Construction method and application of electric energy metering abnormity diagnosis model

Abstract

The invention discloses a construction method and application of an electric energy metering abnormity diagnosis model, training data are enhanced by constructing different sample pairs, and a two-stage model training method is adopted. The method comprises the following steps of: firstly, pre-training an electric energy metering anomaly diagnosis model by taking a difference between output of a minimization model and a corresponding sample label as a target, and determining an interval threshold value based on output distribution of the model in a pre-training stage; secondly, with the purpose of maximizing the interval between the model output corresponding to the positive example pair and the negative example pair and the interval threshold value, performing multi-round training on the electric energy metering abnormity diagnosis model, and updating the interval threshold value based on the output distribution of the model in the previous training stage in the multi-round training process so as to be used for the next training stage; the method greatly improves the accuracy of the model, and enables the model to have the capability of sensing unknown abnormal power data on the basis of considering the recognition performance of known categories under the conditions of insufficient power data and non-uniform sample distribution.

Description

technical field [0001] The invention belongs to the field of abnormal diagnosis of electric energy measurement, and more specifically relates to a construction method and application of an abnormal electric energy measurement diagnosis model. Background technique [0002] Power metering devices are an important source of data for analyzing the operating status of power equipment. On the one hand, the abnormality of electric energy metering data reflects the failure of power equipment or the abnormality of line. In view of the wide distribution of electric energy metering equipment, it is difficult for maintenance personnel to inspect each device in a short time. Therefore, automatic remote monitoring of electric power equipment can be realized through abnormal diagnosis of metering data. Once the measurement data is abnormal, the abnormal equipment or line can be quickly located to avoid safety accidents caused by equipment failure. On the other hand, the abnormality of el...

AI Technical Summary

Problems solved by technology

[0008] In view of the above defects or improvement needs of the prior art, the present invention provides a construction method and application of an abnormal diagnosis model of electric energy metering, the purpose of which is to solve the problem that the prior art cannot use The technical problem that the electric energy metering anomaly diagnosis model has the ability to perceive unknown abnormal power data on the basis of taking into account the recognition performance of known categories

Images