Aplicação de uma plataforma de clps para detecção de falhas em um sistema de controle a eventos discretos

Abstract

This work consists of the development of a centralized fault management system, which enables the assignment of a programmable logic controller platform for training artificial neural networks (ANN) as a fault detection tool in discrete event control systems, allowing different implementation strategies for distributed control systems, controller arrangements, application of fault diagnosis and treatment techniques in automation systems, as well as reference for other systems that have time constraints. When an error is not identified and promptly corrected, the entire monitoring of a network is compromised, and its subsystems become vulnerable to sudden stops during operation. This work proposes the use of an ANN, which is responsible for the task of modeling large systems with several functionalities, being a viable solution to analyze the operation time of each system and the latency to handle interruptions. The strategy applies to modeling the system and identifying the processes according to the operation time of each subsystem belonging to the set. The redundant programmable logic controller will act as a compensator, so that there are no premature stops in the operation, during maintenance in the main system with error. The results of the simulations performed serve as a basis in the case study. Fault detection strategies in a simulated industrial plant are also validated using software.