摘要
针对电网发生故障时存在误报和漏报等不确定情况,提出一种基于专家规则和信息论的电网警报处理方法。该方法对警报处理过程中的警报划分集合关系,通过基于专家规则的正反向推理得到候选故障原因的特征警报集,并引入信息熵分析显示的警报集和特征警报集的相关联程度,科学地量化分析处理过程中的不确定程度。提取的交互信息量和信息比两个特征指标,在考虑时序特性与不考虑时序特性两种情况下,均能实时准确评估候选故障原因。实际算例验证了该方法的实用性和有效性。
On the bases of the rules of experts and information theory, an approach to process the power system alarm is presented, which can tackle the uncertainty in the process of alarm processing, e.g. misleading alarms, missing alarms, etc. This method initially clarifies the aggregation relationship and establishes the characterized alarms of can- didate faulty equipments by means of forward and backward reasoning based on rules of experts. Then the Correlation of the given alarms and characterized alarms is analyzed in order to quantify the uncertainty. Two metrics (i.e. the amount of interactive information and the information ratio) are extracted to obtain the real fault cause, and the metrics could accurately evaluate the amount of information of the candidate fault causes in real time (with and without considering the temporal constraint). Results of practical numerical example indicate that the proposed method is available and effective.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2014年第12期69-73,90,共6页
Proceedings of the CSU-EPSA
关键词
电网警报
专家规则
信息熵
时序特性
评估指标
power system alarm
rules of expert
information entropy
temporal constraint
evaluation metrics