数学形态学快速辨别变压器励磁涌流与内部故障的方法研究

发布时间：2018-07-03 20:24

本文选题：励磁涌流 + 内部故障　；参考：《华南理工大学》2014年硕士论文

【摘要】：作为电力系统中的重要主设备之一的变压器，其运行状况对于系统的安全可靠运行具有举足轻重的作用。纵联差动保护因具有高灵敏性、快速性和明确的选择性而被用作变压器主保护。但励磁涌流影响变压器差动保护的正确动作，如何快速辨别励磁涌流与内部故障已变成提高变压器差动保护正确动作率的关键问题。目前针对区分励磁涌流与内部故障的研究已经取得了一定成效，但是由于电力系统的运行状态是非常复杂的以及其他各种干扰因素的影响，现有的方法还存在一些不足。另一方面，考虑到数学形态学运算只涉及加减运算和极值运算，，并且非常适合于突变信号的特征提取，本文提出了采用数学形态学快速辨别变压器励磁涌流与内部故障的方法。该方法是根据励磁涌流与内部故障电流在波形特征上的区别，利用数学形态学提取相应的梯度特征和对称性特征来区分励磁涌流和内部故障。考虑到电流波形的复杂性，该方法由两部分组成，一部分是不考虑CT饱和情况时的电流识别，目的是快速识别出变压器内部故障使保护迅速动作以减少故障电流对变压器的损害，另一部分是考虑CT饱和情况时的电流识别，这可以保证电流识别的可靠性。为了验证所提方法的可行性和有效性，本文不仅利用PSCAD/EMTDC软件搭建相应模型得到了大量仿真数据，还采用了动模实验得到的仿真数据。试验结果证明了本文所提出方法能够有效区分励磁涌流与内部故障，尤其在出现CT饱和、合闸于轻微故障变压器等情况下，识别速度和可靠性优于二次谐波制动原理和间断角原理这两种现有使用的方法。此外，本文还研制了基于FPGA芯片的硬件电路板来实现所提出的电流识别方法。识别程序具体是运行在移植到Nios II的uC/OS II实时操作系统上的。硬件电路实现的测试结果与仿真数据验证结果一致，同时验证了所提出的电流辨别方法在实现硬件应用时的可行性。
[Abstract]:As one of the most important main equipment in power system, the operation condition of transformer plays an important role in the safe and reliable operation of the power system. The longitudinal differential protection is used as the transformer main protection because of its high sensitivity, rapidity and clear selectivity. However, inrush current affects the correct operation of transformer differential protection. How to quickly distinguish inrush current from internal fault has become a key problem to improve the correct operation rate of transformer differential protection. At present, the research on distinguishing inrush current from internal fault has made some achievements. However, due to the complex operation state of power system and the influence of other disturbance factors, there are still some shortcomings in the existing methods. On the other hand, considering that mathematical morphology only involves addition and subtraction operations and extreme operations, and is very suitable for feature extraction of abrupt signals, this paper proposes a method to quickly distinguish inrush current and internal faults of transformer by mathematical morphology. According to the difference between inrush current and internal fault current in waveform characteristics, the gradient feature and symmetry feature are extracted by mathematical morphology to distinguish inrush current from internal fault. Considering the complexity of the current waveform, the method consists of two parts, one is current recognition without considering CT saturation, The purpose of this paper is to quickly identify the internal faults of the transformer and make the protection operate quickly to reduce the damage of the fault current to the transformer. The other part is to identify the current in the case of CT saturation, which can ensure the reliability of the current recognition. In order to verify the feasibility and effectiveness of the proposed method, this paper not only uses PSCAD / EMTDC software to build the corresponding model to obtain a lot of simulation data, but also uses the simulation data obtained from the dynamic simulation experiment. The experimental results show that the proposed method can effectively distinguish inrush current from internal fault, especially in the case of CT saturation and closing on a slight fault transformer. The speed and reliability are better than the second harmonic braking principle and the discontinuous angle principle. In addition, the hardware circuit board based on FPGA chip is developed to realize the proposed current recognition method. The recognition program runs on the UC / OS II real-time operating system transplanted to Nios II. The test results of the hardware circuit are consistent with the simulation data, and the feasibility of the proposed current discrimination method in the implementation of the hardware application is also verified.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM407

【参考文献】