稻米食味值影响的神经网络模型预测.doc

Prediction Impact of Vacuum Drying Parameters on Rice Taste Value with Neural Network Model Xu Zemin1 , Wu Wenfu2,Yin Liyan2 College of Biological Science and Technology, ChangChun University, Chang Chun, 130022, China; College of Biological and Agricultural Engineering, Jilin University, Chang Chun, 130022,china Abstract: Process control over rice vacuum drying directly affects the taste quality of dried rice. Drying temperature, initial rice moisture content and vacuum degree are the main parameters that affect rice-drying process. This paper makes research on the change law of taste quality of brown rice versus drying parameters during vacuum drying process through establishment of neural network model between brown rice taste value and drying parameters, which will provide reference for objective evaluation on rice taste, dried rice quality and the design of vacuum dryer. The results show that, BP neural network model predicts the average relative error of 2.88%, correlation coefficient of 0.96, BP neural network can serve as a new model for description of rice taste value of vacuum drying. Key words: neural network; rice; vacuum drying; taste quality; model 1 Introduction Rice is a heat sensitive material and its drying process control is extremely important. If the drying parameters are inappropriate, rice quality will be adversely affected. How to realize artificial intelligence control of vacuum drying process is the key to ensure rice taste quality [1-4]. The taste is an important evaluation index of rice quality. Taste value is an index that reflects the degree of rice taste, which can be measured with taste analyzer. The larger the index, the better the taste, with the maximum value of 100%. Taste analyzer is a taste measuring instrument developed by Japanese scholars by establishment of relationship between the taste and the physical and chemical indicators of rice content (amylose, protein, moisture and fatty acids) based on theory neuro-fuzzy [5]. The