Optimization of support insulators used in HV systems using support vector machine

In this paper support vector machine (SVM) has been used for optimization of electric field along the support insulators used in high voltage (HV) systems. To illustrate the effectiveness of SVM in optimizing electric field problems, two axi-symmetric cases have been considered one having an insulator with a contour that is quarter ellipse and the second one having a porcelain core solid insulator. The training and the test data for the SVM have been prepared by electrostatic field computation using indirect boundary element method (BEM). It is observed that once the SVM is trained it is able to give results with mean absolute error of less than 1.5% when compared with the analytically obtained results. The SVM designed for insulator contour optimization is first trained with the results obtained from electric field computation for some predetermined contours of the arrangements under consideration. Then the trained SVM is used to produce an optimized insulator contour in such a way that the desired stress distribution can be obtained on the insulator surface. The results obtained from the present study show that the trained SVM is adequately efficient to optimize insulator contours in order to have the desired stress distribution along the insulator surface     

Insulator contour optimization by a neural network

Contour optimization of axi-symmetric insulators in multi-dielectric arrangements by a neural network (NN) has been carried out to obtain not only a uniform but also a complex electric stress distribution along the insulator surface. Three case studies have been reported In detail. Multilayer feed-forward networks with error backpropagation and also with resilient propagation learning algorithms have been employed. The training and test data have been prepared by means of electric field calculations. The results show that optimized insulator contours have been obtained with an acceptable degree of accuracy with the help of a trained NN     

Electrode-spacer contour optimization by ANN aided genetic algorithm

Optimization of contours of three dimensional electrode-spacer arrangements used in gas insulated systems (GIS) has been carried out by ANN aided genetic algorithm (GA) to obtain desirable electric stress distribution along the insulator surface. Two cases have been studied and reported in detail. Multilayer feed-forward neural networks with error back-propagation algorithm have been applied to accelerate the execution of GA loop. The training and the test data have been prepared by means of electric field calculations using indirect boundary element method (BEM). The results show that optimized contours have been obtained with acceptable degree of accuracy with the help of GA aided by trained ANN.     

Insulation optimization by electrode contour modification based onbreakdown area/volume effects

This paper shows how to determine the electrode contour with the best insulation performance on the basis of an area effect and a volume effect in the breakdown field strength. Previous electric field optimization techniques have provided us only with an optimum electric field distribution, For the design of power apparatus insulation, we have to consider the dielectric breakdown characteristics rather than the electric field distribution. As a first step, we developed a new optimization technique which enables us to obtain an optimum electrode contour with the highest breakdown strength while taking into account the area and volume effects of breakdown strength of insulating media. From the results, we have confirmed that the proposed optimization method improved the breakdown voltage more than did the electric field optimization. This leads to an effective insulation design of electric power apparatus     

Corona onset as a design optimization criterion for high voltagehardware

Hardware for use on HV systems is designed to be corona free. Recent hardware design activity has been centered around attempts to increase the corona onset voltage by optimizing the electric field distribution on its surface. Corona onset, however, is not only a function of the surface electric field, but also its rate of decay away from the surface, and the temperature and pressure of the gas in which it is immersed. While corona onset is a better criterion, formulas for predicting it have been validated only for simple geometries. The first goal of this work, then, is to validate corona onset conditions for more practical electrode geometries. The second goal is to determine whether the use of corona onset rather than surface electric field as an optimization criterion can result in hardware with a measurably higher corona onset voltage. To test this idea, two electrodes were designed, one using electric field optimization and the second using corona onset optimization. The corona onset voltage of each electrode was then measured and the results compared with predicted values. It was found that the measured results compared favorably with the predicted values and that the use of corona onset optimization can result in a modest but measurable increase in corona onset voltage     

Optimal Design of Support Insulators Using Hashing Integrated Genetic Algorithm and Optimized Charge Simulation Method

This paper proposes a combined approach of optimized charge simulation method (CSM) and hashing integrated adaptive genetic algorithm (HAGA) to the contour design of support insulators. In the combined approach, the HAGA determines the optimal displacement ratio and number of fictitious charges used in CSM, and then obtains the optimal contour design to minimize and uniform the tangential electric field along the dielectric boundary based on the optimized CSM. The proposed HAGA method considerably improves the execution efficiency by accessing the fitness values of the searched solutions saved a priori in a hashing table during the optimization process. To verify the effectiveness of the proposed method, three cases of different contours of support insulators have been studied. Results show that rather uniform and minimal tangential field distributions along the insulator surfaces can be effectively derived by the proposed approaches to achieve promising performance and higher probability free from flashover.     

基于改进遗传算法的SF6断路器匀场设计

提出一种适于求解复杂结构优化问题的改进遗传算法(variable interval genetic algorithm,VIGA)。采用实数编码方式以及混合杂交算子以提高实数编码杂交效率,同时,通过调整子代种群生成方式,提高算法收敛性,不至于过早收敛到局部极值。以2个典型测试函数为例,分别进行了多峰值、多变量测试函数寻优测试,通过将VIGA寻优结果与测试函数解析解、传统遗传算法(genetic algorithm,GA)以及模拟退火优化(simulated annealing algorithm,SA)策略所得优化结果进行对比分析,VIGA的可行性和有效性得到证实。以550 kV单断口SF6断路器灭弧室为研究对象,进行触头结构优化设计,得到灭弧室内部各触头及屏蔽罩沿面电场分布。仿真结果表明,基于VIGA的电场逆问题求解可有效改善超高压断路器灭弧室电场分布,实现结构优化设计以及灭弧室内匀场设计。     

Size effect and statistical characteristics of dc and pulsedbreakdown of liquid helium

Breakdown time lag in liquid helium is measured over a wide range of electrode sizes and pulsed electric field strengths. The breakdown time lag and dc breakdown strength are statistically analyzed by using the Weibull distribution function and weak link theory. It is found that the time lag depends on both electrical stress and the electrode surface area stressed above a critical level. It is supposed that breakdown triggering electrons are generated by field emission phenomena at small protrusion tips on the cathode surface. In higher external electric fields, a less sharper protrusion emits initial electrons with a shorter time lag and may become responsible for liquid breakdown. A theoretical equation is proposed to predict the electrode size and electrical stress dependency of the breakdown time lag, based on Fowler and Nordheim theory. It is shown that the equation is consistent with the Weibull distribution function under multiple stress of electric field and stressing time     

双平行电极场畸变气体火花开关电极优化与实验研究

为适应新型低电感电容,笔者优化设计了应用于电缆绕组型直线脉冲变压器的新型双平行电极场畸变气体火花开关。对触发电极处于主放电通道和未处于主放电通道的两种电极结构进行了仿真模拟,通过模拟发现不同触发电极位置对开关电场分布无明显影响。分别对触发电极不同位置的场畸变火花开关进行实验,开展了不同电极结构的静态耐压、触发延迟实验。实验数据表明,触发电极处于主放电通道上时,开关的静态特性和触发特性较好。     

真空中典型沿面绝缘结构的电场分析

在高电压作用下,由复合绝缘介质构成的沿面绝缘结构的耐电强度远低于其绝缘材料自身的击穿场强,这一现象与其电场的分布特点密切相关。笔者针对真空中平行平板、平面和棒-板电极系统等多种典型沿面绝缘结构的电场分布进行了仿真计算,探讨了电极-介质结合处的间隙、圆台形绝缘子的圆锥角角度、平面电极的高度以及绝缘子介电常数等因素对电场分布的影响。仿真结果表明,接触间隙的存在导致局部电场的加强和电场方向的变化,间隙宽度越大、高度越小,间隙处电场畸变越大;圆锥角越大,绝缘子的介电常数越大,场强畸变也越大。该分析结果有利于真空中沿面绝缘结构的设计。     

静电放电电磁脉冲耦合的非线性优化建模

针对现有机理建模算法普遍存在计算电磁脉冲响应过程过于复杂的问题,为能够给电子设备静电放电电磁脉冲响应计算提供一种简便有效的能量耦合建模方法,设计了脉冲场强测试仪的静电放电辐射实验。用NARX神经网络代替传统NARX网络,依靠遗传算法对网络的初始权值、阈值进行优化,以3.5 kV静电放电实验数据作为建模数据对系统进行非线性辨识,并对4.5 kV静电放电电磁脉冲响应进行预测。建模结果表明,两种模型均能准确预测响应波形,但优化后的NARX神经网络模型精度更高。该建模方法计算过程简单。该方法同样适用于其他电磁脉冲响应建模。     

GIS绝缘子在冲击电压下沿面闪络的分析

利用计算和试验两种方法,对在SF_6气体中的绝缘子施加冲击电压下的电场及其引起的沿面闪络原因进行了剖析,指出在冲击电压作用下,增加屏蔽罩是减弱电极结构所引起的电场垂直分量的一条行之有效的途径;沿面闪络主要取决于电场分布的均匀度和绝缘子表面受周围环境导电微粒污染这两个基本因素;电极形状导致的局部场强增大,靠增加绝缘子沿面闪络距离不能明显提高沿面闪络电压。     

高压SF6断路器灭弧室电场优化与分析

采用有限元软件ANSYS对SF6断路器灭弧室内触头形状进行了优化,进而得到最优的电场分布。主要分为两部分,第一部分得到了有限元法计算电势的弱解方程,并利用有限元软件ANSYS精确建立了断路器灭弧室的CAD模型,具体介绍了断路器灭弧室电场优化的方法;第二部分根据计算结果分析了断路器灭弧室内的电场分布,并详细分析了触头形状的改变对灭弧室内电场的影响。得到的电场优化计算结果为高压断路器灭弧室的优化设计提供了参考数据。     

基于电-热场联合分析的EPR中压电缆终端异常热点仿真分析及优化

25 kV乙丙橡胶(EPR)中压电缆终端因其自身的结构特性,内部电-热场分布不均,局部易出现异常畸变热点问题,而在安装电缆终端时出现的划伤缺陷加大了问题的严重程度,加速缺陷周围绝缘材料的老化,大幅降低了绝缘性能。为解决该问题,提出了一种电导率与电场、温度相关的非线性应力管材料,采用COMSOL仿真方法对比研究了使用高介质材料与非线性材料制作应力管时电缆终端内部的电-热场分布。结果表明,经优化后电缆终端的电-热场分布畸变程度能得到有效缓解;对于存在划伤缺陷的情况,优化后的电缆终端的电-热场畸变程度低于其出现击穿现象的阈值,表明其能够在缺陷情况下相对安全运行。同时采用热成像仪现场测试电缆终端温度分布,结果验证了经优化后电缆终端表面异常发热情况的改善效果。     

直流气体绝缘输电线路的绝缘设计

气体绝缘输电线路(gas insulated transmission line,GIL)与架空线路相似但占地空间小、损耗低,在高压直流输电和特高压直流输电领域具有较大的应用空间。通过分析表面电荷和金属导电微粒对绝缘子沿面放电的影响,指出了绝缘子表面电荷积聚和自由金属导电微粒附着是降低直流GIL绝缘性能的重要原因。采用了使电场分布合理的方法,即半圆锥形盆式绝缘子的优化和表面电阻率阶梯分布的覆膜。设计了包括电极覆膜、微粒陷阱、驱赶电极和屏蔽环的直流GIL的绝缘结构。     

Surface field dynamics in dc film capacitors under an impulsevoltage perturbation

For metalized polymer film capacitors, the dynamics of electric field on the electrode surface and its resulting energy dissipation influence the onset and scale of breakdown discharges within the capacitor unit, which in turn is indicative of the reliability and lifetime. To gain an understanding of the dependence of surface electric field on key system parameters, an equivalent circuit model is developed to simulate the temporal evolution of the, electric field distribution within a dc film capacitor subjected to an external impulse electric stress. A mosaic pattern of electrode segmentation is taken into account by means of an effective surface resistance and its effects on surface electric field and energy dissipation are computed numerically. Also considered are the effects of the number of electrode segments across a given length. In addition, electric energy dissipation is calculated to assess the level of possible temperature rise within the capacitor. By addressing these two issues, it is shown that the numerical code developed and its underlying methodology can serve as a complimentary tool to the present practice of capacitor designs and performance assessment     

Boundary element studies on insulator shape and electric fieldaround HV insulators with or without pollution

Capacitive-resistive field computations are carried out around post-type HV insulators of varying shapes. The boundary element method (BEM) has been employed for electric field computations. Different insulator shapes have been obtained by varying several parameters, which define the shape of the HV insulator contour. For each insulator shape, the maximum stress occurring on the insulator surface has been determined with no surface pollution, uniform surface pollution and also partial surface pollution. For partial pollution, several cases have been studied, in which different sections of the insulator surface are polluted. Furthermore, the effect of electrode radius on the maximum stress on insulator surface has been investigated. The results obtained are presented in this paper in detail     

A self-organizing fuzzy inference system for electric fieldoptimization of HV electrode systems

Development of efficient methods for optimum design of electrode and insulator contour in a HV apparatus has been an important area of research in the recent past. A self-organizing fuzzy inference system is developed for electrode optimization which is an improved version over traditional iterative approach of optimization and more recent methods employing neural networks. This fuzzy system has been implemented for two practical examples of an axisymmetric single-phase gas-insulated system (GIS) bus termination and a disconnecter switch configuration. Improved electric stress distribution demonstrates the usefulness of the proposed method. A new methodology for automatic generation of the fuzzy rule base from the input-output data set is also proposed in this scheme. This fuzzy system is much simpler and less time consuming to implement compared to other optimization techniques presently available for optimization of HV electrodes     

特高压复合绝缘子电场计算及基于神经网络遗传算法的均压环结构优化设计

为了满足超/特高压大容量输电技术的发展需求,研究复合绝缘子的电场分布,寻找合理的电场优化措施已成为影响输电线路安全运行的关键因素。利用有限元法(FEM)建立了工程实际应用情况下特高压(UHV)交流输电线路复合绝缘子的单串、双串和V串电场计算模型,在考虑横担、分裂导线和均压环及其金具连接情况下,对复合绝缘子的电位和沿面电场分布进行了计算。针对特高压复合绝缘子电位分布极不均匀,且最大沿面电场强度大大超过电晕起始场强的情况,分析了不同的均压环配合方式对复合绝缘子沿面电场分布的影响。针对单串绝缘子下的计算结果,引入神经网络遗传算法(BP-GA),对均压环的结构参数进行了优化设计,利用BP神经网络的高度非线性映射能力直接计算遗传算法的适应度函数值,解决了穷举法计算时间冗长的问题。计算结果表明:加装均压环能够大大降低最大沿面电场强度,明显改善复合绝缘子沿面电位和电场的分布;通过遗传算法优化后的均压环结构参数,能使均压环表面场强和绝缘子沿面场强均小于电晕起始场强,为特高压绝缘子均压环的选取提供了一种可靠、实用的设计方法,可为我国特高压输电线路外绝缘优化设计提供新的参考。     

Electric field optimization of the power cable joint by using evolutionary calculation method

This paper examines electric field optimizations at the joint of power cables by using evolution strategy [(1+1)‐ES, (µ+λ)‐ES] and genetic algorithm which are so‐ called evolutionary calculation methods. The object of the paper is to minimize the electric field strength on the inner curved electrode under the constraint that the electric field strength along the interface between the two different insulators should be below the permissible value. The surface charge method that provides accurate solutions on the boundaries for the multilayer substances with a relatively small number of elements is employed for the calculation of the electric field distribution. Three evolutionary calculation methods are employed for the optimization method. Because these methods employ mutation and rearrangement (cross‐over) processes, the shape design can be performed widely and locally. As a result, each method can reduce the electric field strength on the inner curved electrode while the electric field strength along the interface was kept below the permissible value. The search results of each optimization method were compared and (µ+λ)‐ES was found to have the best performance of the three methods. © 2005 Wiley Periodicals, Inc. Electr Eng Jpn, 150(4): 44–53, 2005; Published online in Wiley InterScience ( www.interscience.wiley.com ). DOI 10.1002/eej.10287