食品比热容的BP神经网络预测  

Using BP Neural Networks to Predict Specific Heat Capacity of Food

作　　者：张雷杰[1] 张敏[1] 杨乐[1] 钟志友[1] 

机构地区：[1]上海海洋大学食品学院,上海201306

出　　处：《安徽农业科学》2009年第17期8296-8297,共2页Journal of Anhui Agricultural Sciences

基　　金：国家自然科学基金资助项目(30771245);上海市2007年度教委重点建设课程项目(6700308)

摘　　要：引入BP人工神经网络计算模型，预测了食品比热容随各种物性参数变化的非线性规律。该模型以食品中的水、蛋白质、碳水化合物及脂肪含量为输入量，以其比热客为输出量，研究的食品包括淀粉、橙汁、牛肉等近20种，比热容变化范围覆盖较全面[1．5～4．0kJ/（kg·K）]。结果显示：该模型的最大相对误差为8．40％，最大绝对误差为0．2868 kJ／（kg·K）。因此，利用该模型能够较准确地预测出食品的比热容。The study introduced BP artificial neural network computational model to predict non-linear rule of specific heat capacity of food with the change of all physical property parameters. Contents of water, protein, carbohydrate and fat of foed were set as the input and the specific heat capacity was set as the output. The food contained starch, orange juice, beef etc, and the range of specific heat capacity was relatively comprehensive [ 1.5 - 4.0 kJ/（ kg· K） ] ： The result showed that the maximum relative error of this model was 8.40%, the maximtan absolute error was 0. 286 8 kJ/（ kg· K）. So the ANN model can be used to predict the specific heat capacity of food accurately.

关 键 词：神经网络 比热容 食品 预测 

分 类 号：TS201[轻工技术与工程—食品科学]