摘要
在木耳酸乳工艺参数设计中,使用自适应调节学习步长修正BP神经网络,模拟参数试验,用计算机取代全面试验和寻找最佳工艺参数的方法,并建立数学模型。其最佳工艺参数为:木耳浸提液用量为7%,蔗糖、琼脂添加量分别为6%和0.4%,接种量为5%,发酵温度为42℃,发酵时间7h。通过对比实验,最终优选出工艺参数所得出的结果符合实际实验情况。
Using the adaptively adjusting step-size of learning modify BP ANN can obtain an optimum yoghurt formula instead of overall experiment, and the math model was build. The optimum technological parameters were determined as follows: Jew's-ear juice 7%, sugar 6%, agar 0.4%, lactic acid bacterium 5%, ferment at 42℃ (7 h). The differences between the new method and the traditional one were not obvious.
出处
《乳业科学与技术》
2007年第2期59-62,共4页
JOURNAL OF DAIRY SCIENCE AND TECHNOLOGY
关键词
木耳
酸乳
工艺参数
改进型BP神经网络
jew's-ear
yoghurt
technology parameter
modified BP neural network
作者简介
金声琅.男.硕士.从事食品工艺和安全检测方面的研究。