【作者】 陈丽；

【导师】 栗然；

【作者基本信息】 华北电力大学（河北） ， 电力系统及其自动化， 2006， 硕士

【摘要】 随着电力体制改革的不断深入,降低设备维修成本,使设备更加经济可靠的运行,从而降低发电成本,是发电企业面临的一个重要课题。目前电力变压器等电力设备大多采用计划检修,这种体制存在严重缺陷,因此对变压器的故障进行预测从而实施状态检修势在必行。本文提出了将数据挖掘中的ID3(Interactive Dicremiser version3)算法和蒙特卡洛(Monte Carlo)方法结合用于变压器状态检修辅助决策的方法。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故障发生的规律,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。更多还原

【Abstract】 With the innovation of the electric power system, reducing the cost of equipmentsmaintenance to make equipments operate economically and reliably to reduce the cost ofgenerating electricity is an important problem for electric power plants to face. At the presenttime, Time-Based maintenance is often applied to transformers and other electrical powerequipments, but there are serious shortcomings in this maintenance strategy. So it isimperative under the situation to forecast transformer fault and carry out Condition-BasedMaintenance. In this paper the method which is combining ID3 Algorithm and Monte CarloSimulation is brought forward to help us to make assistant decision. Based on muchobservation, with ID3 Algorithm to simplify attributes of transformer, applying distributedfunction of Monte Carlo Simulation and rules of fault, and through simulating of randomvariable to fault of transformer, we can get probabilities of transformer fault to direct faultdiagnosis and state prediction, which can provide evidence for Condition-Based Maintenance.更多还原