Monitoramento amostral da temperatura, umidade relativa e dióxido de carbono intergranular para estimativa e predição da perda de matéria seca em grãos de trigo armazenados em silos verticais

Abstract

Storage constitutes an ecosystem, which can influence the metabolic activity and quality of wheat grains. Among the main variables that interfere in the respiration of the grain mass are the temperature and intergranular relative humidity. The increase in temperature and intergranular relative humidity can alter the hygroscopic equilibrium humidity and consequently the water content of the grains, causing losses. Thus, the present study aimed to evaluate the temporal/sample monitoring of temperature, relative humidity and intergranular carbon dioxide for estimation and prediction of dry matter loss in wheat grains stored in vertical silos, using a mathematical model and machine learning algorithms. Among the results obtained, the temperature of the intergranular air remained close to the average ambient temperature (26.5 to 26.7 °C), however, when it approached twenty hours of monitoring, it had a reduction to 26.4 °C. While the relative humidity remained between 61.3 to 62.5% throughout the monitoring time. Under these conditions, the stored wheat grains remained with equilibrium moisture content between 11.65 and 11.82%. The concentration of carbon dioxide throughout the storage time ranged from 300 to 550 ppm in the grain mass. Under these conditions, wheat grains had a slight increase in metabolic activity, causing a dry matter loss of 0.035 to 0.041%. The intergranular temperature, relative humidity and carbon dioxide results were similar to the measurements on the surface of the grain mass. It was concluded that the monitoring of temperature, relative humidity and the concentration of carbon dioxide in the intergranular air indirectly and early determined the changes in the quality of wheat grains during storage. Although the metabolic activity of wheat grains was low, as the lots remained in hygroscopic equilibrium with water contents close to 12% (w.b.), the temporal detection characterized the loss of dry matter and the reduction of grain weight. stored. Finally, the Artificial Neural Networks and Multiple Linear Regression model satisfactorily predicted the dry matter loss of stored wheat grain mass.