Aplicação de espectroscopia no infravermelho próximo a fim de realizar medições de umidade e atividade de água em alimentos para cães

Abstract

The segment destined for animal food production has been growing, having a unique importance in the economy. The present study seeks to act in a way that allows greater agility in routine analyses in a quality control laboratory in this area. Initially, samples of the chosen product were collected in the production line and analyzed in the spectrometer and on the bench. From the results obtained, it was possible to build an algorithm in Python language, with the help of available packages, to verify the consistency, normality, and presence of outliers values in the experimental data and the accuracy of those estimated. The analyses demonstrated, in a preliminary way, that the humidity curves associated with the dryer and post-bath had a satisfactory calibration. In contrast, the others – including those referring to water activity – required a greater quantity and range of data to adjust the mathematical parameters that could represent the experimental set more robustly. To assess the economic viability of the proposal, the discounted payback time was calculated for a profit range, between 10% e 80%, associated with production, 14,300 tonne/year, and sale, resulting in values less or equal to 4 years for profit margins considered conservative, with the need to study more assertively which scenario represents reality more feasibly. The possibility of increasing the cash flow associated with the project developed by applying the methodology used for the other products in the catalog was also considered qualitatively.