速冻青稞鱼面储运过程中货架期预测模型构建

Construction of Prediction Model for the Shelf-Life of Quick-Frozen Highland Barley Fish Noodles

通过对不同储运温度下速冻青稞鱼面的感官、理化指标的变化分析,分别建立基于感官威布尔危害分析(Weibull Hazard Analysis, WHA)、化学动力学(Arrhenius)、BP神经网络(Back Propagation)的货架期预测模型,并验证结果。研究结果表明:在-18～25℃范围内,面条货架期随温度的升高逐渐缩短;根据国家标准确定速冻青稞鱼面在-18、-10、-5、0、5、10、25℃下货架期实际终点分别为294、195、100、30、24、18、3 d;相关性分析显示,储藏时间与TBA值呈极显著正相关,与POV值、pH值、感官评分相关性不显著。在-18、-10、-5、0、5、10、25℃温度下,基于感官评价指标的威布尔危害分析的货架期预测终点为296.21、206. 89、101. 94、29.92、23. 46、19. 15、3. 29 d,相对误差为0.27%～9. 67%;基于TBA指标的Arrhenius方程,货架期预测终点为413. 30、170. 87、101. 05、60. 92、37. 40、23. 36、6. 26 d,相对误差为1. 05%～108. 62%。基于感官评价和理化指标(pH、TVB-N、TBA、POV)的BP神经网络模型货架期预测终点为281.45、190.07、103.24、31.46、23.20、17. 17、2.76d,相对误差为4.60%～7.96%。Weibull与BP神经网络预测模型性能较优,可用于速冻青稞鱼面的货架期终点预测。

Pursuant to the analysis of sensory and physic-chemical characteristics indexes for quick-frozen highland barley fish noodles, using Weibull Hazard Analysis( WHA), chemic reaction kinetic model of the Arrhenius equation and Back Propagation, the shelf-life prediction models were established respectively. The demonstration of models was conducted and the results were analyzed. The results showed that within the temperature of-18 ~25 ℃,the shelf-life of noodles was gradually shortening with increase of temperature. The shelf-life end points under different storage temperatures of-18、 10、 5、0、5、10 and 25 based on the State-specified Stand were 294,195,100,30,24,18,3 d. Correlation analysis showed that days of storage and TBA presented a significant positive correlation, but pH, POV, sensory evaluation were no significant relationship between the days of storage. Under different temperatures of-18、-10、 5、0、5、10、25 ℃,the predicted end points of the shelf-life by the prediction model of sensory evaluation index using WHA were 296.21、206.89、101.94、29.92、23.46、19. 15、3.29 d, respectively, and the relative error was in range of 0. 27 % ~9. 67 %. The predicted end points of the shelf-life under the evaluation index of TBA by Arrhenius model were 413. 30、170. 87、101.05、60. 92、37. 40、23. 36、6. 26 d,respectively, and the relative error was in the range of 1.05% ~108.62%. The predicted end points of the shelf-life based on sensory and pH ysic-chemical indexes(pH、TVB-N、TBA、POV) by BP were 281.45、190. 07、103. 24、31.46、23. 20、17. 17、2. 76 d, respectively, and the relative error was in range of 4. 60%-7.96%. The verification results showed that the prediction model based on WHA and BP was better. Thus, this could provide further basic data for early warning of quality and storage of quick-frozen highland barley fish noodles.