The Wavelet Based Detecting of the Signalling Relay Armature Defects

TLDR: To prompt detection of all possible armature defects, to use the wavelet Shannon entropy as an integral parameter, the value of which strongly depends on the appearance of any order disturbances in the systems was proposed in the work and results confirmed their effectiveness.

Abstract: The article deals with the problem of automatic detection defects of armature and contact springs in railway signalling relays. Signalling relays are currently basic components in train control systems that ensure train traffic safety and to ensure their reliable operation, the relays are periodically subjected to routine maintenance. Most of the tests for signalling relays are performed manually, which is time consuming and very expensive. For developing the automatic method for detecting and identifying the relay armature defects the investigations were performed with ten different types of neutral signalling relays. For each type of relay, several groups were formed, which consisted of three relays, in which one group contained fully serviceable and properly configured relays, while the other groups contained relays with some artificially created defects in the form of deformed contact springs or defects in the coil or magnetic core. To identify the features of the basic defects of relays, the time dependence of the transient currents were measured during relay switching on and off. Three segments are clearly distinguishable on the transient current curve: the first segment when the anchor is in the initial position and the magnetic circuit is open, the second one when the armature moves from the initial to the final position, and the third one when the anchor is picked up and the magnetic circuit is closed. In the first and third segments the current in the relay coil increases exponentially with time constants depending only on the impedances of the coil, which allows to use these time constants for detecting defects in the relay coil and magnetic circuit. In the second segment the current in the relay coil is determined by electromechanical processes that occur during the movement of the armature, which include magnetic, friction, deformation forces etc., and defects of moving components lead to a change in the shape of transient current curve and to appearance of some features in the form of sags and swells on it. Du to these features are small, using the wavelet decomposition of the transient current dependences for detecting defects of armature in signalling relays was proposed in the work. For prompt detection of all possible armature defects, to use the wavelet Shannon entropy as an integral parameter, the value of which strongly depends on the appearance of any order disturbances in the systems was also proposed in the work. The effectiveness of the proposed methods have been investigated and obtained results confirmed their effectiveness.