Espectrometria no Infravermelho Médio e Calibração Multivariada por PLS para quantificação de biodiesel metílico e etílico de soja, milho e girassol em misturas com diesel

• Autor(a) principal: Fonseca, Fábio dos Santos
• Data de Publicação: 2011
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFU
• Texto Completo: https://repositorio.ufu.br/handle/123456789/17346; https://doi.org/10.14393/ufu.di.2011.143
• topic: Biodiesel; CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA

Most of the energy consumed worldwide comes from oil, coal and natural gas. These sources are limited, so alternative sources are needed and the search for such sources is very important. Vegetable oils have been mentioned as likely replacements for diesel fuel since the beginning of last century by having physical and chemical characteristics that could be close or suitable for a diesel engine. They can be used in diesel engines with or without processing. Since January 1st of 2010 was established the addition of 5% biodiesel to diesel in Brazil. The transformation of vegetable oil into biodiesel refers only to better adapt to the engine. The Brazilian Biodiesel Production Chain currently has no regulations regarding the matrix to be used, and is currently regulated by the availability of raw materials. There is still no standardized method for analyzing biodiesel in diesel. Pure biodiesel is analyzed first and then added to diesel. In this work, eight multivariable calibration models based on middle-infrared spectroscopy were developed in order to determine the content of biodiesel in diesel fuel (binary blends), in the range of 1 to 20% v/v, generically called BX. Esters (methylic and ethylic) made with soybean, corn, sunflower seed oils and used soybean frying oil were used to prepare the blends with conventional diesel. Results indicated that partial least squares (PLS) models based on MID infrared spectra were proven suitable as practical analytical methods for predicting biodiesel content in conventional diesel blends in the volume fraction range from 1% to 20%. PLS models were validated by independent prediction set and the RMSEPs (Root Mean Square Errors of Prediction) were estimated as low as 0,2266; RMSECs (Root Mean Square Errors of Calibration) were estimated as low as 0,1532. Linear correlations were observed for predicted vs. observed values plots with correlation coefficient (R) greater than 0,999. Additionally, the average preview errors found were less than 2,21%.