基于数据挖掘与改进灰靶的电压暂降严重度分析与评估

针对电压暂降评估问题,提出一种基于互信息与改进灰靶理论的电压暂降严重度数据挖掘分析方法。提出暂降距离和交叉影响域指标,考虑电压暂降的多种特征属性建立严重度评估框架;构建互信息网络模型对电压暂降进行数据挖掘,利用连接权提取关联规则构成电压暂降知识库,对全网电压暂降严重度水平进行初步分析,避免了基于支持度和置信度的传统数据挖掘方法产生的冲突规则问题;提出靶心度的改进计算方法,结合数据挖掘所得的全网暂降严重度水平,建立电压暂降严重度条件搜索模型进行变电站暂降严重度评估,解决了匹配结果多重化的问题,提高了电压暂降评估的准确性。算例分析验证了所提方法的可行性。

Analysis and Evaluation of Voltage Sag Severity Based on Data Mining and Improved Grey Target Theory

Aiming at the problem of voltage sag severity evaluation, a data mining and analysis method based on mutual information and improved grey target theory is proposed. The index of sag distance and cross influence domain is proposed, and the framework of severity evaluation is established considering various characteristic attributes of voltage sag. A mutual information network model is constructed for data mining of voltage sag, and a voltage sag knowledge base is built by using connection weight to extract rules to preliminarily analyze the severity level of voltage sag in the whole grid, which avoids the problem of conflict rules caused by traditional support and confidence based data mining methods. An improved method for calculating the target centricity is proposed. Combining with the sag severity level of the whole grid obtained from data mining, a conditional search model of voltage sag severity is established to evaluate the sag severity of substation, which solves the problem of multiple matching results and improves the accuracy of voltage sag evaluation. The feasibility of the proposed method is verified by an example analysis.