基于多信息融合的电网故障诊断技术研究

发布时间：2018-03-21 06:13

本文选题：电力系统　切入点：故障诊断　出处：《华北电力大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着电网的不断发展，不同区域间的互联也越来越紧密，这就使得系统中发生故障对系统本身的影响也随之扩大。目前，电网自动化程度飞速发展，为故障信息的获取提供了更为便捷的条件。一旦发生复杂故障，控制中心将会有大量的报警信息迅速涌入，这种情况下要求调度员抓住报警实质，迅速正确地判断故障是十分困难的，误判、漏判的发生在所难免。因此，需要依靠实时、高效的电力系统故障诊断系统提供决策参考，为调度员决策提供辅助判据，以确保电力系统的安全运行。绝大多数故障诊断方法是利用保护动作、断路器跳闸等遥信量信息，采用某种智能算法来进行故障元件的识别，这就对开关和保护信息的完整性要求性比较高，，故障报警信息的完整性与准确性程度将对诊断结果产生较大影响。在实际电力系统中，存在着保护和断路器的拒动或误动及其信息传输过程中的干扰导致信息丢失等不确定性因素，故障诊断结果的准确性不可避免地会受到影响。在此背景下，本论文充分考虑故障后电气量信息变化的特征，利用电气量信息实时、准确的特点，在电网故障诊断中引入电气量分析，并通过多信息融合来进行故障的综合诊断。本文首先分析了故障后故障录波信息的特点，建立了基于希尔伯特-黄变换的电气量故障诊断模型，利用快速本征模态分解和希尔伯特变换，将电气量故障信息转换为定量的故障测度，从而进行故障识别。其次，深入考虑故障前后WAMS信息变化特点，建立了一种基于WAMS量测量信息的故障诊断模型，针对故障后不同元件各自的故障特征，分别建立了不同的判据模型，在拓扑分析结果的基础上进行信息的获取，可以有效得识别出故障元件。充分利用WAMS中的量测量信息，可以弥补传统故障诊断方法对于保护信息缺失、异常等情况下诊断不准确的缺点。然后，，考虑保护、断路器动作的可靠性，在解析模型中引入模糊度，基于开关量求取解析故障度，结合希尔伯特-黄变换分析取得的电气量故障概率表征，通过D-S证据理论进行信息融合，并依据模糊K-均值算法进行决策分析。实现了多源信息融合的电网故障诊断，以IEEE39节点系统为例通过仿真验证了所提方法的有效性。最后，基于云南省调OCS信息平台，进行了云南电网智能在线故障诊断系统的电网拓扑分析、故障信息的获取和故障诊断等程序的研究开发。
[Abstract]:With the continuous development of the power grid, the interconnection between different regions is becoming more and more close, which makes the impact of the system fault on the system itself expanded. At present, the degree of power grid automation is developing rapidly. This provides a more convenient condition for the acquisition of fault information. Once a complex fault occurs, the control center will have a large amount of alarm information coming in quickly. In this case, the dispatcher is required to grasp the essence of the alarm. It is very difficult to judge the fault quickly and correctly, and misjudgment is inevitable. Therefore, it is necessary to rely on the real-time and efficient power system fault diagnosis system to provide decision reference and provide auxiliary criterion for dispatcher's decision. To ensure the safe operation of the power system. Most of the fault diagnosis methods are based on remote signal information such as protection action, circuit breaker tripping and so on, and some intelligent algorithm is used to identify the fault components. Therefore, the integrity of switch and protection information is very high, and the integrity and accuracy of fault alarm information will have a great impact on the diagnosis results. In this context, the accuracy of fault diagnosis results will inevitably be affected by the uncertain factors such as protection and the failure or misoperation of circuit breakers and the interference in the process of information transmission, such as information loss. In this paper, the characteristics of electrical information change after fault are fully taken into account. By using the real-time and accurate characteristics of electrical information, the electrical quantity analysis is introduced into the fault diagnosis of power network, and the comprehensive fault diagnosis is carried out through multi-information fusion. In this paper, the characteristics of fault recording information after fault are analyzed, and a fault diagnosis model of electrical quantity based on Hilbert-Huang transform is established. The fast eigenmode decomposition and Hilbert transform are used. The fault information of electrical quantity is transformed into quantitative fault measure to identify the fault. Secondly, considering the characteristics of WAMS information before and after the fault, a fault diagnosis model based on the measurement information of WAMS quantity is established. According to the fault characteristics of different components after the fault, different criterion models are established respectively. Based on the results of topological analysis, the fault elements can be effectively identified, and the quantitative measurement information in WAMS can be fully utilized. It can make up for the shortcomings of traditional fault diagnosis methods in the absence of protection information and abnormal conditions. Secondly, considering the reliability of protection and circuit breaker operation, the ambiguity is introduced into the analytical model. Based on the analytical fault degree of switch quantity and the probability representation of electrical quantity fault obtained by Hilbert-Huang transform, the information fusion is carried out through D-S evidence theory. The fault diagnosis of power network based on multi-source information fusion is realized, and the effectiveness of the proposed method is verified by simulation of IEEE39 node system. Finally, based on Yunnan Province dispatching OCS information platform, the proposed method is implemented. In this paper, the power network topology analysis, fault information acquisition and fault diagnosis program of the intelligent on-line fault diagnosis system of Yunnan power network are studied and developed.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM73

【参考文献】