Secagem e armazenagem de grãos de soja: efeitos sobre a qualidade física e físico-química, modelagem e predição

Abstract

The quantitative and qualitative losses of post-harvest grains bring an imbalance in the grain production sector. To reduce losses, it is essential that the grain mass goes through cleaning and drying processes, to be stored with low levels of water and impurities. The heterogeneity of the batches of grains harvested at the beginning and end of the harvest can also alter the capacity and uniformity of the processes. Thus, the general objective of the study was to evaluate different technologies and management in the post-harvest of soybeans, based on the harvesting of grains with higher water contents, associated with drying and storage conditions and technologies and the effects on physical and physical chemistry of grains. Specifically, the objective was: 1) to evaluate different drying and storage technologies on quality losses in soybeans; 2) evaluate the effects of storage and storage operations on the quality of processed soybeans; 3) verify the use of mathematical models and multivariate analyzes to evaluate the relationship between the anticipation of the soybean harvest and the drying and storage conditions and the influences on the physical-chemical quality of the grains; 4) analyze the prediction of the quality of soybeans in different drying and storage technologies, on a real scale, using Machine Learning models. Among the results obtained, it was observed that: 1) the management of the grain mass in drying silos and continuous dryers reduced losses and guaranteed better grain quality; 2) grain quality losses due to drying management ranged from 0.23 to 3.26% in crude protein and from 0.15 to 3.05% in crude oil yield. Managing drying with a continuous dryer + silo-dryer-CDSD2, continuous dryer + silo-aerator-CDAS3 is an alternative for reducing losses and conserving grain quality, improving yield in relation to the protein and crude oil contents extracted in up to 95%; 3) early harvesting with water content above 23% and the adoption of drying systems with an air temperature of 80 °C in environments with temperatures below 23 °C preserved the physical-chemical quality of the grains; 4) the grains subjected to drying and storage in drying silos maintained the better quality at the end of the process. Although there were differences related to drying and storage technology in relation to changes in grain quality, it was noted that the Artificial Neural Networks model demonstrated superior performance in predicting grain quality. The Artificial Neural Networks model was unanimous in all processes and technologies evaluated. Therefore, it is recommended to carry out post-harvest drying of soybeans and subsequent storage of grains in drying silos, monitoring environmental and intergranular variables. It is recommended that this approach be associated with the use of Artificial Neural Network models to predict losses with greater efficiency in the drying and storage stages.