Application of Artificial Neural Network (ANN) in predicting mechanical properties of canola stem under shear loading
Keywords:canola, cutting energy, stem, shear strength, power consumption, Artificial Neural Network (ANN)
In this study, at first the shear parameters including the maximum shear force, shear strength, shear energy and power consumption of canola stem were calculated through force-deformation curve; and then these mechanical properties were determined and predicted using artificial neural network. For the tests, testing machine Instron (Model Santam STM-5) with 50 N load cell was used. Stems were cut at 3 diameter levels (1 to 3, 3 to 5 and more than 5 mm), 3 cutting speed levels (75, 115 and 150 mm/min ), 3 cutting angles (0°, 30° and 60°) and three replicates. Cutting parameters including maximum cutting force, shear strength; cutting energy; consumed power and cutting work were examined. Tests lasted for each stem until the full cut. Data requirements were obtained from Force-Deformation curve. The results showed that by increasing the diameter and cutting angle, cutting force values, shear strength, cutting energy, cutting power and cutting work increased. Additionally, with increasing cutting speed, the cutting force, shear strength, cutting energy, cutting power and cutting work declined. Feedforward network was employed to predict some of the mechanical properties of canola stem. The results of statistical analysis using artificial neural network showed that the best values for shear energy, shear force, shear strength, shear power and shear work in canola stem were, respectively, in the epochs of 194, 2000, 275, 92 and 350 and also showed that neural networks can be used in intelligent cutting mechanisms and predicting mechanical properties of crops stem.