基于熵权层次法的油浸式变压器绝缘状态评价

为有效评估油浸式变压器的绝缘状态，文中提出一种综合主观专家经验和指标退化程度的熵权层次法。首先，基于油色谱、环境温度和缺陷信息等指标构建用于绝缘状态评价的层次结构，根据专家经验对各个指标进行主观权重分配；接着，采用组合权重的方法，在指标层权重分配上引入熵权权重，从而在保留专家经验对准则层判断的情况下，对底层权重进行动态调整，体现变压器运行过程中的差异性；然后，将变压器的绝缘状态从常见的4种扩展为5种，实现对绝缘状态更准确的感知；最后，4台500 kV变压器的实例分析表明，相较于仅采用主观权重或客观权重，文中提出的方法可以有效识别运行状态正常的变压器并甄别异常变压器。

Insulation condition evaluation of oil-immersed transformer based on entropy weight hierarchy method

In order to evaluate the insulation condition of oil-immersed transformers effectively,a weight allocation method considering expert experience and degradation state is proposed in this paper. Firstly,a hierarchical structure for insulation condition evaluation is built based on dissolved gas in oil,ambient temperature and defect information,and subjective weights is assigned to each index based on expert experience. Then,the method of combined weight is adopted to introduce entropy weight into the weight allocation of index layer,thus weights of the index layer are dynamically adjusted to reflect the differences in the individual operation process while retaining expert experience for the criterion layer. Furthermore,the transformer insulation state is subdivided into five states from common four to realize more accurate perception of insulation state. Finally,case study of four 500 kV transformers shows that the method proposed can effectively identify normal transformers and discriminate abnormal transformers compared with the method based only on objective or subjective weight,which verifies the rationality of the method.