Abnormal data detection method for power distribution network based on high dimensional random matrix

Abstract

The invention provides an abnormal data detection method for a power distribution network based on a high dimensional random matrix. The method comprises: constructing a high dimensional random matrix according to historical data; normalizing the high dimensional random matrix; defining a covariance matrix and obtaining the characteristic root of the covariance matrix; determining the spectral distribution function of the covariance matrix; and using the spectral distribution and ring rate to detect data anomalies. The invention provides the novel method for abnormal data detection of reactive optimization of the power distribution network. The method can accurately calculate abnormal data within any time period, is highly practical, and is high in detection efficiency.

Description

Technical field: [0001] The invention belongs to the technical field of distribution network operation, and in particular relates to a method for detecting abnormal data of a distribution network based on a high-dimensional random matrix. Background technique [0002] Random matrix theory originated from the development and research of quantum physics, a set of quantum statistics theory established by Winger and Dyson in the early 1960s. Random matrix theory is one of the important mathematical tools for statistical analysis of complex systems. Through the statistical analysis of the energy spectrum and eigenstates of the complex system, it can obtain the degree of randomness of the actual data and reveal the behavioral characteristics of the overall correlation in the actual data. [0003] A random matrix is ​​a matrix composed of random variables as elements in a given probability space. The large-dimensional data refers to the data whose sample dimension and sample size...

AI Technical Summary

Problems solved by technology

[0007] However, the existing abnormal data detection methods of distribution network mainly include (1) traditional bad data detection methods, such as weighted residual method, measurement mutation detection method, objective function extreme value detection method, etc.; (2) relatively new Detection methods, such as fuzzy equivalent matrix clustering analysis method, neural network method, etc.

These methods are usually only aimed at the small-scale structured data of the traditional distribution network, and only rely on a single data source and a single detection method, and the reliability of data detection is relatively low

Images