Use of artificial intelligence for the estimation of effective moisture diffusivity, specific energy consumption, color and shrinkage in quince drying

In the study, drying process of quince fruit was accomplished in a microwave-convective dryer (MCD). The experiments were carried out at microwave power levels of 100, 200, and 300 W, air temperatures of 40, 55, and 70 degrees C, and air velocities of 0.5, 1, and 1.5 m/s. Nevertheless, three artificial intelligence techniques consisted of artificial neural networks (ANNs), particle swarm optimizer (PSO), and grey wolf optimizer (GWO) were evaluated to predict the parameters of D-eff, SEC, Delta E, and S-b. In the evaluation the data by ANNs, input parameters of networks consisted the values of air temperature, microwave power, and air velocity. According to the results, the maximum values of effective moisture diffusivity (D-eff) and specific energy consumption (SEC) were 1.71 x 10(-9) m(2)/s and 126.07 kWh/kg, respectively. In addition, minimum values of total change in color (Delta E) and shrinkage (S-b) of quince achieved 10.85 and 33.85%, respectively. For predicting all parameters, three models used in the study represented good predictive capability with R-2 > 0.97. The obtained results showed that the GWO model had better predictive performance than the ANN and PSO models. Practical Application Drying food and agricultural products by application of microwave-hot air blend dryers can be a good alternative to hot air and microwave dryers. Microwave energy infiltrates the product and facilitates heat release from the product and thus reduces drying time compared to single dryers. The main aim of applying such different models, mathematical simulation or modeling in the drying technology of agricultural products is to transform physical qualities and their interactions into numerical quantities and mathematical relationships.