多指标加权分析的变压器故障油色谱诊断方法可靠性评估

Reliability Assessment of the DGA Fault Diagnosis Methodologies for Transformers Based on Multi-indexes Weighted Analysis

为了弥补油色谱诊断中仅凭正确率评价无法反映诊断误判的不足,文中引入故障诊断灵敏度的概念,通过比较各诊断方法的正确率和灵敏度两个概率指标,并进行加权分析,给出了评估变压器故障油色谱诊断方法可靠性的新原则。考虑到各故障类型案例发生概率的差异性,通过引入各类故障类型的可靠性权重的方法,提出了诊断方法的综合可靠性评估指标,使得对基于油色谱分析的变压器故障诊断方法可靠性的评估更加全面、准确。针对全国范围的628例故障案例,采用国内外7种典型油色谱诊断方法进行分析,通过对每种故障类型下各诊断方法正确率和灵敏度的比较以及综合可靠性比较,明确了各方法的优缺点,可为实际应用中选择合适的油色谱故障诊断方法提供依据,同时也为油色谱诊断方法的改进奠定了基础。

In order to addresses the shortage of misdiagnosis with only the accuracy rate index, the concept of fault diagnosis sensitivity rate is introduced, together with the prevailing accuracy rate, a new criterion to evaluate the reliability of the gas-in-oil fault diagnosis methodologies for power transformers is formed by weighting. Considering the differences of occurrence probability among the fault types, a composite reliability evaluation index is proposed by introducing the reliability weightiness of each fault, which makes the assessment of the gas-in-oil diagnosis effectiveness more comprehensive and accurate. Seven typical gas-in-oil diagnostic methods are used to diagnose the 628 fault cases from power transformer operating nationwide in terms of the accuracy and sensitivity rates for each type of faults, further, both the advantages and disadvantages of each method as well as a comprehensive reliability are obtained. The new outcomes present the basis for appropriate selection of the most suitable gas-in-oil fault diagnosis method in practical applications, and lay a foundation for amelioration of the prevailing gas-in-oil diagnosis schemes.