Experimental investigation and modeling of drying process of balangu seeds gum using infrared dryer by genetic algorithm-artificial neural network method

The genetic algorithm (GA) optimization method can be used to overcome the inherent limitations of artificial neural network (ANN). Genetic algorithm–artificial neural network (GA-ANN) method has a high capability to find the optimum value of a complex objective function. In this study, first, to balangu seeds gum drying, an infrared dryer was used. In this infrared dryer, the effect of distance of samples from lamp at three levels of 5, 7.5 and 10 cm and the effect of height of the gum inside the container at three levels of 0.5, 1 and 1.5 cm on drying time and weight loss percentage of balangu seeds gum during drying time, were investigated. The results of balangu seeds gum drying using infrared method showed that with decreases in sample distance from the heat source and also with decreases in thickness of the gum in the sample container, drying time were decreased. With increasing in the lamp distance from 5 to 10 cm, the average drying time of balangu seeds gum increased from 62.6 minutes to 87.6 minutes. With sample thickness increasing from 0.5 to 1.5 cm, the average drying time of balangu seeds gum increased from 45.9 to 109.2 minutes. In the next step, this process was modeled by GA-ANN method with 3 inputs (radiation time, lamp distance from samples surface and thickness of samples) and 1 output (weight loss percentage). The results of modeling with GA-ANN method showed that the network with structure of 3-9-1 in a hidden layer and using the hyperbolic tangent activation function could predict the weight loss percentage of balangu seeds gum during drying in an infrared dryer with high correlation coefficient (0.999) and low mean squared error (0.788).