基于Stacking模型融合的失压故障识别算法

计量故障中的失压故障是目前电力计量系统常见的故障问题之一，传统的失压故障判定以终端告警为依据，判定维度单一，且终端告警存在误报、漏报的情况，导致了故障无法及时发现、无法实时处理。为了解决失压故障识别维度单一和终端漏报误报的问题，采用比较研究法，在前人使用机器学习算法解决故障识别问题的基础上，结合真实计量数据，构建失压关键指标，提出了一种基于Stacking模型融合的计量故障监测算法。经反复实例论证和理论测算，该算法相较于传统的机器学习算法，能够提升失压故障识别的效果，平均精确率0.99以上。该种算法的提出为计量故障识别提供了一种新的解决方案，为失压故障后电量追补提供了一种依据，为提升计量系统管理水平增加了一种手段。

Loss-of-voltage fault identification algorithm based on Stacking model fusion

The fault-loss fault in the measurement fault is one of the common fault problems in the current power metering system.The traditional pressure-loss fault determination is based on1 the terminal alarm，the judgment dimension is single，and the terminal alarm has a false alarm and a false report，resulting in a fault.Unable to find in time，can’t be processed in real time.In order to solve the problem of single-dimensional and terminal false-reporting of loss-of-voltage fault identification，the comparative research method is used.Based on the previous problems of using machine learning algorithms to solve fault identification problems，combined with real measurement data，the key indicators of pressure loss are constructed.A measurement fault monitoring algorithm based on stacking model fusion.Through repeated case demonstration and theoretical calculation，the algorithm can improve the effect of faultless fault recognition compared with the traditional machine learning algorithm，and the average accuracy rate is above 0.99.The proposed algorithm provides a new solution for metering fault identification，which provides a basis for power chasing after voltage loss faults，and adds a means to improve the management level of metering systems.