Adaptive Distributed Structure Models for Recognizing Input Signals in High-Speed Relay-Protection Applications

• Published in: Russian electrical engineering Vol. 94; no. 8; pp. 539 - 545
• Main Authors: Vorobyev, E. S., Ubaseva, M. V., Antonov, V. I., Ivanov, N. G., Soldatov, A. V.
• Format: Journal Article
• Language: English
• Published: Moscow Pleiades Publishing 01.08.2023; Springer Nature B.V
• Subjects: Adaptive filters; Engineering; High speed; Machines; Manufacturing; Processes; Relay; Signal processing; Tuning; Prony’s method; distributed structure adaptive models; relay protection; residual signal filter; canonical signal component filter; adaptive structural analysis; noise filter; digital signal processing; composite filters
• ISSN: 1068-3712; 1934-8010
• DOI: 10.3103/S106837122308014X

In high-speed relay-protection applications, decisions are made based on signal processing over a short observation window. Classical adaptive signal models, which are prone to increasing their order so as to increase their resolution, do not cope with the task of recognition in a short window. The article proposes a new approach to constructing a signal model based on the use of an adaptive filter with a distributed structure. The proposed model is perfected by means of excluding canonical filters of controlled components from the competitive environment that arises in the classical adaptive filter. The filter-tuning scheme with a distributed structure is a multichannel system, in each channel of which a canonical filter is tuned to block of its component; the more accurately each individual canonical filter is tuned, the better the conditions are that are created for tuning the remaining canonical filters and the residual signal filter. The results of the computational experiment confirm the high performance of an adaptive filter with a distributed structure in terms of accuracy and speed when recognizing input signals of relay protection.