On-line diagnosis and evaluation method of insulation state of large electric machine

Abstract

The invention relates to an on-line diagnosis and evaluation method of the insulation state of a large electric machine, aiming at accurately diagnosing and evaluating the insulation state by comprehensively considering the influences of working environmental factors, such as vibration, temperature, humidity, and the like, of the electric machine on partial discharge. The on-line diagnosis and evaluation method comprises the following steps of: firstly building a neural network used for identifying discharge modes; constructing a neural network frame in an MATLAB2007 (Matrix Laboratory 2007); extracting a sample from a sample library to train the neural network; respectively building relational models of various discharge modes and the working environmental factors influencing the discharge modes according to the action relationship of the working environmental factors on the different discharge modes; installing various sensors on the large electric machines, and acquiring data; establishing a database; calculating a characteristic value; correcting to obtain a Qm value at a standard working environment; longitudinally comparing corrected Qm historical data read from the database with the current corrected Qm value, and horizontally comparing Qm among three phases of same equipment and among all electric machines so as to obtain the insulation state according to corresponding rules. The invention provides a diagnosis and evaluation conclusion of an insulation fault.

Description

technical field [0001] The invention relates to a large motor online monitoring and fault diagnosis and evaluation method. Background technique [0002] When large motors are in operation, they are subjected to long-term effects of the working environment and electrical, thermal, and mechanical stresses, and their insulation performance is gradually aged and damaged, which eventually leads to insulation accidents. Such accidents account for about 40% of large motor failures. Therefore, on-line diagnosis and evaluation of the insulation state of large motors has very important practical significance for improving the reliability of large motors. The invention patent with the patent number CN1402015A "Diagnosis Method and Device for Motor Insulation Aging Based on Wavelet Transform" uses an impact source to strike the main insulation surface of a stator bar of a large generator, and an acoustic sensor receives the radiated sound waves. The wavelet transform is performed on th...

AI Technical Summary

Problems solved by technology

Therefore, the partial discharge analysis method that ignores the working environment factors has low reliability and cannot accurately diagnose and evaluate the insulation state

Images