Reliability analysis using experimental statistical methods and AIS: application in continuous flow tubes of gaseous medium

Autor(es)

Outa, Roberto ^[1] ; Chavarette, Fabio Roberto ^[2] ; Goncalves, Aparecido Carlos ^[3] ; da Silva, Sidney Leal ^[4] ; Mishra, Vishnu Narayan ^[5] ; Panosso, Alan Rodrigo ^[6] ; Mishra, Lakshmi Narayan ^[7]

Afiliação do(s) autor(es)

^[1] Fatec Fernando Amaral Almeida Prado, Dept Biocombustiveis, Aracatuba, SP - Brazil

^[2] Univ Estadual Paulista, Dept Engn Fis & Matemat, Inst Ouim, Rua Prof Francisco Degni 55, BR-14800060 Araraquara, SP - Brazil

^[3] Univ Estadual Paulista, Dept Engn Mecan, Ilha Solteira, SP - Brazil

^[4] Fac Tecnol Itaquera Miguel Reale, Dept Proc Soldagem, Itaquera, SP - Brazil

^[5] Indira Gandhi Natl Tribal Univ, Dept Math, Anuppur, Madhya Pradesh - India

^[6] Univ Estadual Paulista, Dept Engn & Ciencias Exatas, Jaboticabal, SP - Brazil

^[7] Vellore Inst Technol, Sch Adv Sci, Dept Math, Vellore, Tamil Nadu - India

• Tipo de documento: Artigo Científico
• Fonte: ACTA SCIENTIARUM-TECHNOLOGY; v. 43, JAN-DEC 2021.
• Citações Web of Science: 0

Resumo

The motivation for the development of this work arose from the observation of maintenance in pressure vessels, which are categorized as highly hazardous security risk products. The costs of detecting failures in the production systems allow the result of the process to be safe and of good quality, using standardized tests internally within the company. The main objective of this work demonstrates the efficiency and robustness of the artificial immune system (AIS) of negative selection in the detection of failures by recognizing the vibration signals and categorizing them in the degree of probability and level of severity of failures. The intrinsic objectives are the application of the elimination of signal noise by the Wiener filter, and the processing of data-Wiener data using experimental statistics. The result of this work successfully demonstrates the precision between the experimental statistical and AIS techniques of negative selection; the robustness of the algorithm in precision and signal recognition; and the classification of the degree of severity and probability of failure. (AU)

• Processo FAPESP: 19/10515-4 - Prognose e detecção de falhas em estruturas mecânicas utilizando computação natural
• Beneficiário: Fábio Roberto Chavarette
• Modalidade de apoio: Auxílio à Pesquisa - Regular