

【作者基本信息】 中南大学 ， 控制科学与工程， 2010， 博士

【摘要】 矩阵变换器因其在结构和功能上的诸多优点,是一种极具发展潜力的电力变换装置,是近年来的研究热点。尽管关于矩阵变换器的研究成果很多,但依然存在一些重要问题未能十分满意地解决,因此限制了矩阵变换器大规模工业应用的步伐。这些问题包括调制策略复杂、系统成本高、稳定性问题、非线性问题和无功功率问题等共性问题以及具体应用问题。本文针对调制策略,非线性补偿,系统稳定性,衍生拓扑——四脚矩阵变换器的调制与控制,以及矩阵变换器在风力发电领域中的低成本应用等方面进行了研究,并取得了系列进展。论文的研究工作包括：在调制策略方面。一方面,将矩阵变换器的两类典型的调制策略——双电压合成调制和空间矢量调制策略统一起来,从数学的角度证明了它们的一致性,并从开关序列安排、原点开关的基本概念、功率因数控制、共模电压、输入电压不平衡抑制等方面揭示了两种调制策略的一致性,使得这两类调制策略中研究成果可以互为借鉴,相互促进；另一方面,针对现有的调制策略中算法复杂等问题,提出了一种简单通用的基于数学构造的矩阵变换器调制策略,其核心在于调制矩阵的求解。根据虚拟整流、逆变过程,可初步得到一组满足输入输出功能的过渡调制矩阵,然后在过渡调制矩阵的基础上通过叠加偏置信号逐步构造出既满足输入输出关系又满足物理约束的调制矩阵。该调制方法简单实用、电压传输比高,通用性强,其通用性体现在偏置信号的选择上,不同偏置信号的选择对应不同性能的调制策略。针对双级矩阵变换器的非线性问题。论文首先深入分析了其非线性根源,发现窄脉冲、器件电压降、滤波电容电压纹波以及死区等均是影响系统非线性的重要原因,而且它们与调制策略、拓扑结构和运行状态密切相关。尤为重要的是,分析发现现有经典算法在理论上必然导致窄脉冲问题的出现,而窄脉冲对系统的安全运行构成隐患,同时还会引起输入输出波形畸变。为保证系统可靠运行和补偿非线性所带来的负面影响,本文提出了一种修正调制算法,克服了已有调制算法在实现上的固有缺陷,有效地避免了窄脉冲问题的出现。并提出保证脉冲局部对称性的死区补偿策略,提高了输入输出波形质量。矩阵变换器稳定性问题是该系统的一个根本问题。在恒功率负载的情况下,矩阵变换器闭环调制可能会导致系统出现不稳定现象。受阻抗比判据的启发,本文提出了一种基于构造的矩阵变换器稳定化方法,即构造适当的修正项以提高矩阵变换器的输入阻抗进而改善系统稳定性,论文给出了构造的依据和基本原则,并介绍了几种不同类型的构造示例。本文还揭示了前期研究人员对相关问题所提方法的本质,并将本文的方法和它们进行了对比分析,发现基于构造的方法具有足够的自由度,能进一步扩大系统的稳定域,且具有通用性。针对矩阵变换器驱动不平衡负载时零、负序负载扰动电流对输出电压的影响问题,本文提出一种基于载波调制的双级四脚矩阵变换器解决方案。在调制策略层面,为了提高功率变换系统的综合性能,提出了一种基于最优马尔可夫链的随机载波四脚矩阵变换器调制策略。该调制策略是一种以随机过渡矩阵为优化变量,以开关次数和输出电压波形质量为综合优化目标的优化调制策略,同时能一定程度地减轻电磁干扰问题。在控制策略层面,通常的办法是将系统分解为正序、负序以及零序部分,然后在各个同步旋转坐标系下分别采用PI调节器。显然系统有多个控制器需要调节,而且坐标变换繁杂。本文提出一种基于鲁棒自适应——反步控制法的控制方案,这是一种基于李亚普诺夫函数的递推控制,可以严格保证系统的稳定性,算法简单,易于实现。关于矩阵变换器在风力发电中的应用问题,本文提出了两种新颖的低成本发电系统方案。第一种方案为基于超级疏松矩阵变换器的永磁同步发电机的风力发电系统。其特征是能量从矩阵变换器的整流级流向电网。为了取得良好的控制效果,采用了反馈线性化、奇异扰动原理以及H∞等控制方法设计了该发电系统的控制器。第二种方案为基于逆—疏松矩阵变换器的永磁同步发电机的风力发电系统。其基本特征是能量从矩阵变换器的逆变器往整流器传输,通过逆变器可以直接控制永磁同步发电机,因而控制简单有效。仿真验证了两种方案的正确性和可行性。更多还原

【Abstract】 Matrix converter has attracted more and more attention recently due to many advantageous features such as bidirectional power flow, controllable input power factor, sinusoidal input and output currents, and compactness in structure;it is a potential power converter. Due to much effort of researchers, lots of achievements have been made up to now. However, there are still many hard problems which include the complexity of modulation strategy, the high cost, nonlinearity, reactive power and its applications involving matrix converter. In this paper, the issues on modulation strategy, nonlinearity compensation, stability, the modulation and control of four-leg matrix converter and the research for matrix converter used in wind energy conversation system (WECS) will be studied in detail.Based on the in-depth analysis of the basic principles of the double line-to-line syntheses and the space vector modulation (SVM), the consistency of duty ratio calculation of the two modulation strategies have been proved from the mathematical point of view. Then, the consistency is revealed from the issues of switching sequence, the concepts of the’origin switch’, power factor and common mode voltage and the ability to suppress the effect to output voltage due to unbalanced grid.Inspired by fictitious rectifier and inverter process of matrix converter, a matrix converter modulation strategy based on mathematical construction is proposed. The core of this method consists in solving the modulation matrix converter. Based on the rectifier and inverter process, it is easy to get a transient modulation matrix, but the physical constraints could not meet here. Then the offsets are added into each row of modulation matrix to satisfy the physical requirements. The offset is a degree of freedom, different offsets lead to different modulations, so the proposed method is a general modulation.Aim to solve the nonlinearity of matrix converter, the nonlinearity sources and features of two-stage matrix converter (TSMC) output voltage are analyzed in details, it is found that the nonlinearity of TSMC has many special features such as narrow pulse, device voltage drops and ripples of filter capacitor which is concerned with modulation strategy, operating mechanism and topologies. To meet the end of reliable operation and compensate the negative affects of nonlinearity, revised modulation strategy and nonlinearity compensation algorithms are proposed. It is found that the classic modulation will lead to narrow pulse which may pose threaten to matrix converter, an improved modulation strategy is proposed by using one time varying distribution factor to overcome the constraints from real practice, and the minimum narrow pulse restriction could be met without special handling. As a result, the waveform quality of TSMC both in output and input is improved greatly. And a dead time compensation method which could keep the local symmetry of pulse is proposed. The experimental results verify the propose method.Stability is a basic requirement of matrix converter, Inspired by the impedance ratio criterion, a general constructive method is proposed, improving the stabilization regions of matrix converter. Based on analysis, it is found that, by using constructive method to modify the output reference voltage and rectifier modulation vector could change the input impedance (or admittance) of matrix converter on D-axis and Q-axis, respectively. As a result, the instability, caused by negative impedance, is eliminated. Plus, stabilization regions are enlarged. Furthermore, the underlying principles of constructive method are stated in detail and several instructive examples which comply with these principles are demonstrated. The essence of some existing methods which are presented to improve the stability of matrix converter is unveiled, and difference and similarity between existing methods and the proposed ones are discussed.In order to reject the zero sequence and negative sequence disturbance due to unbalanced load, a solution using two-stage four-leg matrix converter based on carrier modulation strategy was proposed. In order to improve the comprehensive performance of power converter system, a randomized carried modulation strategy based on Markov chain for four-leg matrix converter was proposed. This modulation is an optimal randomized carried modulation strategy taking the random transient matrix as variables, and taking the output voltage waveform quality and switching times as optimized objectives. And it could reduce the EMI problem of matrix converter.The unbalanced and distorted load currents pose more challenges to four-leg matrix converter’s operation. From the perspective of control, this is a servo control problem with external disturbance. To solve the problem, the most commonly used method is to transform AC system in stationary frame coordinate into DC system in synchronous frame coordinate. Then, the classic proportional integral (PI) controller based on cascaded control will work well; however, this scheme is too complicated A new method using robust adaptive back-stepping control is presented for indirect four-leg matrix converter. It is robust against uncertain loads and control input disturbance due to the interaction between input filter and matrix converter.About the application of matrix converter in WECS (wind energy conversation system), two kinds of WECS based on supper sparse matrix converter (SSMC) and permanent magnetic synchronous generator (PMSG) are proposed. For the first scheme, the key feature is that the energy is transferred from rectifier to inverter of super sparse matrix converter. To design a global stable and robust controller, Feedback Linearization theory and H-infinity control are used. The other WECS is composed of PMSG and a very sparse matrix converter (VSMC). In this application, its distinct feature is that the energy generated by PMSG is transferred from inverter to rectifier of VSMC. The PMSG could be controlled directly, so it is easy to control. Because VSMC has only 12 active power switches, this WECS is low cost and reliable.更多还原