节点文献

分布式发电多逆变器并网孤岛检测技术研究

Research on Islanding Detection Technology for Multi-grid Connected Inverter Based Distributed Generation

【作者】 谢东

【导师】 张兴;

【作者基本信息】 合肥工业大学 , 电气工程, 2014, 博士

【摘要】 由于燃料能源的不断减少及其对环境危害的日益严重,人类把目光投向了太阳能、风能等清洁且储量巨大的可再生能源,基于可再生能源的分布式发电(distributed generation, DG)技术是电力行业今后发展的一个重要方向。孤岛现象是分布式发电系统的一个基本问题。所谓孤岛,是指主电网因故障或进行检修而处于失电状况时,分布式发电系统仍可能持续工作,并与本地负载连接处于独立运行状态。孤岛产生后,会威胁到电网维修人员的安全,影响配电系统的保护开关动作程序,在重合闸时可能对用电设备造成损害。因此,孤岛检测是分布式发电装置必须具备的一种保护功能。目前,并网逆变器型分布式发电系统的孤岛检测技术包括被动式和主动式孤岛检测两大类,但主要是针对这两大类检测方法在单台逆变器并网时的检测算法与参数优化的研究。随着大规模分布式发电系统的应用,实际上常常出现系统连接有多台并网逆变器的情形,多逆变器并网系统孤岛检测技术的研究越来越受到关注。为此,本文针对分布式发电系统在多逆变器并网条件下的孤岛检测技术展开了研究,具体研究内容如下:首先,对基于并网逆变器的孤岛检测技术中与本论文课题相关的一些基本内容展开了研究,主要包括:(1)对基于并网逆变器的孤岛检测方法进行了阐述,分析了被动式及主动式两类本地孤岛检测法,重点分析了目前应用非常广泛的几种主动移频式孤岛检测法的原理及其优缺点。(2)分析了基于Qf×f0坐标系的孤岛检测性能评估方法,给出了主动频率偏移法、Sandia频率偏移法以及滑模频率偏移法等孤岛检测方法的检测盲区(Non-Detection Zone, NDZ)边界曲线的求取方法。(3)建立了分布式发电系统的小信号模型,采用恒电流控制策略以及APS孤岛检测法分析了带正反馈式孤岛检测的逆变器控制策略对系统稳定性的影响,从而给出了解决维护系统稳定性与提高孤岛检测性能间矛盾的方法。其次,对现有多逆变器并网条件下分布式发电系统的孤岛检测方法及其检测性能进行了研究。对主动频率偏移法、Sandia频率偏移法以及滑模频率偏移法这几种主动移频式多逆变器并网孤岛检测法的性能进行了理论分析与仿真验证;对主动移频式多逆变器并网孤岛检测法的检测盲区问题与稀释效应问题展开了研究;对阻抗测量法、相关技术法及负序电流扰动法等多逆变器并网孤岛检测技术的原理及其性能进行了阐述。针对主动移频式多逆变器并网孤岛检测法的检测盲区问题与稀释效应问题,提出了解决策略,即利用数字相关原理检测公共耦合点电压与逆变器输出电流间的相位差角,根据检测到的相位差角对主动移频式孤岛检测算法中的有关参数进行自适应调整,最后对所提策略进行了仿真与实验验证。为进一步降低逆变器输出电流的畸变率,在上述孤岛检测算法改进措施的基础上,提出了基于锁相环的低电流畸变率改进SFS孤岛检测法。对于该种方法,首先分析了采用SFS法时电流畸变率的大小,详细推算了电压频率有偏差及频率无偏差两种情况下该法所对应的电流THD值;然后介绍了该改进SFS孤岛检测法中所使用的基于二阶广义积分器的虚拟两相锁相环的结构与工作原理;最后,分析了基于锁相环的低电流畸变率改进SFS孤岛检测法的原理、参数设置及改进方法,并给出了仿真与实验验证结果。最后,研究开发了基于小波变换与神经网络的多逆变器并网孤岛检测法。针对该种方法,首先简单介绍了小波变换与神经网络的基本理论;然后详细分析了该孤岛检测方法MATLAB仿真设计的原理,给出了6种不同状态下该孤岛检测方法在单逆变器及多逆变器并网条件下孤岛检测的仿真效果;最后,分析了该孤岛检测方法中小波变换Mallat算法及BP神经网络算法的DSP实现过程,给出了该孤岛检测法在孤岛与非孤岛状态的实验验证结果。

【Abstract】 Due to the continuous decrease of fuel energy and it’s increasingly serious damage to the environment, human sights on solar energy, wind energy and other clean and huge reserves of renewable energy, distributed generation technology based on renewable energy is an important direction for future development of the power industry. The phenomenon of islanding is a basic problem in distributed generation system. The so-called islanding refers to the main grid is tripped for breakdown or overhaul, the distributed generation system remains independent running status connected with the local load. After appearance, islanding will threaten the safety of power maintainer and power equipment, affect action programs of protection switch of distribution system, and damage to electrical equipment when reclosing. Therefore, islanding detection is a mandatory protection function of distributed generation device.At present, islanding detection Technologies of grid connected inverter based distributed generation system include two categories, the passive and active islanding detection, but it is mainly for the detection algorithm and parameter optimization of the two kinds of detection methods in single grid connected inverter. With the development of economy and technology, multi-inverter DG systems will be used more extensively, and the researchers have paid more attention to the islanding detection methods of multi-inverter systems. Therefore, the researches of this paper focus on the islanding detection technologies of multi grid connected inverter based DG systems, the specific contents are as follows.First of all, to start the research on some basic content of grid conn c contents are as follows.ected inverter based islanding detection technology in related to this paper, mainly include:(1) The islanding detection method based on grid connected inverter are described, local passive and active methods are analyzed, and emphasized the principle and the advantage and disadvantage of the active frequency drift methods which are used widely.(2) The performance evaluation Methods of islanding detection based on Qf×fo frame are analyzed, the calculating methods of non-detection zones (NDZs) boundary curve for active frequency drift method, Sandia frequency shift method and slip mode frequency shift method are given. (3) A small signal model of DG system is established, the constant current control strategy and APS islanding detection method are used to analyze the inverter control strategies with a positive feedback type islanding detection for it’s effects on the stability of the system, and gives the solution to the conflicts between maintaining system stability and improving islanding detection performance.Secondly, the existing islanding detection methods and their detection performance under the condition of multi grid connected inverter based DG system is studied. carries out the theoretical analysis and simulation verification for the performance of active frequency drift method, Sandia frequency shift method and sliding mode frequency shift method this several active frequency shift type multi grid connected inverter islanding detection methods; NDZ problem and the dilution effect of active frequency shift type multi grid connected inverter islanding detection methods is studied; The principle and performance of the impedance measurement method, the Correlation method and negative sequence current perturbation method for the multi grid connected inverter islanding detection are introduced.For the problem of NDZ and dilution effect of active frequency shift type multi grid connected inverter islanding detection methods, puts forward the strategies to solve these two problems, namely, using digital correlation theory to detect the phase difference between the voltage at the point of common coupling and the output current of the inverter, and carries out the adaptive adjustment for relative parameters of active frequency shift type islanding detection algorithm according to the detected phase difference, finally the proposed strategy has been verified through simulation and experiment. On the basis of the improved algorithm, the low current distortion SFS islanding detection method based on phase locked loop for multi grid connected inverter is put forward to further reduce the inverter output current distortion rate. For this method, the current distortion rate is firstly analyzed when the Sandia frequency shift method is used, the corresponding THD value of current of this method is Calculated in detail for two cases of voltage frequency has deviation and hasn’t deviation; And then introduces the structure and principle of virtual two-phase second-order generalized integrator based phase-locked loop which is used in the improved SFS islanding detection method of multi grid connected inverter; Finally, analyzes the principle, parameter settings and improvement for the low current distortion phase locked loop based islanding detection method, and the simulation and experimental results are presented. Finally, the multi grid connected inverter islanding detection method based on wavelet transform and neural network is developed. For this method, firstly introduces the basic theory of wavelet transform and neural network; Then a detailed analysis of the design principle of MATLAB simulation for this islanding detection method and simulation results of this method under six different states in case of single and multi grid connected inverter are given; finally, analyses the DSP implementation process of Mallat algorithm of wavelet transform and BP neural network algorithm for this islanding detection method, and gives the experimental results in islanding and non islanding state.

节点文献中: