Prediction of effective moisture diffusity and specific energy consumption of eggplant in continuous band drying using some novel methods

Eggplant (Solanum melongenaL.) is being cultivated in North America, Asia and the Mediterranean area. Its limited shelf life is one of the important restrictions in the trade of eggplant as a fresh product. Drying is one of the most current methods used to maintenance agricultural products. This process improves the food stability, since it reduces significantly the water and microbiological activity of the material and minimizes physical and chemical changes during its storage. Dynamic modeling of drying characteristics for various agricultural products, including artificial intelligence techniques that include artificial neural networks (ANNs), particle swarm optimization (PSO) and grey wolf optimizer (GWO) Which has attracted a lot of attention recently, because the ability to learn from these systems to detect fruit and vegetable behaviors is a complex process in which mathematical models simply do not apply in recent decades. The main objective of this research was to determine the effective moisture diffusivity, and specific energy consumption of eggplant slices with a semi-industrial continuous band dryer. Moreover, some novel methods including ANN, PSO and GWO as an approximating tools were developed and evaluated for prediction of Deff and SEC of the process. Freshly harvested eggplant were purchased from a local market and stored in the refrigerator at about 4°C for experiments. The initial moisture content of eggplant was determined by oven method. About 40 g of eggplant slice (4 mm thickness) with three replicates were dried at 70°C for 24 h. Eggplant slice with average initial moisture content of 10.25% (d.b.) was chosen as the drying material.
The dryer consists of an adjustable centrifugal blower, hot air suction tube, heater, control panel, air channel to uniform distribution of hot air, drying chamber, Belt (20 cm, 200 cm), three inverters (LS, Korea), temperature and humidity sensors, electrical motor, removable upper part, base, shafts, three infrared lamps (Philips, Belgium) and belt guide. The experiments were performed at air temperatures of 45, 60, and 75C, air velocities of 1, 1.5, and 2 m/s, and belt linear speeds of 2.5, 6.5, and 10.5 mm/s. Feed and cascade forward neural networks were used in this study. There are two types of multilayer perceptron neural network. Two training algorithms including LevenbergMarquardt (LM) and Bayesian regulation (BR) back propagation algorithms were used for updating network weights. The PSO is a simple, powerful and metaheuristic technique that can be applied to solve optimization problems. In the PSO model, every solution is showed as a particle that is alike to a bird flying via the space of a potential solution. In order to mathematically model the social governance of wolves when designing Grey Wolf Optimizer (GWO), assume the fittest solution as the alpha ( ). Consequently, the second and third best solutions are named beta ( ) and delta ( ), respectively. In the present study, the application of Artificial Neural Network (ANN), particle swarm optimization (PSO) and grey wolf optimizer (GWO) for predicting the and was investigated. Based on several statistical operates [such as coefficient of correlation ( ) and mean-square error ( ), mean absolute error ( )], for predicting and was found that the GWO ( =0.9915, =0.9986, Respecively) performs better than the PSO (with =0.9927, =0.9890) and ANN (with = 0.9618, =0.9773) models. Drying behavior of eggplant slices at different air temperatures of 45, 60, and 75C, air velocities of 1, 1.5, and 2 m/s and belt linear speeds of 2.5, 6.5, and 10.5 mm/s was studied. The moisture ratio was reduced exponentially with drying time as expected. When the temperature was increased, the drying time eggplant fruit reduced. In other words, at high temperatures, the transfer of heat and mass was higher and the water loss was more excessive. Effective moisture diffusivity and specific energy consumption were calculated after drying of turnip fruit. Maximum values of for eggplant were 1.14×10-8 m2/s. The lowest amount specific energy consumption ( ) was calculated at the boundary of 130.62 MJ/kg.