A classification method for electric power on-line data collection

Abstract

The invention provides a classification method for electric power online data collection, the method includes the following steps: (1) collecting data and constructing a database; (2) selecting data and samples from the original database; (3) using the sorted Data training linear SVM, and save the training results; (4) judge whether the event is a transformer fault event through the trained model; (5) explain the classification results, and adjust the power quality. The invention applies the SVM classifier based on the stochastic gradient descent algorithm to the classification and identification of transformer fault events, and can effectively solve the problem of online collection and classification of measured data in a power system. The present invention uses a stochastic gradient descent algorithm to iteratively update each sample. Even if the sample size is large, only tens of thousands or thousands of samples may be used to iterate to the optimal solution. Therefore, it is more suitable for today's growing requirements for online data collection and processing of electric power.

Description

technical field [0001] The invention relates to the field of information technology, in particular to a classification method for electric power online collection data. Background technique [0002] In the power industry, digital technology has been widely used in recent years. The amount of power grid information collected and recorded in real time by various systems has exploded. Many large power grid systems have a daily data volume of tens of gigabytes, or even hundreds of gigabytes. Therefore, how to make full use of data, quickly and effectively analyze, process, and refine to discover useful knowledge has become one of the key issues facing the power industry. [0003] The basic requirements for power system operation are: (1) to ensure safe and reliable power supply; (2) to have qualified power quality; (3) to have good economy. To meet these basic requirements, the normal operation of the transformer is inseparable. The power transformer is an important substation...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the deficiencies of the prior art, the present invention provides a classification method for electric power online data collection. Support vector machine is a pattern recognition method based on statistical learning theory, which can effectively solve limited samples, nonlinear and high-dimensional patterns The identification problem has become a standard tool in the field of machine learning and data mining, but the traditional SVM is not suitable for online processing of measured data

Images