旋转超声磨削平行砂轮复合变幅器设计与试验*

为了获得大负载平行砂轮复合变幅器的设计方法,将复合变幅器简化为复合变幅杆与多阶环盘组成。利用Mindlin中厚板理论,求解了各振动环盘的位移、转角、弯矩和剪力解析表达式,并通过复合变幅杆各振动单元间的力、位移的边界条件和连续条件,以及复合变幅杆和环盘两者结构间力与位移的耦合条件,建立了复合变幅器的数学模型和频率方程。针对频率方程进行编程求解 ,设计了平行砂轮复合变幅器。通过有限元仿真分析、实物阻抗特性试验、谐振试验验证。试验结果表明：平行砂轮复合变幅器的试际谐振频率、节圆位置与理论设计的误差小于5%,验证了其理论设计方法的正确性。为研制大负载平行砂轮复合变幅器提供了一种有效的设计方法。     

散热片对变压器声辐射的影响及优化分析*

油浸式配电变压器表面布有大量散热片结构,使得油箱振动声辐射变得复杂。该文研究了散热片结构对油浸式变压器辐射声场的影响,在此基础上提出了散热片结构尺寸的优化措施。建立了考虑流固耦合的变压器振动和声学分析模型,提出了散热片声场效应评判准则,基于模态分析和频响分析探讨了散热片对变压器振动声辐射影响的振动效应、声源效应和声障效应。实验数据和仿真分析结果表明,变压器散热片对变压器振动声辐射的影响不可忽视,尤其是声源效应。采用遗传算法对散热片分布和尺寸参数进行优化分析,计算结果表明,通过优化散热片可以有效降低变压器表面均方振速和辐射声功率,可为变压器噪声控制和优化提供依据。     

拼接式编码器测角误差分析及试验

根据望远镜测角系统的精度需求,分析了影响拼接式编码器测角精度的误差源,以及多读数头测角消差原理,确定了采用相位相差90°的4读数头的测角方式。在某望远镜方位轴系转台进行了逆时针和顺时针方向的测角试验,通过对测角数据进行谐波分析并补偿后,得到两组实验测角误差RMS分别为0.34"、0.38"。实验表明,相位相差90°的均布4读数头的测角方式消除了由轴系误差、编码器安装位置误差和钢带安装盘部分圆度误差对测角精度的影响,实现了亚角秒级测角的目的。研究结果可用于大口径望远镜设计阶段的误差分析与分配、预估测角精度,为降低设计和加工误差提供参考。     

Analysis of polychlorinated biphenyls in transformer oils by automated on-line coupling reversed phase liquid chromatography-gas chromatography using the through oven transfer adsorption desorption (TOTAD) Interface

An automated method for the direct analysis of polychlorinated biphenyls (PCBs) in transformer oil is presented. The proposed method uses the TOTAD (through oven transfer adsorption desorption) interface for the on-line coupling of reversed phase liquid chromatography and gas chromatography (RPLC-GC). In this fully automated system, the oil is injected directly with no sample pre-treatment step other than dilution with n-propanol and filtration. In the LC step, PCBs are separated from other components of the oils using methanol/water (90:10 v/v) as mobile phase, at a flow rate of 1?mL?min^?1. The LC fraction containing the PCBs is automatically transferred to the GC by the TOTAD interface and GC analysis enables the separation of the PCB congeners. The proposed method is compared with two other methods: the European Norm (UNE-EN-61619) and that of the American Society for Testing and Materials (ASTM) (D4059-00). The proposed method practically eliminates the time-consuming sample preparation step and avoids errors caused by sample manipulation. The total PCB concentrations obtained with the three methods are similar.     

Investigation on Noise Pollution Comprehensive Treatment of Distribution Transformer in Living Area

In the current paper, which deals with the noise pollution excited by distribution transformers in the living area, a comprehensive treatment scheme is put forward for the purpose of reducing the sound pressure level emitting into the environment. In accordance with the associated test standard, the sound pressure levels of distribution transformer and surrounding environment are not only tested but analyzed as well. The measurements were carried out with the frequency analysis of the 1/3 octave resolution, with the center frequencies at 125 Hz, 250 Hz, 400 Hz, and 500 Hz. As illustrated, on the basis of the measurement results, the frequency of noise at 500 Hz of distribution transformer causes the major noise pollution in the surrounding environment. This measurement result is in line with the noise frequency characteristics of distribution transformer. There are two transmission routes of noise: i) the noise excited by distribution transformer transmits by means of the wall of distribution room, and ii) part of noise spreads through the ground of distribution room. Accordingly, acoustic shield and vibration isolation device are applied for the reduction of the low frequency noise emitted through the above two paths. Aimed at applying the appropriate acoustic material and vibration mounting, the evaluation of the noise reduction and vibration absorption is carried out in accordance with the sound and vibration insulation theory. Following the noise treatment, the transformer and environment noise are measured again. The corresponding findings shed light on the fact that the sound level satisfied the requirement of limits of the ordinance. The proposed noise treatment scheme can be applied to the existing power distribution facilities for controlling the sound levels that reach a point where it is comparatively more unobjectionable.     

Prediction of Outdoor Noise Propagation Induced By Single-Phase Power Transformers

Outdoor power transformers are one of the most pervasive noise sources in power transmission and distribution systems. Accurate prediction of outdoor noise propagation plays a dominant role for the evaluation and control of noise relevant to the transformer stations. In this paper surface vibration tests are carried out on a scale model of a single-phase transformer tank wall at different excitation frequencies. The phase and amplitude of test data are found to be randomly distributed when the excitation frequency exceeds the seventh mode frequency, which allows the single-phase power transformer to be simplified as incoherent point sources. An outdoor-coherent model is subsequently developed and incorporated with the image source method to investigate noise propagation from single-phase power transformers, due to the occurrence of multiple reflections and diffractions in the propagation path of each point source. The proposed model is used to calculate the sound field of the power transformer group by exploiting the additional phase information. In comparison with the ISO9613 model and the boundary element method, it is found that the proposed coherent image source method leads to more accurate prediction results, and hence better performance for the prediction of the outdoor noise induced by single-phase power transformers.     

基于LabView的光纤电流互感器一体化测试系统设计

光纤电流互感器日益成为电力系统中最重要的设备,在实用化研究中,光纤电流互感器的检测精度是重要的指标之一,而被测电流的谐波和光纤电流互感器内部的噪声往往影响其测量的准确性。因此对光纤电流互感器谐波的检测以及对其解调算法实用性的验证非常重要。为进一步的谐波治理提供依据及验证解调算法能否有效运用问题,在详细分析现有的谐波检测理论和两种解调算法原理的基础上,设计了一种基于虚拟仪器LabView的光纤电流互感器谐波检测及解调算法测试一体化系统,在快速检测电流各次谐波电平的同时解调出被测电流的信息并对比验证两种解调算法的可行性与优越性,验证结果表明第二种解调算法的精度明显优越于第一种解调算法,更能适用于光纤电流互感器的信号解调。     

一种变压器油色谱峰识别算法的设计

在变压器油色谱峰识别领域,传统的一阶导数法需要斜率阈值来实现色谱峰识别,因此自动化程度低及容易失真。针对这些缺点,本文在一阶导数法的基础上进行了改进,将迭代移动平均及归一化分析技术应用到色谱峰识别中,通过对信号曲线及方波曲线多次迭代移动平均确定最优的归一化识峰参数,结合色谱峰的绝对保留时间及识峰窗口实现对色谱峰的准确识别。实验结果表明：该算法可以准确识别色谱峰,对噪声、色谱峰的峰宽及峰形变化不敏感,具有很强的自适应性,满足变压器油中气体在线监测装置的现场使用要求。     

参数寻优支持向量机在基于光声光谱法的变压器故障诊断中的应用

为了解决变压器气相色谱分析法故障诊断中存在的操作繁琐、消耗待测气体和载气、检测周期长等缺点,提出了利用光声光谱技术检测变压器油中CH₄,C₂H₂,C₂H₄,C₂H₆,H₂五种特征气体的含量并计算C₂H₂/C₂H₄,CH₄/H₂,C₂H₄/C₂H₆三对比值数据。将五种SVM类型和四种核函数采用交叉组合建立20种不同的支持向量机模型,并采用启发式算法对于惩罚因子c和g的取值进行参数寻优,以建立变压器故障诊断准确率最高、最快运行速度的支持向量机模型。启发式算法主要对比研究了粒子群算法和遗传算法在寻优精度与速度上的效果。仿真实验结果表明C-SVC模型、RBF核函数、遗传算法寻优构成的支持向量机模型对变压器故障的诊断准确率最高,测试集达到97.5%,训练集达到98.333 3%,并且遗传算法的寻优速度快于粒子群算法2倍左右。该方法具有操作简单、非接触性测量、不消耗载气、检测周期短、稳定性和灵敏度高等优点。可以代替传统的气相色谱分析法进行变压器故障诊断,满足变压器故障诊断的实际工程需要。     

电网末端电力变压器原边缺相故障分析

分析了电网末端电力变压器原边缺相故障特征,并提出了故障判据。