外加电压对明胶膜液荷电与润湿性能的影响及其预测模型

Effect of Applied Voltage on the Charging and Wettability Properties of Gelatin Film-forming Solution and Its Predictive Model

中文关键词: 明胶可食用涂膜  荷质比  表面张力  接触角  深度神经网络

英文关键词: gelatin edible coating  charge-to-mass ratio  surface tension  contact angle  deep neural networks

中文摘要:

      目的 探究不同荷质比明胶溶液的润湿性能，并建立预测模型。方法 以明胶可食涂膜为研究对象，利用感应荷电施加外源静电场以改善膜液润湿性能，探究电场电压对明胶液滴荷质比与表面张力，以及液滴在疏水表面接触角的影响，并通过机器学习建立荷质比与表面张力/接触角之间预测模型。结果 随着电压升高，明胶液滴荷质比不断增大，且仅以司盘20为表面活性剂(tw0组)时液滴具有最高的荷质比(−50 nC/g)。在0~7 kV内，明胶液滴的表面张力随电压升高从35.99~40.65 mN/m降至31.38~35.65 mN/m，其中tw0组表面张力下降最为明显。明胶液滴在石蜡表面的接触角也随电压升高而减小，在表面活性剂吐温20与司盘20质量比为1∶1时具有最小值，即电压7 kV时接触角为64.99°。深度神经网络预测模型决定系数接近于1，均方误差小于0.08，平均绝对误差小于0.15，具有最好的预测效果。结论 静电喷涂能够有效改善膜液在食品表面的润湿性能，利用深度神经网络能够建立膜液液滴荷质比与表面张力/接触角的良好预测模型。

英文摘要:

      The work aims to investigate the wettability properties of gelatin solutions with different charge-to-mass ratios and establish a prediction model. With the edible gelatin film-forming droplet as the research object, an external electrostatic field was applied through induction charging to improve the wettability properties and explore the effect of electric field voltage on the charge-to-mass ratio and surface tension of the gelatin droplets and the contact angle of droplets on hydrophobic surfaces. Machine learning was used to establish predictive models between the charge-to-mass ratio, surface tension, and contact angle of gelatin droplets. The charge-to-mass ratio of gelatin droplets continued to increase with the increase of voltage, and the gelatin droplets had the highest charge-to-mass ratio (−50 nC/g) when only Span20 was used as the surfactant (tw0 group). In the range of 0-7 kV, the surface tension of gelatin droplets decreased from 35.99-40.65 mN/m to 31.38-35.65 mN/m with the increasing voltage, with the tw0 group showing the most significant decrease in surface tension. The contact angle of gelatin droplets on the surface of paraffin also decreased with the increasing voltage and had a minimum value at a 1:1 ratio of surfactant Tween 20:Span20 (64.99° at a voltage of 7 kV). The decision coefficient of the deep neural network (DNN) prediction model was close to 1, the mean square error was less than 0.08, and the average absolute error was less than 0.15, indicating the best prediction performance. Electrostatic spraying can effectively improve the wettability properties of the film on the surface of food and enhance the preservation effect of the film. The DNN has the best prediction effect on the relationship between the charge-to-mass ratio of gelatin droplets, surface tension, and contact angle.