Modeling and Optimization of Parameters affecting Drying of Corn Kernels (Zea Mays) in Convective Tray Dryer

Abstract

The present work involves the application of artificial neural network (ANN) and response surface methodology (RSM) for corn kernels drying in a convective tray dryer. The process parameters which affect most the drying of corn kernels namely: air velocity (2-5 m/sec), amount of feed (50-100 g), temperature (50-80 ℃) and drying time (30-60 min). These parameters affect the drying process tremendously. Predictive modelling has been done using RSM and ANN. With ANN, experimental results have been evaluated to train, test and validate so that behaviour of the system can be predicted. There were 10 neurons used for ANN model; LM (Levenberg-Marquardt) showed as a suitable training function; MSE as a performance function; GDM as a learning suitable function for simulation of drying operation. The coefficient of determination values (R) is 0.99916 for training, 0.99071 for validation, 0.96091 for test and 0.97527 for all the results, value of MSE is 0.0002. Experimental and predicted moisture removal regression coefficient is 0.8085 and 0.9348 with RSM and ANN model, respectively. ANN model shows better behaviour than Box-Behnken design.