基于神经网络平台的牡蛎肉超高压杀菌工艺条件优化

Study on the processing optimization of ultra high pressure sterilization from oyster meat base on a neural network method

摘要: 为研究超高压对牡蛎杀菌的影响,利用Box-Behnken实验设计建立数据集,以菌落总数为参考指标,借助JMP7.0软件的神经网络平台建立神经网络模型,优化牡蛎肉超高压杀菌处理工艺条件。实验结果表明,牡蛎肉超高压杀菌的最佳工艺条件为压力350MPa,保压时间20min,处理温度30℃,在此条件下超高压处理后的牡蛎肉菌落总数从3.2718lg cfu/g减少到2.44342lg cfu/g（灭菌率85.13%）,大肠菌群从7MPN/100g减少至0。验证实验结果表明,所建立的神经网络模型具有较好的预测能力,可准确地预测杀菌效果,预测值与实验值的相对误差较小（0.201%）。

Abstract: The results showed that the best process condition was the pressure was 350 MPa for 20 min and the temperature was 30℃,the oyster meat total bacteria after pressure treatment in this condition decreased from3.2718 to 2.44342 lg cfu/g(The sterilization rate was 85.13%),and Coliforms decreased from 7 to 0(MPN/100g).Validation test results showed that the established neural network model had better prediction ability which could accurately predict the sterilization effect,and the relative error of predicted values and experimental values was small(0.201%).