【作者】 杨波；

【导师】 何湘宁；

【作者基本信息】 浙江大学 ， 电力电子与电力传动， 2010， 博士

【摘要】 世界能源危机和环境污染已经成为制约人类社会可持续发展的严重问题。以太阳能、风能等为代表的分布式发电和并网逆变技术在发展可再生能源、实现我国节能减排中具有重要的战略意义和应用价值。在分布式发电系统中,并网逆变器作为能量传输的通道,如何实现尽可能高的变换效率和电能质量是一个重要的研究课题。论文围绕这两个关键问题,在总结、归纳当前研究现状和技术发展的基础上,主要从以下几个方面进行深入研究：无变压器型并网逆变器电路拓扑及其调制方式、并网逆变器死区补偿和消除方法、两级式并网逆变器建模和控制策略、并网逆变器电流控制技术等。首先,论文在分析无变压器型并网逆变器对地共模漏电流产生原理的基础上,推导了无变压器型并网逆变器共模谐振电路的数学模型,给出了无变压器型并网逆变器的无共模漏电流条件。根据无共模漏电流条件,论文提出一种对称H6结构的无变压器型并网逆变器拓扑结构,并基于该拓扑,实现了单极性SPWM和倍频SPWM调制策略。单极性调制有利于提高拓扑的变换效率和降低系统的散热难度；倍频调制则提高了拓扑的等效工作频率,使并网电流谐波含量和THD下降,有效改善了并网电能质量。论文还详细研究了该对称H6结构无变压器型并网逆变器在非理想条件下存在的并网电流与输出电压的相移问题、开关器件的死区时间问题和开关器件寄生结电容的平衡条件。其次,论文在分析电压源型逆变器死区效应产生原理的基础上,推导了理想条件下死区效应的数学模型,详细分析了功率器件开通和关断时间、导通压降等固有非线性特征及零电流箝位现象对逆变器死区效应的影响,并获得了死区电压导致的并网电流基波损失、基波相移和低次谐波注入的数学表达式。在上述基础上,论文提出一种基于无效器件原理的在线自适应死区消除方法。该方法通过在线跟踪负载变化,自适应地计算合理的电感电流过零区域宽度,能够较准确地实现电流过零区域和非过零区域中死区效应的分别消除,有效降低输出电压和电流的低次谐波含量和THD,改善并网电能质量,该方法对电流过零区域内电压、电流畸变的抑制效果尤为明显。利用DSP实现的在线自适应算法降低了对电流检测器件特性和精度的依赖,增强了该方法的工程实用性和可靠性。然后,论文从两级式并网发电系统出发,推导了高增益有源箝位软开关交错并联Boost变流器稳态特性更为准确的数学模型。模型证明该变流器的电压增益表达式与占空比、耦合电感匝比、耦合电感等效漏感、变流器开关频率以及负载大小有关,具有较高的精度,有利于提高两级式系统的控制精度和最大功率点跟踪效果。针对作为功率后级的并网逆变器,论文提出一种带双补偿环节的直接电流控制方法。该方法建立在PI控制器基础之上,通过增加两个前馈补偿环节,在不影响双环控制系统稳定性的前提下,能有效消除直流母线电压和电网电压各次谐波及随机扰动对并网电流THD的影响。为实现并网电流对正弦参考的无静差跟踪,论文还提出一种预测电流控制的改进模型,证明了传统预测电流控制的基本模型仅是该改进模型的一个特例。应用该改进模型的预测电流控制方法除能提高并网电流的跟踪精度,减小并网电流的谐波含量和THD外,还能有效抑制模型参数特别是滤波电感参数发生匹配失真时对系统稳定性造成的影响,降低并网电流发生振荡的概率,使控制系统鲁棒性加强。最后,利用上述研究成果,论文完成了并网逆变器在电梯能量回馈单元、光伏并网发电系统和绿色节能电子负载三种工程实例中的应用工作。产品测试和应用实践证明,研发产品的主要技术参数已达到或超过目前工业界主流产品的性能水平。更多还原

【Abstract】 The energy crisis and environmental pollution around the world have become two severe issues that impede the sustainable development of human society. The distributed generation typified by solar and wind energy and the grid-connected inverting technologies possess both strategic significance and application prospect in developing renewable energy and realizing energy saving and pollutant reduction. The power energy is transmitted by grid-connected inverters in distributed generation systems. Hence, the realization of high efficiency and power quality in such system is a significant subject of studies. Focused on the two crucial tasks, and based on the overview of the existing studies and technologies, this dissertation has made thorough research on the following aspects:topology and modulation strategy of the transformerless grid-connected inverter, dead-time compensation and elimination method of the grid-connected inverter, modeling and control scheme of the two-stage grid-connected system, and current control technology of the grid-connected inverter.Firstly, based on the generation principle of the common-mode leakage current, a mathematical model of the common-mode resonant circuit in transformerless gird-connected inverters is derived in this dissertation. The condition to meet no common-mode leakage current is obtained by adopting the model. And a novel transformerless grid-connected inverter topology with symmetrical H6-type configuration is proposed, which can guarantee not to generate the common-mode leakage current. The unipolar SPWM and double frequency SPWM strategies can be both applied in the proposed inverter. The high efficiency and convenient thermal design are achieved by adopting the unipolar SPWM. Moreover, the higher equivalent frequency and lower current ripple are achieved depending on the double frequency SPWM. Therefore the harmonic contents and THD of the output current are reduced greatly, and the power quality is improved accordingly. Furthermore, the influences of the phase shift between the output voltage and grid-connected current, the configurations of the dead time among all power switches and the balance conditions of the switches’parasitical junction capacitance are discussed detailedly. Secondly, the generation principle of dead-time effects is analyzed, and an ideal model of dead-time effects is derived in this dissertation. Furthermore, the nonideal characteristics of the turn-on time, turn-off time and conduction voltage of power switches are discussed in detail. The mathematical functions for describing the loss and phase shift of the fundamental component and low-frequency harmonic contents of the output current, which are induced by the dead-time voltage, are formulated. Based on the above analyses, a novel online adaptive dead-time elimination method is proposed. Through tracking the load change on-line and determining dynamically the current direction, the zero-crossing current region and the non-zero-crossing current region are precisely distinguished and the optimal width of the zero-crossing current region is adaptively calculated. Therefore, the dead-time effects are eliminated respectively in the two different regions of the whole sinusoidal modulation cycle. The proposed method can significantly restrain the output distortion in the zero-crossing current region and reduce the low-frequency harmonic contents and THD. Moreover, the online adaptive arithmetic is implemented by DSP chip and has little dependence on the precision of the current detection devices, which improves the practicability and reliability of the proposed method.Then, the topological structure and control strategy of two-stage grid-connected system are explored in this dissertation. As the first power-processing stage, an accurate steady-state model of the high step-up interleaved Boost converter with active-clamp circuits is derived, which correlates the voltage gain of the converter with the duty cycle, the turns ratio and leakage inductance of the winding-coupled inductors, the switching frequency and the load. A full-bridge inverter with bidirectional power flow is used at the second power-processing stage, which employs the direct current control strategy based on PI regulators. And two feedforward compensation units are added in to perform in the control loops to restrain the disturbances of the DC-bus voltage and grid voltage, and then achieve the low THD of the output current. Meanwhile, the proposed control method has no influence on the system’s stability. Furthermore, an improved predictive current control expression is proposed, which calculates the inverter voltages required to force the output current to follow the reference without the steady-state error. The algorithm not only has the advantages of precise current tracking, low harmonic contents and THD of the output current, but also reduces the probability of output current oscillation and enhances the system stability and robustness to the model parameter mismatch especially in the filter inductance estimate.Lastly, based on the above research results, the engineering applications of grid-connected inverters are achieved in the three projects of energy regenerative unit for the elevator system, grid-connected photovoltaic power system and regenerative electronic load. Furthermore, the production tests and the practical application have proved that the major technical parameters and performances of the grid-connected inverters have reached or exceeded those of the similar products in present industry.更多还原