摘要
为提高草莓果浆品质,本试验研究了复合酶添加量、超声功率和酶解时间对草莓果浆可溶性固形物含量(SSC)和花色苷含量的影响,通过中心组合试验设计,利用响应面模型(RSM)和多层感知神经网络模型(MLPANN)对草莓果浆超声酶解工艺参数进行优化,并比较2种模型的预测效果。结果表明,复合酶添加量和酶解时间对草莓果浆SSC影响显著,酶解时间和超声功率对草莓果浆花色苷含量影响显著; RSM和MLP-ANN模型对草莓果浆SSC和花色苷含量预测的RMSE分别为0.269 8、0.675 8和0.038 7、0.007 7,决定系数(R2)分别为0.937 4、0.928 2和0.984 7、0.999 4,MLP-ANN模型的预测能力优于RSM模型;采用MLP-ANN模型优化得到的超声酶解法制备草莓果浆的最佳工艺条件为:复合酶添加量0. 06%,超声功率180 W和超声时间30 min,在此条件下,草莓果浆的SSC为10. 2%,每100 g草莓果浆花色苷含量为13. 46 mg。与未经超声酶解的果浆相比,SSC提高1. 12倍,花色苷含量提高2. 35倍,故超声酶解法可有效提高草莓果浆品质。
To improve the quality of strawberry pulp,the effects of enzyme dosage,ultrasonic power and enzymolysis time on the soluble solids content(SSC)and anthocyanin content were studied.Response surface methodology(RSM)and multilayer perception neural network(MLP-ANN)models were used to optimize the enzymatic hydrolysis parameters of strawberry pulp.The optimizing effects of RSM and MLP-ANN were also compared.The results showed that the addition of complex enzyme and the time of enzymolysis had a significant effect on the SSC of strawberry pulp.The enzymolysis time and ultrasonic power had significant effects on the anthocyanin content of strawberry pulp.The RMSE values of strawberry SSC and anthocyanin content predicted by RSM and MLP-ANN models were 0.269 8,0.675 8 and 0.038 7,0.007 7,respectively.The determination coefficient(R 2)was 0.937 4,0.928 2 and 0.984 7,0.999 4,respectively.The prediction ability of the MLP-ANN model was better than that of the RSM model.The optimum technological conditions for preparing strawberry pulp by using MLP-ANN model were as follows:the addition of complex enzyme was 0.06%,the ultrasonic power was 180 W,and the ultrasonic time was 30 minutes.Under these conditions,the SSC of strawberry pulp was 10.2%,and the anthocyanin content per 100 g of strawberry pulp was 13.46 mg.Compared with the pulp without ultrasound digestion,the SSC increased by 1.12 times and the anthocyanin content increased by 2.35 times.Therefore,the quality of strawberry pulp can be effectively improved by the ultrasonic enzymatic method.
作者
丛嘉昕
宋江峰
李大婧
刘春泉
包怡红
CONG Jia-xin;SONG Jiang-feng;LI Da-jing;LIU Chun-quan;BAO Yi-hong(College of Forestry,Northeast Forestry University,Harbin 150036,China;Institute of Farm Product Processing,Jiangsu Academy of Agricultural Sciences,Nanjing 210014,China)
出处
《江苏农业学报》
CSCD
北大核心
2018年第6期1354-1361,共8页
Jiangsu Journal of Agricultural Sciences
基金
农业部公益性行业(农业)科研专项(201503142-05)
关键词
草莓
超声酶解
响应面
神经网络模型
strawberry
ultrasonic enzymolysis
response surface methodology
neural network model