基于极化谱特征参数的油纸绝缘设备老化状态研究

发布时间：2018-03-09 08:12

  本文选题：回复电压　切入点：变压器　出处：《福州大学》2014年硕士论文　论文类型：学位论文

【摘要】：本课题是在国家自然科学基金项目《基于电路分析法和恢复电压响应的油纸绝缘老化诊断方法》(基金编号为61174117)资助下提出的,以油纸绝缘变压器的绝缘系统为主要研究对象,为了深入研究极化谱与回复电压特征参数和变压器绝缘状态之间的关系,搭建计算机模型和制作模拟实验平台,完成大量的仿真和实验,对油纸绝缘系统的回复电压特性进行探讨,并将本文获得的回复电压诊断判据思路与实际测试结果相结合,相互佐证了利用回复电压法对变压器绝缘状态诊断的可行性。因此,基于极化谱和回复电压特征参数的绝缘状态诊断法具有继续深入的探索价值和广泛的实际应用前景。本文研究的主要工作在于：(1)以扩展德拜介质响应模型为出发点,对如何准确辨识模型参数进行研究,首次在该参数辨识中引入分层思路,并提出分层粒子群优化算法,实现基于回复电压现场测试数据的等值模型参数辨识,通过与未改进的粒子群优化算法对比分析可知,其具有使用灵活、精确高效等特点。(2)在介质响应等值电路模型和参数辨识的基础上,仿真研究绝缘状态、极化谱、回复电压特征参数和等值模型参数之间的内在联系,并提出模型支路数与绝缘状态之间的对应关系,即绝缘状态越差,引用模型支路数则应有所增加；成功通过等值模型对回复电压极化谱的非标准化进行了仿真,揭示绝缘状态造成非标准极化谱的内部影响因素,提出对非标准极化谱的模型辨识研究思路。(3)根据实际油纸绝缘变压器绝缘结构,提出“Ⅱ”型模拟绝缘系统,基于模拟实验平台,实验研究了充电电压、充放电时间比等外设取值对测试数据的影响,并从水分、纸老化、绝缘纸层数和气隙(气泡)出发,进行了回复电压特性分析,目前运用回复电压法对含气隙的绝缘系统的研究还未见其他学者有所涉及,所以该研究方向具有一定的研究价值。(4)将现场测试结果与本文研究所得的绝缘诊断思路相结合,提出根据极化谱、回复电压峰值时间和初始斜率三者综合考量的诊断理念,具体项目如下：a)对相同电压等级不同容量的油纸绝缘变压器进行现场测试,提出容量接近的变压器测试数据才具有可比性；b)根据实际变压器不同绝缘状态下获取的极化谱和特征参数,提出该法可用于具体故障相的精确判别；c)根据现场测试,说明模型支路数确与绝缘状态具有相关性；d)首次将实际测试所得的非标准极化谱成功用于建立等值模型,并指出局部受潮是其非标准化的原因之一。
[Abstract]:This subject is proposed under the project of National Natural Science Foundation of China, "diagnosis method of Oil-Paper Insulation Aging based on Circuit Analysis and recovery Voltage response" (fund number 61174117). The main research object is the insulation system of oil-paper insulated transformer. In order to deeply study the relationship between polarization spectrum and characteristic parameters of recovery voltage and insulation state of transformer, computer model and simulation experiment platform are built, and a large number of simulations and experiments are completed. The characteristics of recovery voltage of oil-paper insulation system are discussed, and the diagnostic criterion of recovery voltage obtained in this paper is combined with the actual test results, which proves the feasibility of diagnosing the insulation state of transformers by the method of recovery voltage. The insulation state diagnosis method based on polarization spectrum and recovery voltage characteristic parameters has further exploration value and wide application prospect. The main work of this paper is to expand the Debye medium response model. This paper studies how to identify the model parameters accurately, introduces the delamination idea into the parameter identification for the first time, and puts forward the layered particle swarm optimization algorithm to realize the equivalent model parameter identification based on the field test data of the recovery voltage. By comparing with the unimproved particle swarm optimization algorithm, it can be seen that it has the characteristics of flexibility, accuracy and high efficiency. On the basis of the equivalent circuit model of dielectric response and parameter identification, the insulation state and polarization spectrum are simulated and studied. The relationship between the characteristic parameters of the recovery voltage and the equivalent model parameters is put forward, and the corresponding relationship between the number of the model branches and the insulation state is put forward, that is, the worse the insulation state, the more the number of the quoting model branches should be increased. The non-standardization of the reverse-voltage polarization spectrum is successfully simulated by the equivalent model, and the internal influence factors of the non-standard polarization spectrum caused by the insulation state are revealed. According to the insulation structure of oil-paper insulated transformer, the "鈪，

本文编号：1587705