Detecting development scenarios of dynamic events in electric power networks smart-grid. Part 1 "Method".

Abstract

A method for constructing a semantic S-detector for the senior hierarchical level of pattern recognition in dynamic objects is
proposed. The S-detection method is based on a uniform synthesis of all hierarchically subordinate detectors and information
preparation schemes, based on the tasks of the upper hierarchical level of recognition of scenarios for the development of processes
in an object. The method can be used to solve problems of relay protection, emergency control automation, to identify rare events in
impulse, short-term input signals when implementing the concept of developing of electric power smart-grid. The synthesis of the Sdetector of scenarios for the development of transient processes was the result of the development of a structural-information SImethod for processing information components. A mathematical SI-model is presented that allows you to control the mutual
correspondence of the structure of the formation tree in the object and the device recognition tree. The task is to improve the stability
of the device recognition algorithm. The Part 1 describes the theoretical provisions of the application of the recognition method. It is
proposed to unify the trees for generating recognition scenarios in the industry, similar to the IEEE high-voltage network diagrams.
The S-detector algorithm based on the formation of a semantic signal and on the selective search for the required amount of
information for automatic decision-making is presented. The facts of the appearance of elementary information events in the input
signals of the detector are controlled by Ξ,H-selectivity windows, which are sequentially opened depending on the previous
development of events. The block diagram of the S-detector of the scenario is presented, which can be synchronous, multichannel,
with a high signal processing speed, without losing rapidly changing information when slow and infraslow information is detected.
Combines useful properties of instantaneous and inertial detection algorithms. The output of the S-detector is a series of responses
with different weights. This allows elementary detectors to be combined into sequential chains at different hierarchical levels of
recognition. Part 2 provides an example of the application of the scenario detection method to solve the problem of increasing the
stability of the operation of a selective protection relay against single-phase earth faults based on medium-frequency signals in 6-35
kV networks with an arc suppression Petersen’s coil. The results of modelling the operation of the S-detector algorithm on the signals
of real emergency files recorded by high-frequency recorders of transient processes in the network are presented.