SVM在变压器继电保护方面的研究能研究

发布时间：2018-05-06 09:43

本文选题：支持向量机SVM + 变压器继电保护　；参考：《安徽理工大学》2017年硕士论文

【摘要】：当前,电力系统在国民经济发展过程中扮演着着及其重要的角色。并且对社会的各个行业的进步都有着长远的影响。其中变压器保护更是成为整个电力网络系统研究中的重中之重。当前,除了变压器自身的系统的复杂性、人为损坏及外在不可抗因素之外,针对加强变压器的高灵敏性,验证故障类型及故障位置,提升变压器的可靠性和精确性,更是当前变压器保护探索的热点方向。变压器继电保护故障诊断问题的处理关键在于如何快速鉴别故障类型,确定故障的具体位置及其故障元件。作为电力系统中的重要环节,目前变压器保护大多还是采用传统的差动保护的方法,对于变压器中励磁涌流造成的保护拒动、误动情况,依然难以有效解决。当前对于变压器保护的误动等保护动作准确率不高等问题依然很严峻。因此,针对上述因素,本文结合目前国内外针对变压器继电保护的研究现状,提出一种基于支持向量机(SVM)的变压器故障诊断模型。首先,根据支持向量机的基本原理,分析核函数参数优化的算法,寻求最优的支持向量机相关参数和惩罚参数组合。进而,建立一种基于PSO-SVM算法的变压器故障诊断模型,并针对励磁涌流形成原因,进行建模仿真。通过PSCAD/EMTDC电磁仿真软件对变压器故障诊断模型进行仿真,并获取变压器样本数据,进而利用MATLAB中LIBSVM工具包对样本数据进行训练、测试,最终获取故障诊断结果。本系统可以更加简单高效的预测出故障类型,增强了变压器故障诊断的正确性,大大减少了励磁涌流引起的变压器的误动和拒动,使得变压器保护模型更加简单化,智能化。为以后变压器保护发展方向指出了一条新方向。
[Abstract]:At present, power system plays an important role in the development of national economy. And has a long-term impact on the progress of all sectors of society. Transformer protection has become the most important part in the research of power network system. At present, in addition to the complexity of the transformer's own system, artificial damage and external irresistible factors, in order to enhance the high sensitivity of the transformer, verify the fault type and fault location, and improve the reliability and accuracy of the transformer, Is the current transformer protection exploration hot spot direction. The key to deal with the fault diagnosis of transformer relay protection lies in how to quickly identify the fault type, determine the specific location of the fault and its fault components. As an important part of power system, transformer protection mostly adopts the traditional differential protection method at present. It is still difficult to effectively solve the protection failure and misoperation caused by the inrush current excitation in the transformer. At present, the protection accuracy of transformer protection is not high. Therefore, in view of the above factors, this paper puts forward a transformer fault diagnosis model based on support vector machine (SVM) combined with the current research situation of transformer relay protection at home and abroad. Firstly, according to the basic principle of support vector machine, the optimization algorithm of kernel function parameters is analyzed, and the optimal combination of related parameters and penalty parameters of support vector machine is obtained. Furthermore, a transformer fault diagnosis model based on PSO-SVM algorithm is established. The transformer fault diagnosis model is simulated by PSCAD/EMTDC electromagnetic simulation software, and the transformer sample data are obtained. Then the sample data are trained and tested by LIBSVM toolkit in MATLAB, and the fault diagnosis results are finally obtained. The system can predict fault types more simply and efficiently, enhance the correctness of transformer fault diagnosis, greatly reduce the misoperation and rejection of transformer caused by inrush current, and make the transformer protection model more simple and intelligent. A new direction is pointed out for the future development of transformer protection.
【学位授予单位】：安徽理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TM774

【相似文献】