Resultados de incerteza de calibração para sensores infravermelho do tipo MEMS termopilha

• Ano de defesa: 2021
• Autor(a) principal: Vitor Furtado Paes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ENG - DEPARTAMENTO DE ENGENHARIA MECÂNICA; Programa de Pós-Graduação em Engenharia Mecanica; UFMG
• Palavras-chave em Português: Termografia; Sensores térmicos; Calibração de termopilhas; Incertezas de medição; Engenharia mecânica; Termômetro e termometria; Incerteza; Sensores
• Link de acesso: http://hdl.handle.net/1843/39338

The calibration of IR sensors is still a subject of debate. The scientific literature still strugles with the systematic evaluation of measurement uncertainties, which hinders the comparison between calibration results and the reliability of temperature measurements. Despite this, MEMS (Micro-Electro-Mechanical Systems) thermopile arrays are primarily used for quantitative temperature measurements, then a study that puts the accuracy of this type of sensor to the test is critical. Given this background, this work presents the calibration results of tree units of commercial MEMS thermopile arrays provided by the same manufacturer. A calibration methodology is proposed, based on the combination of mathematical models and experimental procedures divided into the following parts: (1) Experimental data collection; (2) Application of algorithms (radiation heat transfer, regression models and Non-uniformity Correction models; (3) Estimation of uncertainty sources and estimation of expanded uncertainties. Uncertainty estimations are based on the GUM method (Guide to the Expression of Uncertainty in Measurement) and took into account the uncertainty sources of the blackbody radiator, radiation heat transfer model, mathematical models integration, and IR thermometers. The results led to a maximum deviation of 0,46 ◦C, in the RBF model and 0,49 ◦C in the Sakuma-Hattori model for a measurement range of 30 ◦C to 80 ◦C. Uncertainties related to the blackbody radiator, mainly uniformity, and temperature, were dominant in the uncertainty budget, followed by the mathematical model’s propagation of errors and temporal noise. The uncertainty results were compared with the scientific literature and the manufacturer’s data. The maximum calibration uncertainty provided by this work (± 1, 9 ◦C) is lower than the one provided by the manufacturer (± 2, 5 ◦C). The results also indicates the importance of a pixel-by-pixel uncertainty analysis to verify the instruments’ reliability.