低压交流串联电弧故障检测

发布时间：2018-08-26 12:51

【摘要】：能源互联网快速发展,电能的生产与利用方式越来越多样,现有电力系统电气保护体系中遗留的漏洞愈加凸显。交流串联电弧故障ASAF,作为现有电气保护漏洞,尚未得到合理解决,由其导致的安全事故频发。实时准确检测ASAF,并切断故障电路是避免电弧持续燃烧以酿成火灾等事故的有效途径。自主搭建实验平台,通过示波器采集实验电路电流传感器输出电压信号OVSCS,获得研究数据。数字化和智能化的ASAF检测算法主要由特征提取算法和故障识别算法组成。采用短时过零分析提取ASAF发生时OVSCS短时过零率特征。为降低噪声对提取结果的干扰,并最大限度提取有效特征,应用信息熵理论优化降噪阈值。为弥补短时过零分析不能保留OVSCS幅度信息的短板,采用多个降噪阈值提取OVSCS短时过零率,构成OVSCS短时过零率矩阵。设计模式匹配和隐马尔可夫模型HMM相结合的方法实现ASAF识别。为降低识别算法计算复杂度和建模难度,根据实验观察结果,定义ASAF在3种不同时间尺度下所表现出的状态,并依次经信号分帧、模式匹配、HMM识别、宏观状态判别4个过程完成ASAF识别。以低成本微控制器为硬件核心,以检测算法为软件基础,设计与实现电弧故障检测装置AFDD和ASAF检测试验装置。所实现AFDD体积小,成本低,使用附加条件少。所实现试验装置具有高精度调节电弧电极间距、自动重复实验、判别ASAF发生类型的功能。设计实验测试所实现AFDD。实验结果表明所设计AFDD在不间断供电工作状态下,可实时、准确检测ASAF,并可处理偶发交流串联电弧故障AASAF,且可靠性良好,发生“好弧”时不误判。
[Abstract]:With the rapid development of the energy Internet, the production and utilization of electric energy is becoming more and more diverse, and the loopholes left over in the existing electrical protection system of power system are becoming more and more prominent. As an existing electrical protection loophole, AC series arc fault (ASAF,) has not been solved reasonably, and the safety accidents caused by it occur frequently. Detecting ASAF, in real time and cutting off the fault circuit is an effective way to avoid continuous arc burning and lead to fire accidents. The experiment platform is built independently, and the research data are obtained by using oscilloscope to collect the output voltage signal of the current sensor of the experimental circuit by OVSCS,. Digital and intelligent ASAF detection algorithm is mainly composed of feature extraction algorithm and fault identification algorithm. Short-time zero-crossing analysis was used to extract the characteristics of short-time zero-crossing rate of OVSCS when ASAF occurred. In order to reduce the interference of the noise to the extraction results and extract the effective features to the maximum extent, the information entropy theory is applied to optimize the threshold of noise reduction. In order to make up for the short board whose OVSCS amplitude information can not be retained in short time zero crossing analysis, OVSCS short time zero crossing rate is extracted by using multiple noise reduction thresholds, which constitutes the OVSCS short time zero crossing rate matrix. Design pattern matching and hidden Markov model (HMM) are combined to realize ASAF recognition. In order to reduce the computational complexity and modeling difficulty of the recognition algorithm, the state of ASAF under three different time scales is defined according to the experimental results. ASAF recognition is completed by four processes of macroscopic state discrimination. The low cost microcontroller is used as the hardware core and the detection algorithm is used as the software base to design and implement the arc fault detection devices AFDD and ASAF. The realized AFDD is small in size, low in cost and less in additional conditions. The experimental device has the function of adjusting arc electrode spacing with high precision, automatically repeating experiments and discriminating the occurrence types of ASAF. Design and Test Institute to implement AFDD. The experimental results show that the designed AFDD can detect ASAF, in real time and accurately under the condition of uninterrupted power supply, and can deal with occasional AC series arc fault AASAF, with good reliability and no misjudgment when "good arc" occurs.
【学位授予单位】：沈阳工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM501.2

【参考文献】