Análise de Métodos para Identificação de Tipos de Corrosão e de Substâncias Corrosivas através de Ruído Eletroquímico

• Ano de defesa: 2017
• Autor(a) principal: Alves, Lorraine Marques
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Elétrica; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Elétrica
• Palavras-chave em Português: Corrosion; Corrosive Waste; Electrochemistry; Noise; Recurrence Quantification Analysis; Análise de quantificação de recorrência; Corrosão; Resíduos corrosivos; Eletroquímica; Ruído; Wavelets (Matemática); Medidas Elétricas, Magnéticas e Eletrônicas; Instrumentação; 621.3
• Link de acesso: http://repositorio.ufes.br/handle/10/9576

Corrosion is a problem that affects several sectors of modern life, and in the industrialcontext is the main source of unplanned costs. Among the most common types, localizedcorrosion is present in most industrial processes and is the most difficult to detect. Thepresence of corrosion is a consequence of the exposure of structures and equipment tosubstances, which by electrochemical action, are capable of causing wear and surface orstructural failure. The consequences of corrosive processes can be considerably reduced bythe use of methods for detecting, analyzing and monitoring of hazardous areas in order toprovide useful information for accident prevention and maintenance planning.In this work, some techniques for the classification of different types of localized corrosion(pitting, crevice and watermark) and the occurrence of passivation are analyzed. Themethods analyzed were also applied for the detection of different types of corrosivesubstances. Such methods are based on the use of machine learning techniques and in theextraction of important information from electrochemical noise signals, which are signalsfrom corrosive processes.The techniques of electrochemical noise signal analysis are not still fully established,highlighting the importance of performing comparative studies in different contexts. In thiswork, the methods analyzed are based on the use of the Wavelet Transform and RecurrenceQuantification Analysis. For the corrosion type classification, the mean accuracy was95,86% using the Wavelet Transform and 91,02% using the Recurrence QuantificationAnalysis. In the classification of corrosive substances, the mean accuracy was 87,57% usingthe Wavelet Transform and 90,49% using the Recurrence Quantification Analysis. Theresults showed that the methods analyzed are promising in the classification of localizedcorrosion types and electrochemical sensing to identify the presence of corrosive substancesin several industrial processes.