Regulagem da vazão de saída de um silo de trigo: testes experimentais e modelagem

Abstract

In a local wheat flour factory, the mass flow during the emptying of a grain silo is controlled in an inefficient manner, which makes the transport pipelines prone to clogging if they are fed too much grain. There is also the risk of the waste of electric energy if the volume of grains provided is not large enough. In order to make the flow control more trustworthy, finding a better method to measure the outflow as a function of the size of the aperture of a valve, which controls the volume of grains leaving the silo was necessary. With that in mind, the mass flow exiting one of the factory’s silos was measured multiple times, while varying the filling of the silo and the size of the aperture of the valve. Following the collection of the data, five mathematical models were chosen, and their fitting to the experimental data was evaluated using the Curve Fitting Toolbox, from the software MatLab. It was taken into account, during the evaluation, that the the models were primarily developed for scale sized systems, in a controlled environment, as opposed to this works experiments, which were done in a large scale silo, inside a real industrial plant. By evaluating smaller partitions of the data separately, it was found that some points were influenced by external factors during their measurement, and their mass flow showed up as larger than the real flow exiting the silo. Furthermore, those points were ignored, and three applying ranges were defined using the models and coefficients that provided the best fit to the remaining experimental data. Thus, it was possible to relate the aperture of the valve to the emptying outflow, for mass flows ranging from 16,50 to 39,60 kg/s, with a minimum adjustment factor R² of 0,8653. It is important to mention that this work was successful in fitting some of the models evaluated to data collected from a setup that was different from the ones that were used for the development of the mathematical models. That difference being the addition of a rectangular pipe under the silo, where the grain flow passes and is controlled, before being send into the factory. The consequence of that setup difference was observed and interpreted as some divergences between the coefficients reported in previous works and the ones found by this work’s experimental data fitting.