Comparison of Mathematical Equation and Neural Network Modeling for Drying Kinetic of Mendong in Microwave Oven

6th Asian Physics Symposium

Maulidah, Rifa'Atul^1 ; Purqon, Acep^1

Physics of Earth and Complex Systems, Institut Teknologi, Bandung, Indonesia^1

关键词: Artificial neural network modeling;    Heat and mass transfer;    Levenberg-Marquardt;    Mathematical equations;    Neural network model;    Optimal parameter;    Predictive modeling;    Training function;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/739/1/012106/pdf
DOI  :  10.1088/1742-6596/739/1/012106

来源: IOP

【 摘 要 】

Mendong (Fimbristylis globulosa) has a potentially industrial application. We investigate a predictive model for heat and mass transfer in drying kinetics during drying a Mendong. We experimentally dry the Mendong by using a microwave oven. In this study, we analyze three mathematical equations and feed forward neural network (FNN) with back propagation to describe the drying behavior of Mendong. Our results show that the experimental data and the artificial neural network model has a good agreement and better than a mathematical equation approach. The best FNN for the prediction is 3-20-1-1 structure with Levenberg- Marquardt training function. This drying kinetics modeling is potentially applied to determine the optimal parameters during mendong drying and to estimate and control of drying process.