基于缺陷概率预测的主变巡视策略优化

Main transformer inspection strategy optimization based on prediction of defect probability

设备巡视是预控设备运行风险,夯实电网安全运行基础的重要手段。通过以某省级电网公司35 k V及以上油浸式主变的历史数据为基础,构建了基于Logistc回归方法的主变个体缺陷概率预测模型,并提出了基于主变个体缺陷概率预测的设备巡视重点制定方法,通过实证数据验证了基于主变个体缺陷概率预测开展设备巡视的有效性。

Device inspection is a significant means to control the risk and maintain the safe running of power grid. A defect probability prediction model based on logistic regression method is established on the basis of historical data of 35 kV and above oil-immersed transformers in a provincial power grid company. A method to generate the inspection strategy based on the model predicting the defect probability of individual transformer is proposed. The availability of the device inspection based on defect probability prediction for individual transformer is proved with real evidence data and empirical analysis.