摘要
提出了将决策树方法中的ID3(InteractiveDicremiserversion3)算法和蒙特卡洛(MonteCarlo)模拟法结合用于变压器故障预测的方法。基于变压器的大量观测数据,首先利用ID3算法简化引起变压器故障的属性特征,再应用蒙特卡洛方法的分布函数和ID3算法得出的故障发生规则,通过随机变量对变压器故障进行模拟,从而得到设备故障的概率,以此来指导变压器的故障诊断或状态预测,为状态检修提供依据。并对变压器油色谱监测中得到的各气体含量数据进行分析和处理,介绍了方法的运用并对其可行性进行了验证。
The method combining ID3 （Interactive Dicremiser version3） Algorithm and Monte Carlo Simulation is applied to forecast the transformer faults. Based on the observed and measured data, the ID3 Algorithm is used to simplify the attributes of transformer. The rules of fault generation are obtained by distributed function of Monte Carlo Simulation and ID3, The random variables are used to simulate the fault of transformer. The probabiiities of fault are Obtained. The data of dissolved gases ： in transformer oil is analyzed and processed.
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2006年第2期12-15,共4页
Journal of North China Electric Power University：Natural Science Edition
关键词
状态检修
故障预测
ID3算法
蒙特卡洛模拟法
变压器
condition-based maintenance
fault prediction
ID3 Algorithm
Monte Carlo Simulation
transformer
