摘要
为了解决电力变压器套管绝缘状态评估中存在不确定性因素的问题,提出了一种新的状态评估方法。该方法全面考虑评估指标等级分类边界的随机性和模糊性,采用由云推理确定指标客观权重,由未确知理论确定主观权重的组合赋权法。建立了核向量空间模型,利用核函数将样本映射到高维特征空间的方法,在高维特征空间定义样本数据的指标有向线段和变压器套管绝缘等级标准的理想指标有向线段。进而通过计算两线段之间的夹角加权余弦获得样本与标准模式的贴近度,将绝缘评估问题转化为向量空间的模式识别问题,最终得出变压器套管所处的绝缘状态。实际算例表明该套管在第1次测试时,其绝缘状态贴近正常的贴近度最大即处于正常状态;第2次和第3次测试时,贴近故障的贴近度最大即处于故障状态。
In order to reduce uncertain factors affecting insulation condition assessment of power transformer bushing,we proposed a novel condition assessment method.The randomness and fuzziness of evaluation index classification boundary are fully considered in the method.This combined weighing method includes the objective weight and the subjective weight decided by cloud model and unascertained theory,respectively.A kernel vector space model is established,and the kernel function is used to map the input data into the high-dimensional feature space,in which the index-directed line segment of data sample and ideal directed line segment of insulation grade standard sample are defined.By calculating the weighted cosine value between directed line segment in each data sample and ideal directed line segment in grade standard sample,the approaching degree is obtained.By using the approaching degree,the insulation condition assessment is transformed into the pattern recognition in vector space,from which the insulation condition grade is obtained.An actual example shows that,in the first test of a bushing,its insulation condition is close to the normal state in the largest degree,meaning that it is in normal state;yet in the second and third tests,it is close to the fault state in the largest degree,meaning that it is in fault state.
出处
《高电压技术》
EI
CAS
CSCD
北大核心
2013年第5期1101-1106,共6页
High Voltage Engineering
基金
国家重点基础研究发展计划(973计划)(2009CB724505)
国家创新研究群体基金(51021005)~~
关键词
电力变压器套管
绝缘状态评估
云模型
组合赋权
核向量空间模型
贴近度
power transformer bushing
insulation condition assessment
cloud model
combination weighing
kernel vector space model
approach degree
