Page 481 - Decision Making Applications in Modern Power Systems

Chapter 17

Pattern-recognition methods

for decision-making in
protection of transmission lines

2
1
Mohammad Pazoki , Anamika Yadav and Almoataz Y. Abdelaziz 3
1 2
School of Engineering, Damghan University, Damghan, Iran, Department of Electrical
3
Engineering, National Institute of Technology, Raipur, India, Electrical Power and Machines
Department, Faculty of Engineering, Ain Shams University, Cairo, Egypt

17.1 Introduction
Pattern recognition is a scientific procedure that focuses on the automatic dis-
covery of regularities and relations in input patterns. Then, it assigns a label or
a real-valued output to each input patterns. The input patterns can be provided
based on the main goal of the procedure. They can be any type of events, mea-
surements, processes, and captures that need to be categorized or estimated.
Today the demand for practical functioning of pattern recognition has been
increased in parallel with promotion of core processors. Appearance of new
concepts in industrial systems and handling of large amount of information
generates a novel trend in the field of pattern recognition. Development and
implement of various systems of pattern recognition confirm the creation of
growing trend in the application and research. The system of pattern recogni-
tion can be categorized according to the fields of applications. For example,
classification of document image in the field of optical character recognition,
personal identification based on the face or finger print in the field of biomet-
ric, computer-aided diagnosis in the field of medical, automatic target recogni-
tion in the field of military, sequence analysis in the field of bioinformatics,
sorting objects in the field of industrial automation, power system protection
in the field of power electrical engineering, etc. The problems of pattern rec-
ognition have four common solutions, which are classified into four classes:
statistical, structural, hybrid methods of the statistical and structural, and artifi-
cial intelligence (AI) approaches. The last one is recommended for more com-
plex problems where there is not a linear and direct relation between the input
patterns and target vector.

Decision Making Applications in Modern Power Systems. DOI: https://doi.org/10.1016/B978-0-12-816445-7.00017-7
© 2020 Elsevier Inc. All rights reserved. 441