Desenvolvimento de um fotômetro multi LED NIR para determinação de metano em gás natural

• Autor(a) principal: Barbosa, Mayara Ferreira
• Data de Publicação: 2019
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFPB
• Texto Completo: https://repositorio.ufpb.br/jspui/handle/123456789/19299
• topic: Gás natural; Sistema automático para preparação de misturas; Espectrometria NIR; Calibração multivariada; Fotômetro à base de LEDs; Natural gas; Automatic system for preparation of gas mixtures; NIR spectrometry; Multivariate calibration; Photometer based on LEDs; CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA

Natural gas (NG) is a fossil fuel considered clean due to the lower emission of polluting gases when compared to coal or petroleum. NG can be used as an input or raw material, being well known for its versatility and efficiency. As it is present in residences, commercial establishments, industries and in the transportation, the necessity and the demand for the quality control of this type of fuel grows, mainly with respect to its composition, since, as influences occur in its calorific power. The composition of NG marketed in Brazil follows a standard of compliance determined by the Agência National de Petróleo, Gás Natural e Biodiesel (ANP), guaranteeing the energy efficiency of this fuel. Thus, this work proposes the development of a multi LED NIR photometer that indicates the composition of the natural gas with respect to its major component (methane). Therefore, an automatic system was used for the preparation of gas mixtures and evaluation of the NIR region, by means of the multivariate calibration associated with a bench and a miniaturized NIR spectrophotometer. In the first step, using the bench equipment, the iPLS, PLS and GA - MLR models obtained RMSEP values of 2.08, 1.45 and 1.00 % mol mol-1 and correlation coefficient (r) equal to 0.97, 0.90 and 0.88, for determination of methane, ethane and propane, respectively. In the second part, a miniaturized NIR spectrophotometer was evaluated for the same purpose in natural gas samples and in a biogas sample. For methane, ethane and propane, the metrics SW - MLR, GA - MLR and iPLS, provided RMSEP values of 3.38, 3.05 and 1.58 % mol mol-1 with r values of 0.97, respectively. Therefore, the construction of the photometer in the NIR region converged to LEDs at wavelengths in 2020, 2140, 2200 and 2350 nm. The detector used was the PbS and the control of the photometer was performed via Arduino board with a program written in Matlab.The LED with the highest intensity and sensitivity to hydrocarbons was 2350 nm, and this one was used to determination of methane in gas mixtures with ethane and propane. The photometer presented linearity in the absorbances obtained as a function of the methane concentration in the gaseous mixtures with the other hydrocarbons, with R2 equal to 0.9852, in the calibration range of 50-100 % mol mol-1 .The ANOVA results showed significance in the regression without lack of fit with LOD and LOQ values of 2.92 and 8.84 % mol mol-1 , respectively. In certified and commercial samples prediction, the methodology did not present statistically significant differences with the reference methodology (GC), obtaining a confidence limit of ± 2.85 % mol mol-1 .The methodology with the photometer was also validated by means of a recovery test, in which a mean and standard deviation of 100.2 ± 0.9 were obtained for tests involving certified and commercial samples.