| |
|
SR21 Redes Neurais, Algoritmos Genéticos e Lógicas Nebulosa/Neural |
| |
|
Title:
|
TÉCNICAS NÃO DESTRUTIVAS DE MONITORAMENTO E DETECÇÃO DE FALHAS ESTRUTURAIS UTILIZANDO REDES NEURAIS
|
| |
|
Summary :
|
|
ABSTRACT: THE USE OF NEURAL NETWORK AS NON-DESTRUCTIVE EVALUATION TOOL FOR FAULT DETECTION BASED ON VIBRATION MEASUREMENT IS GAINING MORE IMPORTANCE IN THE CONTEXT OF OPERATION CON-DITION AND HEALTH MONITORING OF MECHANICAL SYSTEMS. THE PROPOSE CONSIST IN THE MONITORING OF THE VIBRATION CONDITION OF THE STRUCTURE THAT WILL BE EVALUATED IN ORDER TO DEFINE A SET OF DATA ABLE TO IDENTIFY IT AND SUITABLE FOR THE TRAINING PROCESS. HOWEVER, THERE EXIST THE NECESSITY OF GENERATION OF A LARGE SET OF DATA CONTAINING THE DIFFERENT PATTERNS REPRESENTING ALL FAILURE CON-DITION THAT IS LIKELY TO OCCUR IN THE STRUCTURE. THIS WORK STUDIES THE CAPABILITY OF GENERALIZATION OF NEURAL NETWORK USING THE BACKPROPAGATION ALGORITHM TO CLASSIFY THE INPUT PATTERS OF A SYS-TEM WITH DIFFERENT STIFFNESS RATIO. THIS WILL BRING TO DEFINITION OF NEURAL NETWORKS CAPABLE TO USE A REDUCED SET OF DATA DURING TRAINING PHASE AND, ONCE IT IS SUCCESSFULLY TRAINED, IT COULD IDENTIFY INTERMEDIATE FAILURE CONDITION. SEVERAL CONDITION AND INTENSITY OF FAULT HAS BEEN STUD-IED BY USING NUMERICAL DATA. THE NEURAL NETWORK DEMONSTRATED A GOOD CAPACITY OF GENERALI-ZATION FOR ALL CASE STUDIED. IN A NEXT STEP THIS GENERALIZATION CAPABILITY WILL BE TESTED WITH EX-PERIMENTAL DATA.
KEYWORDS: NEURAL NETWORK, BACKPROPAGATION, GENERALIZATION, FAULT DETECTION.
|
| |
|
Author :
|
DEMARCHI, DANIELA 0
LOPES JUNIOR, VICENTE 0
PEREIRA, JOÃO ANTONIO
|
| |
|
Paper View :
|
|
COBEM99
| Organizing Committee | Keynotes
| Technical
Sessions | Author
| Symposia and special sessions
| Referees |
Work
Titles | Venue
|
