Modelling of the drying of eggplants in thin-layers

In this paper, the thin‐layer drying behaviour of eggplant slices (6 mm thick layers) in a convective‐type cyclone dryer is reported. Thin‐layer drying experiments were conducted at drying air temperatures of 55, 65 and 75 °C and dry air velocities of 1 and 1.5 ms^?1. Data on sample mass, temperature and velocity of the dry air were recorded continuously during each test. In order to estimate and select a suitable form of the drying curve, eight different semi‐theoretical and/or empirical models were fitted to the experimental data and comparisons made of their coefficients of determination as predicted by non‐linear regression analysis. The Page model best described the drying curve of eggplant, with a correlation coefficient, r = 0.9999.     

THIN-LAYER DRYING OF DATES

Thin‐layer drying rates of two date cultivars, namely, Sukkari and Sakie were experimentally determined at three drying temperatures 70, 80, and 90C. Three drying models were evaluated for the thin‐layer date data. These were the exponential model, Page equation, and the approximation of diffusion model. The Page equation fitted the data best. Two empirical expressions of Page equation for predicting moisture ratio as a function of drying time and drying temperature for the two date cultivars are presented.     

Microwave drying of mango ginger (Curcuma amada Roxb): prediction of drying kinetics by mathematical modelling and artificial neural network

Curcuma amada (Mango ginger) was dried at four different power levels ranging 315–800 W to determine the effect of microwave power on moisture content, moisture ratio, drying rate, drying time and effective diffusivity. Among the fifteen thin layer drying models considered for evaluating the drying behaviour, the semi‐empirical Midilli et al., model described the drying kinetics very well with R² > 0.999. Drying rate and effective diffusivity increased as the microwave power output increased. Activation energy was estimated by a modified Arrhenius type equation and found to be 21.6 kW kg^?1. A feed‐forward artificial neural network using back‐propagation algorithm was also employed to predict the moisture content during MW drying and found adequate to predict the drying kinetics with R² of 0.985.     

Artificial neural network and mathematical modeling comparative analysis of nonisothermal diffusion of moisture in wood

The objective of this study was to develop an optimum artificial neural network (ANN) capable of predicting the direction and magnitude of the moisture flux through wood under nonisothermal steady-state diffusion. A comparison between experimental measurements and the predicted values of three mathematical models reported in the literature and of the proposed neural network is presented and discussed. When developing the ANN model, several configurations were evaluated. The optimal ANN model was found to be a network with six neurons in one hidden layer. This well-trained network correlated the forecasted to the experimental data with low-level errors compared to previously developed models and also predicted the flux-reversal phenomenon thus confirming that ANN modeling has a much better predictive performance. It was also shown that the numbers of the training data were linked to the performance of the network during estimation. However, the powerful predictive capacity of this modeling method was still supported although a limited experimental data set was trained.     

鲍鱼热风干燥动力学及干燥过程数学模拟

研究了鲍鱼在不同热风干燥温度下的干燥动力学特点,并构建了干燥过程的数学模型。热风干燥温度选取60、65、70、75、80℃;风速恒定为1m/s。干燥方法采取间歇干燥,分两个阶段进行。利用理论模型—扩散模型,和常见经验模型—Newton模型、Henderson and Pabis模型、Logaritmic模型、Two-terms模型、Page模型及Modified Page模型,对鲍鱼干燥过程的两个阶段分别进行描述。实验结果表明:鲍鱼热风干燥只经历降速阶段,水分扩散在鲍鱼干燥的过程中起主导作用。通过对实验数据进行统计分析,得到适合鲍鱼热风干燥的模型为Page模型(第一阶段干燥)和Two-terms模型(第二阶段干燥),模型的预测值与实际值比较吻合(Page模型r2>0.999,s<1%;Two-terms模型r2>0.997,s<2%),可以用来描述鲍鱼的热风干燥过程。     

Characteristics of thin‐layer drying and rehydration of nata de coco

Nata is white gelatinous bacterial cellulose produced by Acetobacter aceti ssp. xylinum through fermentation of coconut water. Drying behaviour of nata de coco was conducted with a hot air dryer for a temperature range of 50–90 °C. The experimental data obtained were fitted into seven empirical and/or semi‐theoretical thin‐layer drying models using nonlinear regression analysis. Results showed that Logarithmic model, Wang and Singh model fitted better than Verma et al. model although all these three models were suitable for predicting the drying process of nata de coco. The rehydration capacity was lost severely when the moisture content of the samples was nearly zero, but it can almost rehydrate to its initial state when the moisture content was above 8%. The drying kinetic properties and rehydration capacity of nata de coco indicated that most of the moisture in nata de coco was free water.     

Modeling Thermal Conductivity of Iranian Flat Bread Using Artificial Neural Networks

The objective of this work was to develop Artificial Neural Network (ANN) based thermal conductivity (K) prediction model for Iranian flat breads. Experimental data needed for ANN models were obtained from a pilot-scale set-up. Breads were made from three different cultivars of wheat and were baked in an eclectic oven at three different baking temperatures (232°C, 249°C and 260°C). A data set of 205 conditions was used for developing ANN and empirical models. To model K using ANN, 16 different MLP (multilayer perceptron) configurations ranging from one to two hidden layers of neurons were investigated and their prediction performances were evaluated. The (4-3-5-1)-MLP network, that is a network having two hidden layers, with three neurons in its first hidden layer and five neurons in its second hidden layer, had the best results in predicting the thermal conductivity of flat bread. For this network, R², MRE, MAE and SE were 0.988, 0.6323, 1.66×10^{? 3}, and 8.56×10^?4, respectively. Overall, ANN models (with R² ≥ 0.95) performed superior than the empirical model (with R² = 0. 870).     

Estimation of sweet cherry antioxidant activity and anthocyanin content during ripening by artificial neural network–assisted image processing technique

This paper presents a new approach for estimating antioxidant activity and anthocyanin content at ripening stages of sweet cherry by combining image processing and artificial neural network (ANN) techniques. The system was consisted of a CCD camera, fluorescent lights, capture card and MATLAB software. Anthocyanin content and antioxidant activity were determined by pH differential and 2, 2‐diphenyl‐1‐picrylhydrazyl methods, respectively. It was found that anthocyanin content was constantly increased during ripening stages, and antioxidant activity decreased during the early stages of development but increased from stage five. Several ANN models were designed and tested. Among these networks, a two hidden layer network with 11‐6‐20‐1 architecture had the highest correlation coefficient (R = 0.965) and the lowest value of mean square error (MSE) (215.4) for modelling anthocyanin content. Similarly, a two hidden layer network with 11‐14‐9‐1 architecture had the highest correlation coefficient (R = 0.914) and the lowest value of MSE (0.070) for modelling antioxidant activity.     

Modelling of heat and moisture transport during drying black grapes

In this work, heat and moisture transport occurred during drying of black grapes in a laboratory dryer were investigated. In the experiments, the air was passed through the chamber at a variety of flow rates (0.5, 0.75, 1.0 and 1.25 m s^?1) and temperatures (40, 50, 60 and 70 °C). Thermal and moisture diffusivities were determined. The possibility of expressing the moisture removal from the grapes was searched with the fourteen one‐layer drying models selected from the literature. Among all the models, the Page model was found the best for explaining the experimental results. The effects of the drying temperature and air velocity on the constants and coefficients of Page model and diffusivities were also investigated by multiple combinations of the different equations as the linear, power, logarithmic, exponential, polynomial, inverse polynomial and Arrhenius type by non‐ linear regression analyses. Models obtained were also analysed statistically using t‐test, RMSE, MBE and χ².     

ARTIFICIAL NEURAL NETWORK MODELING OF APPLE DRYING PROCESS

Artificial neural network (ANN) modeling and several mathematical models were applied to predict the moisture ratio in an apple drying process. Four drying mathematical models were fitted to the data obtained from eight drying runs and the most accurate model was selected. Two sets of ANN modeling were also performed. In the first set, the data obtained from each pilot were modeled individually to compare the ANN predictions with the best mathematical model. In the second set of ANN modeling, the simultaneous effect of all the four input parameters including air velocity, air temperature, the thickness of apple slices and drying time was investigated. The results showed that the ANN predictions were more accurate in comparison with the best fitted mathematical model. In addition, none of the mathematical models are able to predict the effect of the four input parameters simultaneously, while the presented ANN model predicts this effect with a good precision.

PRACTICAL APPLICATIONS

Today, modeling of chemical engineering processes is widespread in the process industries. An accurate modeling results in a precise prediction of the products of a process which could be very expensive or even unsafe to evaluate by experimental experiences. Because artificial neural network modeling is more or less proved to be one of the best modelings against mathematical ones, we suggest it to be considered for industrial processes such as drying in the food industry.     

Evaluation of drying models of apple (var. McIntosh) dried in a convective dryer

Drying behaviour of apple particles was investigated in a laboratory type dryer. The effect of drying air temperature, airflow velocity, initial height of layer, particles shape and size on the dehydration characteristics of apples was investigated. Increase in drying air temperature and increase in the airflow velocity caused a decrease in the drying time and an increase in drying rate. Increase in initial height of layer and increase in the sample thickness caused an increase in the drying time and decrease in drying rate. Drying time of the cubes was shorter and their drying rate was higher than for slices. The experimental dehydration data of apple particles obtained were fitted to the semi‐theoretical, empirical and theoretical models. The accuracies of the models were measured using the correlation coefficient (R), mean bias error (MBE), root mean square error (RMSE), reduced chi‐square (χ²), and t‐statistic method. All models described the drying characteristics of apple particles satisfactorily (R > 0.9792). The Logarythmic model can be considered as the most appropriate (R > 0.9976, MBE = ?10^?11?4.5 × 10^?6, RMSE = 0.00287–0.01746, χ² = 8.5 × 10^?6?3.1 × 10^?4, t‐stat = 7.3 × 10^?9?1.2 × 10^?3). The effect of drying air temperature, airflow velocity, characteristic dimension of the particle and initial height of layer on the drying models parameters were also determined.     

Thin Layer Drying of Pigeon Pea

Semi-theoretical and empirical models were examined to test their applicability to describe the thin-layer drying characteristics of whole pigeon pea. The Page and single-term diffusion equations were found simple and adequate for describing the drying behavior of the grain. The diffusion model with several terms matched the observed data closely over the entire drying period. The drying parameters of pigeon pea were a function of temperature and might be related by an Arrhenius-type equation. The drying behavior of pigeon pea consisted of two periods and the drying rate plot was similar to that of other food grains.     

Modelling of air drying of fresh and blanched sweet potato slices

Effects of blanching, drying temperatures (50–80 °C) and thickness (5, 10 and 15 mm) on drying characteristics of sweet potato slices were investigated. Lewis, Henderson and Pabis, Modified Page and Page models were tested with the drying patterns. Page and Modified Page models best described the drying curves. Moisture ratio vs. drying time profiles of the models showed high correlation coefficient (R² = 0.9864–0.9967), and low root mean squared error (RMSE = 0.0018–0.0130) and chi‐squared (χ² = 3.446 × 10^–6–1.03 × 10^–2). Drying of sweet potato was predominantly in the falling rate period. The temperature dependence of the diffusion coefficient (D_eff) was described by Arrhenius relationship. D_eff increased with increasing thickness and air temperature. D_eff of fresh and blanched sweet potato slices varied between 6.36 × 10^–11–1.78 × 10^–9 and 1.25 × 10^–10–9.75 × 10^–9 m² s^–1, respectively. Activation energy for moisture diffusion of the slices ranged between 11.1 and 30.4 kJ mol^–1.     

Far-infrared drying characteristics and quality assessment of Ligularia fischeri

The purpose of this study was to investigate drying models and drying characteristics of Ligularia fischeri by using far-infrared drying. The far-infrared drying tests on L. fischeri were performed at air velocities of 0.6 and 0.8 m/s, and drying chamber temperatures of 40, 45, and 50°C. Four thin layer drying models were used to estimate drying curves. Drying characteristics were analyzed on the basis of drying rate, color, antioxidant activity, and contents of polyphenolics and flavonoids. The goodness of the models was estimated using the coefficient of determination, the root mean square error, and the reduced chi-square. The results revealed that increases in drying temperature and air velocity caused a decrease in drying time. The Page and Thompson models were considered suitable for the far-infrared drying of L. fischeri. After drying, the antioxidant properties of L. fischeri were decreased under all drying conditions.

     

Weight loss of frozen bread dough under isothermal and fluctuating temperature storage conditions

Evaporative weight loss from food leads to both loss of saleable weight and quality deterioration so it need to be minimized. The effect of isothermal and fluctuating conditions on frozen dough weight loss was measured and compared with kinetic, physical and artificial neural network (ANN) models. Frozen dough samples were regularly weighed during storage for up to 112 days in loose-fitting plastic bags. The storage temperatures were in the range of −8 to −25 °C with fluctuations of ±0.1 °C (isothermal), ±1, ±3 or ±5 °C about the mean. For each combination of temperature and fluctuation amplitude, the rate of dough weight loss was constant. The rate of weight loss at constant temperature was nearly proportional to water vapour pressure consistent with standard theories for evaporative weight loss from packaged foods but was also accurately fitted by Arrhenius kinetics. Weight loss increased with amplitude of temperature fluctuations. The increase could not be fully explained by either the physic model based on water vapour pressure differences or the kinetic model alone. An ANN model with six neurons in the input layer, six neurons in hidden layers and one neuron in the output layer, achieved a good fit between experimental and predicted data for all trials. However, the ANN model may not be accurate for product, packaging and storage systems different to that studied.     

Neural network prediction of physical property changes of dried carrot as a function of fractal dimension and moisture content

The relationship between microstructural and physical properties of dried foods is difficult to quantify. This study uses artificial neural network analysis (ANN) to predict shrinkage and rehydration of dried carrots, based on inputs of moisture content and normalized fractal dimension analysis (ΔD/D₀) of the cell wall structure. Several drying techniques were used including conventional hot air (HAD), low pressure superheated steam (LPSSD), and freeze drying (FD). Dried carrot sections were examined by light microscopy and the fractal dimension (D) determined using a box counting technique. Optimized ANN models were developed for HAD, LPSSD, HAD + LPSSD, and HAD + LPSSD + FD, based on 1–10 hidden layers and neurons per hidden layer. ANN models were then tested against an independent dataset. Measured values of shrinkage and rehydration were predicted with an R² > 0.95 in all cases.     

油菜籽微波预处理的干燥动力学及模型研究

为了探明油菜籽微波预处理过程中的水分变化情况,建立油菜籽微波预处理干燥模型,对油菜籽在料层厚度12 mm、不同微波功率（1.0、1.5、2.0 kW）以及微波功率1.5 kW、不同料层厚度（6、12、18 mm）下预处理后的含水率、水分比和失水速率的变化情况进行了研究,并以常用的3种干燥模型指数模型、单项扩散模型和Page模型进行了数据拟合。结果表明:微波功率越高、料层越薄,油菜籽水分流失越快,微波预处理时间越短;微波预处理过程中油菜籽水分变化情况与Page模型拟合度最好。     

MODELING DEAD-END ULTRAFILTRATION OF APPLE JUICE USING ARTIFICIAL NEURAL NETWORK

An artificial neural network (ANN) was developed to model the dead-end ultrafiltration process of apple juice. Molecular weight cutoff, transmembrane pressure, gelatin–bentonite concentration and time were the input variables, while filtrate flux and filtrate volume were the output variables of the ultrafiltration process. According to error results and correlation values for two types of network (one or two hidden layer configurations), configurations with two hidden layers had comparatively better performance. The highest correlation coefficient with the minimum prediction error was calculated for two hidden layers with 6-5 nodes configuration. Trained ANN (4-6-5-2) predicted filtrate flux and filtrate volume with 2.33 and 1.38% mean relative error, respectively. The results suggest that the ANN modeling can be effectively used to optimize filtration process.

PRACTICAL APPLICATION

Membrane separation processes including ultrafiltration have gained importance in the food industry. Today, fruit juices are widely clarified by means of ultrafiltration process instead of tedious and laborious conventional clarification treatments. Membrane fouling which results in flux decline is the main problem associated with the ultrafiltration of fruit juices. In order to perform an efficient ultrafiltration process, optimization is required to obtain maximum filtrate volume per unit time. Artificial neural network (ANN) modeling offers great advantage on improving the performance of ultrafiltration process by accounting the effects of different variables, i.e., feed properties, transmembrane pressure and membrane pore size on filtrate volume as the main output of the filtration process. ANN modeling of ultrafiltration may be an alternative to previously proposed empirical and semiempirical models.     

CHARACTERISTIC DRYING CURVE and MATHEMATICAL MODELING of THIN-LAYER SOLAR DRYING of PRICKLY PEAR CLADODE (OPUNTIA FICUS INDICA)

Thin‐layer solar drying experiments were conducted for the prickly pear cladode grown in Marrakech, Morocco. the experimental drying curves obtained show only a falling rate period. the results verified, with good reproducibility, that the drying air temperature is the main factor in controlling the drying rate. the expression of the drying rate equation was determined empirically from the characteristic drying curve. Eight different drying models were compared according to their correlation coefficient (r²) to estimate solar drying curves. the Page model could satisfactorily describe the solar drying curves of cladode with an r² of 0.9995. the coefficient of this model could be explained by the effect of drying air temperature with an r² of 1.0000.     

PREDICTION OF MECHANICAL PROPERTIES OF CUMIN SEED USING ARTIFICIAL NEURAL NETWORKS

In this paper, two artificial neural networks (ANNs) are applied to acquire the relationship between the mechanical properties and moisture content of cumin seed, using the data of quasi-static loading test. In establishing these relationship, the moisture content, seed size, loading rate and seed orientation were taken as the inputs of both models. The force and energy required for fracturing of cumin seed, under quasi-static loading were taken as the outputs of two models. The activation function in the output layer of models obeyed a linear output, whereas the activation function in the hidden layers were in the form of a sigmoid function. Adjusting ANN parameters such as learning rate and number of neurons and hidden layers affected the accuracy of force and energy prediction. Comparison of the predicted and experimented data showed that the ANN models used to predict the relationships of mechanical properties of cumin seed have a good learning precision and good generalization, because the root mean square errors of the predicated data by ANNs were rather low (4.6 and 7.7% for the force and energy, respectively).

PRACTICAL APPLICATIONS

Cumin seed is generally used as a food additive in the form of powder for imparting flavor to different food preparations and for a variety of medicinal properties. Physical properties of cumin seeds are essential for the design of equipment for handling, harvesting, aeration, drying, storing, grinding and processing. For powder preparation especially the fracture behavior of the seeds are essential. These properties are affected by numerous factors such as size, form and moisture content of the grain and deformation speed. A neural network model was developed that can be used to predict the relationships of mechanical properties. Artificial neural network models are powerful empirical models approach, which can be compared with mathematical models.