变压器绕组变形诊断新方法的研究

发布时间：2018-10-31 09:08

【摘要】：变压器是电力系统的核心设备之一,其运行状况直接影响电力系统的安全稳定运行。绕组变形是变压器的主要故障类型之一,严重威胁电力系统正常运行。为准确诊断变压器绕组变形,减少变压器故障,在总结分析变压器绕组变形检测方法的研究现状和存在的一些问题的基础上,利用变压器运行和检修过程中测量到的电气信息,从不同角度研究了变压器绕组变形的诊断方法,研究内容包括以下几个方面:找出了一种基于频响阻抗法的变压器绕组变形检测方法。频率响应分析法在实际测试时,因现场噪声的干扰及接线长度等的影响,导致诊断的准确性受到限制。研究了变压器绕组在低频段和中频段的阻抗频率特性,以频率响应分析法的测试接线为基础,测量相关的电压和电流值后,计算和处理以获得阻抗频率曲线,对比曲线间的差异实现对变压器绕组变形的诊断。其诊断结果可以和频率响应分析法的判定结果互相验证,减少在频率响应分析法测试中由于现场环境干扰等造成的误判。仿真结果表明,该方法能有效识别变压器绕组不同类型的变形。提出了一种基于完整集成经验模态分解和变量预测模型模式识别的变压器绕组变形在线故障定位方法。暂态过电压陡峭的上升沿中包含大量的高频成分,当其冲击变压器时,电压行波在绕组变形位置的能量损耗和正常位置存在差异,这种差异会体现在行波的某些频率成分中。在绕组末端测量暂态过电压信号,采用具有自适应白噪声的完整集成经验模态分解方法处理数据以得到本征模态分量,计算相关系数作为故障特征量,经基于变量预测模型的模式识别方法训练后得到预测变量模型,对变形位置进行定位。仿真结果表明,本方法能可靠反映绕组变形位置,具有较高的实用价值。提出了一种基于变分模态分解和概率密度估计的变压器绕组变形在线检测方法。短路电抗是判断变压器绕组是否变形的重要依据,其在实际测量中受现场噪声干扰而呈现一定的随机性,从而影响对绕组状态的判断。首先,将变分模态分解应用于电气信号消噪处理,得到基波模态分量;然后,利用该基波模态分量在线计算短路电抗值;最后,对检测周期内求得的短路电抗样本,通过参数估计得到短路电抗参数正态分布的概率密度函数,根据其正态分布均值的估计值计算短路电抗偏差系数,评估绕组的当前状态。仿真验证了该方法能稳定地得到短路电抗估计值,避免噪声和设备测量误差对检测的干扰,从而可靠检测绕组变形。
[Abstract]:Transformer is one of the core equipments in power system. Its operation condition directly affects the safe and stable operation of power system. Winding deformation is one of the main fault types of transformer, which seriously threatens the normal operation of power system. In order to accurately diagnose transformer winding deformation and reduce transformer fault, on the basis of summarizing and analyzing the present research situation and some existing problems of transformer winding deformation detection method, the electrical information measured during transformer operation and maintenance is used. The diagnosis method of transformer winding deformation is studied from different angles. The research contents include the following aspects: a method of transformer winding deformation detection based on frequency response impedance method is found out. When the frequency response analysis method is tested, the accuracy of diagnosis is limited because of the interference of field noise and connection length. The impedance frequency characteristics of transformer windings in low and medium frequency bands are studied. Based on the test wiring of frequency response analysis, the relative voltage and current are measured, and the impedance frequency curves are calculated and processed. Contrast the difference between curves to realize the diagnosis of transformer winding deformation. The diagnostic results can be verified with the results of the frequency response analysis method, and the misjudgment caused by the field environment interference in the frequency response analysis method can be reduced. Simulation results show that this method can effectively identify different types of deformation of transformer windings. An on-line fault location method for transformer winding deformation based on integrated empirical mode decomposition and variable prediction model pattern recognition is proposed. The steep rise edge of transient overvoltage contains a large number of high-frequency components. When the transient overvoltage is impacted on the transformer, the energy loss of the voltage traveling wave in the winding deformation position is different from that in the normal position, which will be reflected in some frequency components of the traveling wave. The transient overvoltage signal is measured at the end of the winding. The integrated empirical mode decomposition method with adaptive white noise is used to process the data to obtain the intrinsic mode component, and the correlation coefficient is calculated as the fault characteristic. After training the pattern recognition method based on the variable prediction model, the prediction variable model is obtained, and the deformation position is located. The simulation results show that the method can reliably reflect the position of winding deformation and has high practical value. A method for on-line detection of transformer winding deformation based on variational mode decomposition and probability density estimation is proposed. Short-circuit reactance is an important basis for judging whether transformer windings are deformed or not. In actual measurement, short-circuit reactance presents a certain randomness due to field noise interference, thus affecting the judgment of winding state. Firstly, the variational mode decomposition is applied to the de-noising process of electrical signal to obtain the fundamental mode component, and then the short-circuit reactance is calculated on-line by using the fundamental mode component. Finally, the probability density function of the normal distribution of short-circuit reactance parameters is obtained by parameter estimation for short-circuit reactance samples obtained during the detection period, and the deviation coefficient of short-circuit reactance is calculated according to the estimated value of the mean value of short-circuit reactance distribution. Evaluate the current state of the winding. The simulation results show that the proposed method can stably obtain the short-circuit reactance estimation and avoid the interference of noise and measurement error, thus the winding deformation can be detected reliably.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM407

【参考文献】