Skip to main content
ARS Home » Plains Area » Brookings, South Dakota » Integrated Cropping Systems Research » Research » Publications at this Location » Publication #258526
Title: Artificial neural network modeling of DDGS flowability with varying process and storage parameters

Author
item BHADRA, RUMELA - South Dakota State University
item MUTHUKUMARAPPAN, K - South Dakota State University
item Rosentrater, Kurt

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 9/3/2010
Publication Date: 10/7/2010
Citation: Bhadra, R., Muthukumarappan, K., Rosentrater, K.A. 2010. Artificial neural network modeling of DDGS flowability with varying process and storage parameters. ASABE/CSBE Intersectional Meeting, Saskatoon, Saskathewan, October 7-9, 2010.

Interpretive Summary: Computer modeling techniques were used to predict flowability behavior of distillers dried grains with solubles (DDGS) prepared with varying CDS (10, 15, and 20%, wb), drying temperature (100, 200, and 300°C), cooling temperature (-12, 0, and 35°C) and cooling time (0 and 1 month) levels. These models were used to predict various granular flow parameters, including aerated bulk density, packed bulk density, Hausner Ratio, Angle of Repose, Total Flow Index, Total Flood Index, and Jenike Flow Function. Various models were developed in order to predict single response variables or multiple response variables simultaneously. The performance of these models was then compared based on goodness of fit and error produced. Results from these models will be presented and discussed. Finally, the best combination of model parameters and input variables for predicting DDGS flowability will also be presented. Modeling of DDGS flowability has not been previously done, and hence this work will be a step towards solving this industrial challenge.

Technical Abstract: Neural Network (NN) modeling techniques were used to predict flowability behavior in distillers dried grains with solubles (DDGS) prepared with varying CDS (10, 15, and 20%, wb), drying temperature (100, 200, and 300°C), cooling temperature (-12, 0, and 35°C) and cooling time (0 and 1 month) levels. Response variables were selected based on our previous research results, and included aerated bulk density, packed bulk density, Hausner Ratio, Angle of Repose, Total Flow Index, Total Flood Index, and Jenike Flow Function. Various neural network models were developed using multiple input variables in order to predict single response variables or multiple response variables simultaneously. The NN models were compared based on R2, mean square error (MSE), coefficient of variation (CV), and Akaike information (AIC) produced. In order to achieve results with higher R2 and lower error, the number of neurons in each hidden layer, step size, momentum learning rate, and number of hidden layers were varied. Results comparing different NN models with various input, output, and model parameter options will be presented and discussed. Finally, the best combination of model parameters and input variables (i.e, with highest R2, least MSE, least CV, and least AIC) will be chosen for predicting DDGS flowability. Modeling of DDGS flowability using NN has not been previously done, and hence this work will be a step towards application of intelligent modeling procedures to this industrial challenge.