Drying Characteristics And Mass Transfer Simulation Of Sigillate Walnut(Juglans Sigillata Dode) During Convection Drying

With fresh sigillate walnut(Juglans sigillata Dode) as the research object, some basic physical properties (the length, width, thickness, geometric mean diameter, sphericity, surface area, unit mass, volume, bulk density and particle density) of sigillate walnut and its kernel as a function of moisture content were determined first in this work. Then the drying behavior of sigillate walnut during convection drying was researched at three hot air temperature levels(40,50and60℃) and two air velocity levels (0.8and1.4m/s). Four different models (the Newton, Page, Henderson-Pabis and the BP neural network model) were fitted to the experimental data to select a suitable prediction form of the drying process. Finally, the mass transfer model was established to predict the moisture distribution inside sigillate walnut during drying. The conclusions as follow:There were significant differences among the individual quality of sigillate walnut and its kernel as the variable coefficientes of unit mass and volume were relatively larger than others. The average sphericity of sigillate walnut was93.06%, which indicated that the sigillate walnut can be treated as spheres and expected to roll rather than slide on the surface. Most of the studied physical properties decreased linearly with decreasing moisture, while the sphericity of sigillate walnut and its kernel almost remained constant. Similar slope was observed with the three axis dimensions and geometric mean diameter at the studied moisture contents regardless of the nut or kernel. Besides, with the moisture varying the geometric mean diameter and length was close to each other, which meant that the geometric mean diameter could be replaced by length to some extent.The constant rate period was absent during the drying processes of sigillate walnut. The falling rate stage, which included the part I and part II, was the main part of the process.The higher drying temperature caused faster drying rate in the part I of falling rate stage, and the longer the duration of this stage, which caused the whole drying time of sigillate walnut reduced. At the studied drying conditions, the drying intensity and specific power consumption (SPC) were significant affected by temperature, the higher the temperature, the greater the intensity, the lower the SPC, while the air velocity had little effect on these test index. In conclusion, the condition of sigillate walnut during convection drying was suggested:The temperature and air velocity should be higher in the part I of falling rate stage while lower in part II.The relative error between the experimental and calculated from the BP neural network model was less than5%, while they were0～53.06%,0～64.32%and0～78.06%for the Newton. Page and Henderson-Pabis model respectively. It was thus clear that the BP neural network model was more accurate for prediction of the drying process, whose predicted value was closer to the actual drying process. So the BP neural network model could be taken as an important auxiliary means to guide the actual production.The numerical model was developed to simulate the unsteady one-dimensional mass transfer during the convection drying process of sigillate walnut, which was useful for prediction the moisture distribution inside sigillate walnut as well as the design of drying process.The Deff calculated from the numerical model varied from1.14×10-9to1.73×10-9m2/s in the studied condition. Moisture content of the surface decreased much faster than that of internal at the early period, while then the situation just the reverse, after the moisture content of each location in the sample decreased steadily at later period. The moisture gradients from the surface to the centre decreased as the drying time increase. In addition, analysis suggested that the shell is one of the most main resistances of moisture transfer during the whole drying process of sigillate walnut.