【作者】 陈朝阳；

【导师】 杨文荣；

【作者基本信息】 河北工业大学 ， 电气工程， 2022， 硕士

【摘要】 为保障电力变压器安全可靠运行,需要制定动态检修策略,并量化故障带来的风险,开展变压器铁心与绕组机械状态在线监测具有重要意义。为了更好地实现振动分析法在变压器机械故障诊断中的应用,本文以一台型号为SC-30/10.5干式变压器和对应的试验样机为研究对象,开展理论分析、仿真建模、实验测试、特征提取、故障分类和在线诊断等方面研究,论文主要研究内容如下:从变压器铁心与长圆形绕组的受力展开研究,对叠片铁心的磁致伸缩力与麦克斯韦力进行受力分析,建立长圆形绕组辐向和轴向振动数学模型。建立变压器三维振动仿真模型,进行电场-磁场-结构力场的耦合计算;通过改变铁心和长圆形绕组的边界条件与几何参数实现变压器铁心松动、绕组松动和翘曲故障的仿真,通过对变压器铁心和绕组进行电磁场、振动特征分布和振动频谱图的分析发现,振动特征分布提取是变压器机械故障表征的关键,为后续变压器铁心与长圆形绕组机械故障诊断奠定了基础。过电压下的变压器铁心松动故障诊断。测试了变压器上铁心轭面上不同方向的振动信号,从振动熵、时域、频域及振动分布角度分析上铁心轭面振动特性规律,选择搭叠区与非搭叠区振动幅值最大的点作为松动故障诊断测点。分析预紧力对铁心不同方向振动的影响,将面外法线方向振动幅值与过电压相结合作为故障特征,实现对铁心正常、轻微松动与严重松动的诊断,对准确诊断铁心松动故障有实际意义。针对长圆形绕组稳定性差、易发生位移和变形的问题,提出一种基于灰狼优化算法的变压器长圆形绕组机械故障诊断的方法。对变压器样机的绕组进行正常、松动以及翘曲三种状态试验,获取薄弱位置点的小波包能量特征,采用灰狼算法优化支持向量机参数对变压器长圆形绕组机械状态的诊断,最终优化后的诊断综合准确率达到90%,提出的基于灰狼算法的机械故障诊断方法具有所用样本点少、准确率高的优点,为准确评估长圆形绕组机械状态提供可行性方案。更多还原

【Abstract】 To ensure safe and reliable operation of power transformer,dynamic maintenance strategies need to be developed.To quantify the risk posed by faults,it is important to carry out online monitoring of the mechanical condition of transformer cores and windings.In order to realize the application of vibration analysis method in transformer mechanical fault diagnosis,a model SC-30/10.5 dry-type transformer and the corresponding test prototype are used as the research object in the paper.It carries out research on theoretical analysis,simulation modeling,experimental testing,feature extraction,fault classification and online diagnosis,etc.The main research contents of the paper are as follows:The study is carried out from the force on the transformer core and the oblong winding.Magnetostrictive and Maxwell forces are analyzed in the laminated core.Mathematical models are developed for the radial and axial vibrations of the oblong winding.It builds three-dimensional vibration simulation models of transformers.The coupling of the electricmagnetic-structural force fields is calculated.The boundary conditions and geometric parameters of the core and oblong winding are changed,which realizes the simulation of transformer core loosening,winding loosening and warpage faults.The electromagnetic fields,vibration characteristics distributions and vibration spectrograms are analyzed in the transformer core and oblong winding.It is found that vibration feature distribution extraction is the key to transformer mechanical fault characterization,which lays the foundation for transformer core and oblong winding mechanical fault diagnosis.A loosening core fault in a power transformer under overvoltage is diagnosed.The vibration signal is tested on the yoke surface of the upper core of the transformer in different directions.The vibration characteristics of the upper core yoke surface are analyzed in terms of vibration entropy,time domain,frequency domain and vibration distribution.The points with the largest vibration amplitude in the overlapping area and the non-overlapping area are selected as the measuring points for loosening fault diagnosis.The effect of the preload force on the vibration of the core in different directions is analyzed.It is proposed to combine the amplitude of vibration in the normal direction with overvoltage as a fault characteristic.This is of practical significance for accurate diagnosis of core loosening faults.To address the problem of poor stability and easy displacement and deformation of oblong winding,a method of mechanical fault diagnosis of transformer oblong winding is proposed based on Gray Wolf Optimizer.The winding of the prototype transformer is tested in three states: normal,loose,and warped.The wavelet packet energy feature of the weakness is obtained.The Gray Wolf Optimizer is used to optimize the Support Vector Machine parameters for the diagnosis of the mechanical condition of the transformer oblong winding.The final optimized diagnosis has a comprehensive accuracy of 90%.The proposed mechanical fault diagnosis method based on the Gray Wolf Optimizer has the advantages of using few sample points and high accuracy,which provides a feasible solution for the accurate assessment of the mechanical condition of the oblong winding.更多还原