<![CDATA[Recirculation drying process ofmaterial on pneumatic conveying recirculated dryer (PCRD) are very complexand not linear, so it is very difficult to predict the final required moisture content.The purpose of this study was to develop a model of Artificial Neural Networks (ANN) to predict the final moisture content of the material on the PCRD machine. In this study, PCRD machine has been designed with variability in recirculation, and ANN Graphical User Interface (GUI) application using Neural Network in computer software. AAN models have been designed using the structure of a network with 11 input neurons, hidden multilayers neurons, and one output neuron with backpropagation learning algorithm. Training and testing of models using 54 and 27 data set observations respectively. The validity test results of the model obtained the value of r2 trainning was 0.99 or 99%, and r2 of the testingwas 0.96 or 96%. This indicated that the models are very valid to predict the final moisture content of the materialon the PCRD machine. The results also revealed RMSE, MAE, MRE value of ANN optimization model was 0.118% wb, 0.056% wb, and 0.644% respectively. While the value of RMSE, MAE, MRE ofthe process of the model testing was 0.226% wb, 0.129 % wb, and 1.496% respectively.Keywords: prediction, moisture content, models, pneumatic recirculated conveying dryer, artificial neural network]]>
Copyrights © 2017
