基于改进的连续局部枚举采样和径向基函数响应面法的变压器静电环结构优化设计  

Optimized Design of Transformer Electrostatic Ring Based on Radial Basis Function Response Surface Method with Enhanced Successful Local Enumeration Sampling

作　　者：刘刚[1] 高成龙 胡万君 朱章宸 刘云鹏[1] Liu Gang;Gao Chenglong;Hu Wanjun;Zhu Zhangchen;Liu Yunpeng(Hebei Provincial Key Laboratory of Power Transmission Equipment Security Defense North China Electric Power University ,Baoding 071003 China)

机构地区：[1]华北电力大学河北省输变电设备安全防御重点实验室,保定071003

出　　处：《电工技术学报》2023年第23期6266-6278,共13页Transactions of China Electrotechnical Society

基　　金：国家重点研发计划(2021YFB2401703);中央高校基本科研业务费专项资金(2022MS073)资助项目。

摘　　要：变压器的安全稳定运行与其内部绝缘结构密切相关,为了提高变压器的绝缘性能,该文以提高最小绝缘裕度为优化目标,通过径向基函数响应面模型优化变压器主绝缘结构。为提高响应面模型预测优化的准确性,选用采样更均匀改进的连续局部枚举(ELSE)方法进行试验设计,优化后的变压器最小绝缘裕度比优化前提高了12.56%。将优化结果分别从采样、模型和优化方法三个角度进行对比,与拉丁超立方采样(LHS)采样下的径向基函数响应面模型对比,ESLE方法的引入使得优化精度提高到218倍;与ESLE采样下的二次多项式响应面模型对比,径向基函数响应面模型的引入降低了模型预测误差,并使优化精度提高到78倍;与遗传算法对比,在相同的优化精度下,优化效率是遗传算法的10倍。上述结果表明,该优化模型具有很高的预测精度和优化能力,同时相比于遗传算法优化,在保证优化质量的同时,显著地提高了优化效率。该文所提出的方法为变压器主绝缘结构优化问题提供了较好的解决方案。The transformer main insulation structure needs be reasonably designed to ensure its safe and reliable operation.The response surface method(RSM)is an efficient tool to the optimization problem of transformer main insulation structure.In the field of traditional transformer structure optimization,the RSM is widely used due to its superior optimization accuracy and efficiency in comparison to the empirical formula method and the global optimization algorithm.This paper proposes combining the(ELSE)enhanced successful local enumeration(ELSE)method with upgraded sampling uniformity and the radial basis function(RBF)-RSM with superior nonlinear fitting capability to solve the main insulation structure optimization problem,whose optimization objective is to improve the minimum insulation margin,so as to obtain a more effective design scheme for the transformer main insulation structure.Firstly,a 500 kV transformer is modeled using the ansys parametric design language(APDL).Secondly,it is necessary to identify optimization variables,which are selected from the electrostatic ring structure.Subsequently,the training set and test set data are obtained using the ESLE sampling method.Then,the RBF-RSM is derived using the training set data.Finally,the resulting RSM is optimized by an intelligent algorithm,which can figure out the variables’values with the optimal minimum insulation margin.Due to the parametric modeling,the training set and test set objective function results are obtained via Matlab and Ansys calls,which significantly reduces the human workload and enhances the engineering applicability of the above process.The results of the optimization based on ESLE and RBF-RSM are compared with the pre-optimized results,and the minimum insulation margin is improved by 12.56%,which indicates that the proposed method has practical utility.In order to verify the advantages of the proposed approach,firstly,from the sampling method,the ESLE method is compared to the LHS sampling method in RBF-RSM optimization,and the usage of

关 键 词：响应面方法 改进的连续局部枚举 径向基函数 绝缘裕度 变压器 结构优化 

分 类 号：TM403.3[电气工程—电器]