A data cleaning method and system for a GIS partial discharge big data system

Abstract

The present invention provides a data cleaning method and system of a GIS partial discharge big data system. First, establish a GIS partial discharge big data system, collect data, formulate a unified data format, develop a conversion module, and normalize the format of heterogeneous data sources Then, through exploratory analysis, process abnormal data, establish corresponding cleaning rules for different data types, and form a data cleaning rule base, organize data into uniform and accurate data sets, and continuously optimize data Clean the rule base; finally, mine and analyze the data. By adopting the invention, the format conversion and data cleaning of the partial discharge detection data can be carried out, and the advanced algorithm can be used for analysis and mining, so as to improve the diagnosis accuracy of the partial discharge detection data.

Description

technical field [0001] The invention relates to the technical field of partial discharge big data processing of gas insulated switchgear (GIS), in particular to a data cleaning method and system of a GIS partial discharge big data system. Background technique [0002] With the rapid development of power grid scale and the continuous increase of power load, the reliability and safety requirements of electrical equipment operation are also increasing, and it becomes more meaningful to perform status detection on the insulation performance of electrical equipment. Effective and accurate detection and evaluation of electrical equipment is the premise of condition-based maintenance and life-cycle management, and the basis for ensuring safe and reliable operation of electrical equipment. [0003] After the power equipment is put into operation, corona and surface partial discharge will occur due to design defects, surface pollution and poor contact. When discharge occurs, physical...

AI Technical Summary

Problems solved by technology

[0005] 1. For missing values, outliers, and duplicate values ​​in the data, a simple delete operation is used to lose a lot of data that can be used through data processing;

[0006] 2. The data format is not unified, and the data formats generated by various testing instruments are not uniform. It is necessary to develop a set of data mining methods for each type of data, which is inefficient and lacks comprehensive application of data;

[0007] 3. Using manual classification of data is prone to errors and low efficiency, and there is no combination of classification algorithms to classify data

Images