基于数据挖掘理论的电力系统暂态稳定评估

将数据挖掘理论中的关联规则分析与分类分析相结合，提出了一种基于数据挖掘理论的暂态稳定评估方法。文中选择反映电力系统运行状态的特征变量，建立暂态稳定评估模型；考虑到电力系统数据量大的特点，采用聚类分析、特征变量提取、连续数据离散化等数据预处理手段，以提高问题判断的准确性、可靠性和实用性。利用关联分类法可产生反映电力系统运行状态和暂态稳定性的关联规则，这些规则可被用来对系统进行暂态稳定的预测和评估。通过对3机9节点系统的计算，验证了该评估方法的有效性。

A NOVEL APPROACH FOR TRANSIENT STABILITY ASSESSMENT BASED ON DATA MINING THEORY

On the basis of the data mining theory,a novel approach to assess the transient stability of power system, i.e. associative classification method, is presented. The feature variables describing the system states are selected for transient stability assessment. A mass of initial data of power system need to be preprocessed by clustering analysis, feature variables extraction, and data discretization, etc, to improve the veracity, reliability and utility of the assessment. By using associative classification, some associative rules reflecting the relationship between the operation status and transient stability of power system can be generated, which can be used to forecast and assess the transient stability. As an example, the 3-machine 9-node power system is used for simulation. The result shows the validity of the proposed approach.