A Fault Diagnosis Method for Regional Power Network Based on Wavelet Transform and Rough Set Theory

Abstract

The invention discloses a regional power grid fault diagnosis method based on wavelet transform and rough set theory. The fault original decision table; then use the genetic algorithm based on rough set theory to reduce the fault original decision table to obtain the reduced fault decision table, and extract the fault diagnosis rules from the table; The voltage and current signals are collected, and then the sequence of suspicious fault components is obtained according to the collected voltage and current signals; finally, according to the fault diagnosis rules and the sequence of suspicious fault components, it is judged whether each suspicious fault component has a fault, thereby completing the fault diagnosis of the regional power grid. The invention avoids the problem of possible misjudgment of the fault diagnosis of the traditional power grid in the case of protection malfunction, and can perform real-time and accurate fault diagnosis of the regional power grid.

Description

technical field [0001] The invention belongs to the field of fault diagnosis, in particular to a method for fault diagnosis of regional power grid based on wavelet transform and rough set theory. Background technique [0002] The scale of the modern power system network continues to expand, power generation and transmission have entered the era of high voltage and large units, and the degree of inter-regional grid networking has become closer. Once a fault occurs, it is necessary to quickly diagnose the faulty component and restore the power supply to ensure the safe and stable operation of the power system and avoid greater losses. At the same time, as the power grid monitoring devices become more and more intelligent, a large number of alarm information will flood into the dispatch center when there is a failure, and the uncertain or incomplete information and redundant information will be analyzed by the dispatcher. Judgment brings a huge workload, so it is of great sign...

AI Technical Summary

Problems solved by technology

Among them, the expert system is the earliest artificial intelligence technology applied to power grid fault diagnosis, which can improve the accuracy of decision-making and reduce the workload of operation dispatchers, but it has poor learning ability, difficult to cover comprehensive knowledge base, and low fault tolerance. Or the diagnosis speed of complex fault conditions is very slow, etc.

In the fault diagnosis of artificial neural network, there are difficulties in the training and learning workload and limited fault tolerance.

However, the optimization evolution process of genetic algorithm, particle swarm algorithm and other optimization methods is affected by the initial value and some initial parameters, and there are cases where the convergence speed is slow and the convergence is not satisfied.

Rough set theory can deal with incomplete information and redundant information well, and is suitable for power grid fault diagnosis. However, in the research, there are multiple complex fault information, which leads to too large decision table and too much reliance on pre-formed decision table for diagnosis decision rules. question

Images