The processes of transportation of bulk materials from silos and hoppers are significant in various industrial applications because of their influences on material characteristics and working parameters of the production process. In this paper, a rotating valve feeder, with eight vanes was investigated for transport action of bulk materials, such as wheat, maize and rice, which were ground, using the sieve sizes of 1, 3 and 5 mm. The rotating valve feeders under investigation have proven to be useful in transportation processes despite their construction simplicity. All investigations were done experimentally and numerically, using coupled Discrete Element Method (DEM) and Computational Fluid Dynamics calculation (CFD). The influences of different types of bulk materials and its particle size, on the performances of the rotating valve feeder during material transport were explored. The artificial neural network was developed (in the form of a multi-layer perceptron model) in order to optimize the granular flow of the bulk material, showing the high prediction capability of bulk density, dosing time and granular material flow, with the coefficient of determination equal to 0.999 during the training period. The decreasing of the sieve opening diameter caused the decrease in bulk density of the ground material, but statistically significant only for rice, as seen from the experiments and the results of the neural network model. The 5 mm sieve ensured the material with the highest flowability, significantly increasing the granular flow and decreasing the dosing time. The granular particles were modelled as the spheres in the DEM/CFD simulation, with a small-sized triangular surfaces. The DEM/CFD prediction of the mass transport for rice, wheat and maize was quite adequate, obtaining the coefficient of determination being 0.997; 0.998 and 0.849, respectively.