【作者】 张涛洪；

【导师】 徐圣言；

【作者基本信息】 江苏大学 ， 农产品加工工程， 2002， 硕士

【摘要】 冷冻食品是指经过冷冻加工的食品。冷冻食品工业虽然历史很短,发展却异常迅猛,并且将会有更大的发展空间。冷冻食品工业的发展需要高质量、高品位的冷冻食品。冷冻食品的质量主要取决于冷冻加工过程中冰晶的生成与成长的情况。因此,研究食品的冷冻,确定食品在复杂条件下的冷冻过程,对于优化冷冻食品加工工艺、改进冷冻加工设备、促进冷冻食品工业的进一步发展具有无法取代的重要意义。食品在冷冻过程中,与环境发生复杂的热交换。热交换的速度与食品的导热系数、温度扩散系数、形状、散热面积、食品与介质之间的温差以及介质的性质、流动速度等诸多因素有关。按照速率的不同,冷冻食品的降温过程可以分为三个阶段:初始、冻结、及终了阶段。国内外很早就有关于冷冻过程计算的模型,但直到目前为止,这些模型对冷冻过程的计算要么过于简单、粗略,要么太经验化,要么不具备实时应变的功能。为了得到一种准确、简便、而又具有实时性与一定通用性的计算食品冷冻过程的模型,作者提出了“在ANN下基于MATLAB的有限差分法”的动态模拟技术,关于食品的动态模拟技术主要包括如下三点内容:选择基于MATLAB的有限差分模型进行数值计算,包括建立数学模型、用有限差分法进行离散、修正冻结阶段食品物理性质参量的取值、及在MATLAB中编程运算等;设计实验验证数值模型的精确性,包括在现行条件下设计实验、及比较分析实测值等;在人工神经网络(ANN)技术支持下,实现食品冷冻过程的动态模拟,包括在EXCEL中作正交数据组、用Neroshell 2选择最好性能的Ward网络、及分析ANN的检验结果等。实验结果证实了本研究所建立的数值模型准确的反映了食品的冷冻规律;ANN软件Neroshell 2对产品数据的检验结果表明本文研究的ANN技术完全适用于实时预测在复杂多变的冷冻加工条件下的食品的冷冻加工过程。最后展望了食品冷冻过程的动态模拟技术的发展前途,并指出了本文研究所存在的问题。更多还原

【Abstract】 Frozen food is the food processed with freezing. The industry of frozen food develops rapidly, although its history is not long, and it has a promising future. To the development of frozen food industry, frozen food with high quality is necessary. The quality of frozen food mainly depends on the case in which ice crystal forms and grows up during the freezing process. Therefore, for optimizing the freezing process of food, improving the freezing equipment and promoting the further development of frozen food industry, it is significant to study the food’s freezing and determine the freezing process of food under the complex conditions.During the process of freezing, there is complex heat transfer between food and environment. The speed of heat transfer relies on thermal conductivity, diffusion coefficient of temperature, shape and area of dissipation of food, temperature difference between the medium and food, the property and flow velocity of the medium, and so on. By speed, the temperature drop process of frozen food can be categorized three phases: initial phase, freezing phase and final phase. Models of freezing process established early are simple, superficial, or experiential, or unchangeable under a range of conditions.To make a correct, simple and convenient, on-line and versatile model of freezing process of food, a dynamic simulation, finite difference method with MATLAB based on ANN, is proposed. The simulation includes three parts: choosing a finite difference model to numerical calculate, design experiments to verify the model, and carry out the dynamic simulation of the food freezing with artificial neural network (ANN).Experiments confirm that the model correctly reflects the freezing rule of food. And the result of the ANN software, Neroshell 2, verifying the data shows that the ANN studied in this paper is applicable to predict the process of food freezing under a range of complex conditions on-line.At last, the outlook of the dynamic simulation of food freezing is prospected and the defect in this paper is presented.更多还原