flow rate determination of granular material by using sound and multivariate data analysis

in this study, using sound created by wheat grain passing through a pipe, wheat flow rate was measured. the developed device consists of a hopper, metering device, sound sensor, delivery tube, dc motor and power supply. several wheat mass flow rates were tested and the sound signal created by the passage of the grain through the discharge tube was measured and transferred to a computer using data acquisition card (dac). utilizing matlab signal processing toolbox and wavelet transfer functions, it was possible to extract frequency characteristics of the sound signals used as distinguishing features of the different flow rates. artificial neural networks-multilayer perceptron (ann-mlp) and discriminate analysis (da) were used to classify different wheat flow rates. results showed that by using da and ann-mlp it was possible to determine the wheat flow rates with 97% and 89% accuracy from each other respectively.