中文摘要：

为了解决发电机局部放电评估困难的问题,构建了一个局放评估框架,针对基于时间和相位的局放特征参数分别进行趋势和相似分析,在此基础上进行了综合评估。在局放趋势分析中,引入了金融时序分析方法,基于平均放电量和放电次数定义一特征参数强度次数IN,它能比常用局放特征参数更清晰地揭示局放发展过程中的异常;在局放相似分析中,针对发电机的局放指纹,定义了特征参数相似率。实例分析表明,IN和相似率能有效地反映局放的变化情况,为局放的正确评估提供支持。

英文摘要：

Based on the feature parameters extracted from the partial discharge （PD） data, a scheme for assessing the severity of partial discharge in generators was established. Feature parameters of PD could be classified into categories, or time-resolved PD parameters and phased-resolved PD parameters. In the contribution, the former was used to analyze the trend of PD and the latter to study the stability of discharge mode. Accordingly, for two kinds of PD parameters, trend analysis and similarity analysis were carried out respectively. Further, with the results obtained, final evaluation of PD was given out. The performance of commonly used statistical parameters is vague while displaying the trend of PD, and for improving the situation, by introducing the financial time sequence analyzing method, a novel parameter indicating the development trend of partial discharge, namely Intensity Number （IN）, was proposed. IN was derived from mean discharge intensity and discharge number. In comparison to other parameters of PD, IN could display and discover the abnormal case in the partial discharge developing process more clearly. Besides, for studying the stability of partial discharge type, especially in the case of generators, a new parameter similarity rate was defined. 26 phase-resolved parameters constituted the fingerprint of PD and the similarity rate was based on it. With the distribution of similarity rate, stability analysis of discharge was conducted. Through practical data analysis, it was demonstrated that, IN as well as similarity rate could reflect the development of PD effectively, and provide strong support for further PD assessing study.