基于机器学习决策树的计量设备异常分析

Abnormity analysis of measurement equipment based on decision tree

为加强对现场计量设备的实时监测与管理、提前发现设备隐患并及时处理,提出一种基于机器学习决策树的计量设备异常分析的诊断方法。该方法通过建立计量设备异常分析模型,分析计量设备发生故障的规律,对计量设备在运行过程中产生的大量数据进行典型特征提取,分析计量设备故障特征的相关度以及权重系数,从而分析计量设备发生故障的概率,同时结合运维人员的现场设备故障核查反馈的结果,不断优化完善计量设备分析模型,提高了计量设备异常发现的及时性和设备故障查处的命中率,有效降低了运维人员的工作难度。

In order to enhance the real-time monitoring and management of on-site metering devices,and identify potential hidden dangers in advance,a diagnosis method based on machine learning decision tree is proposed. This method establishes the anomaly analysis model of measuring equipment,measuring equipment failure analysis law,through a large amount of data generated by the measurement equipment in the process of operation of measuring equipment of typical feature extraction,calculation of fault characteristic measuring equipment correlation degree and weight coefficient,so as to analyze the probability of failure of metering equipment,combined with fault verification the equipment maintenance personnel feedback results,optimize the metering equipment analysis model,to improve the hit rate and abnormal metering equipment timely and equipment fault,effectively reduce the difficulty of operation and maintenance personnel.