短路电流相控分断的Prony零点预测算法研究

发布时间：2018-06-08 04:23

本文选题：短路电流 + 相控分断　；参考：《华中科技大学》2015年硕士论文

【摘要】：目前电力系统正向高压、联网、大容量方向发展,线路的短路电流也逐渐增大。因此,现有断路器已不能满足系统的开断需要,调整电网结构、发展限流装置等传统抑制措施会对系统规格及稳定性造成负面影响。而短路故障电流相控分断技术不仅可以减小触头烧蚀,延长设备电寿命,同时也可以提高现有断路器的短路开断水平,已经成为智能化开关电器的研究热点之一。快速、准确的还原短路电流波形,预测目标零点是实现短路故障电流相控分断技术的基础和难点。本文分析了现有预测算法的利弊,首次提出Prony算法在该领域的应用可行性。深入研究后指出了Prony算法的应用缺陷,并提出了一种更加稳定的类-Prony算法,通过过程优化,增强了算法的应用稳定性。最后结合Matlab中的算例检验结果,分析了采样频率、波长等因素对类-Prony算法的精度影响,形成了五种预测算法的综合对比结果。本文首先利用低压模拟短路故障的电流录波数据,对Prony算法及类-Prony算法的还原效果进行了对比。并通过多组数据处理结果的统计分析,得到类-Prony算法实际应用时可能的误差分布。最后结合三组实际电力系统短路故障录波数据,对本文类-Prony算法进行了进一步的零点预测效果检验。两种录波数据处理结果均显示该算法可以高精度地还原短路电流,零点预测误差满足要求。本文最后设计了可用于相控开关的控制装置,将常规设备选相策略与短路故障策略进行整合,应用更加灵活。并基于该装置的核心芯片,对类-Prony算法的执行时间进行了硬件检测:1kHz采样率下,算法可以在1ms内完成零点预测。同时,考虑到硬件求解精度的影响,本文算法在实际DSP芯片中的还原预测结果也通过算例得到了验证。
[Abstract]:At present, the power system is developing in the direction of high voltage, interconnection and large capacity, and the short-circuit current of the line is increasing gradually. Therefore, the existing circuit breakers can not meet the needs of the system, adjusting the power network structure, developing current limiting devices and other traditional suppression measures will have a negative impact on the system specifications and stability. The short-circuit fault current-controlled breaking technology can not only reduce the ablation of the contact, prolong the electrical life of the equipment, but also improve the short-circuit level of the circuit breaker, which has become one of the research hotspots of intelligent switchgear. Fast and accurate reduction of short circuit current waveform and prediction of target zero are the foundation and difficulty of short circuit fault current phase controlled breaking technology. In this paper, the advantages and disadvantages of existing prediction algorithms are analyzed, and the feasibility of Prony algorithm in this field is put forward for the first time. After in-depth research, the application defects of Prony algorithm are pointed out, and a more stable class of -Prony algorithm is proposed. The stability of the algorithm is enhanced by process optimization. Finally, the influence of sampling frequency and wavelength on the precision of the class-Prony algorithm is analyzed with the example of Matlab, and the comprehensive comparison results of the five prediction algorithms are formed. In this paper, the reduction effect of the Prony algorithm and the similar -Prony algorithm are compared by using the current recording data of the low-voltage analog short-circuit fault. Through the statistical analysis of the results of multi-group data processing, the possible error distribution of the class -Prony algorithm is obtained when it is applied in practice. Finally, combined with three sets of actual power system short-circuit fault recording data, the paper makes a further zero-point prediction effect test for the class -Prony algorithm in this paper. The results of two recorded data processing show that the algorithm can restore the short-circuit current with high accuracy and the zero prediction error can meet the requirement. In the end, the paper designs a control device which can be used in the phase control switch. The conventional equipment phase selection strategy and the short circuit fault strategy are integrated, and the application is more flexible. Based on the core chip of the device, the execution time of the class -Prony algorithm is detected by hardware at the sampling rate of 1 kHz. The algorithm can achieve zero prediction in 1ms. At the same time, considering the effect of hardware solution accuracy, the algorithm in the actual DSP chip restore prediction results are also verified by an example.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM713

【相似文献】