鲟鱼片冷藏过程中品质变化规律与预测模型

测定冷藏条件下鲟鱼(Acipenser sinensis)片的感官分值、菌落总数的对数值(lg菌落总数/(CFU/g)])、挥发性盐基氮的质量分数(w(TVB-N))、鲜度指标(K)等品质指标,研究其品质变化规律,建立动力学和径向基神经网络组合模型。结果表明:随着贮藏时间的延长,lg菌落总数/(CFU/g)]、w(TVB-N)、K均呈现明显的上升趋势,感官分值呈现下降趋势,各指标之间显著相关(P0.05)。动力学和径向基神经网络组合模型用于鲟鱼片冷藏过程中品质变化的预测,其预测值与试验值之间的相对误差的绝对值小于1%。组合模型充分利用了动力学模型的线性预测和径向基神经网络的非线性预测的优势,预测准确度较高。

Modeling quality changes in sturgeon (Acipenser sinensis) fillets during storage

To evaluate and predict the freshness of sturgeon (Acipenser sinensis) fillets during refrigerated storage, changes in quality sensory assessment, logarithmic value of total aerobic counts (lgTAC/(CFU/g)]), mass fraction of total volatile base nitrogen (w (TVB-N)) and K value] were investigated and corresponding model for quality prediction were established.Results showed that lgTAC/(CFU/g)], w (TVB-N) and K value increased with the extension of refrigerated storage.There were high correlations among sensory assessment, lgTAC/(CFU/g)], w (TVB-N) and K value in all samples (P<0.05).The model combined with kinetic model and RBF neural network predicted changes of sensory assessment, lgTAC/(CFU/g)], K value and w (TVB-N) with the absolute value of relative error between predicted values and experimental values all within ±1%.The model take advantages of both the Arrhenius model (for the linear part of prediction) and RBF neural network (to predict nonlinear part), which could significantly improves the accuracy of quality prediction of sturgeon fillets during refrigerated storage.