摘要
主要研究近红外光谱技术对黄酒发酵过程中总酸和酒精度快速检测的可行性。针对黄酒发酵过程中关键参数检测费时费力的现象,提出基于联合区间偏最小二乘结合遗传算法(si PLS-GA)的波长选择方法,由交互验证法确定最佳主成分因子数及筛选的变量数。结果表明,使用5、8个主成分的si PLS-GA算法在总酸和酒精度的建模中取得了最好的预测效果。最好的模型预测精度如下:总酸和酒精度的预测集相关系数分别为0.9177、0.9803,预测均方根误差(RMSEP)分别为0.2132、0.7816。经过对比发现,近红外光谱技术结合si PLS-GA算法能够很好的对黄酒发酵过程中的总酸和酒精度的含量变化进行检测,且效果较好。
The feasibility of NIRS was investigated to rapidly detect the total acid and alcohol in the fermentation process of Chinese rice wine. The detection of key parameters is of importance in the fermentation process of Chinese rice wine,however,it is largely time- consuming and laborious.In this paper,synergy interval partial least squares combined with the genetic algorithm（ si PLS- GA） was put forward to estimate fermentation process parameters in Chinese rice wine,and the cross- validation method was used to select the optimal PLS factors and the variables.The results showed that the best prediction result was achieved by si PLS- GA using 5 and 8 factors in modeling the total acid and alcohol.By employing the best model,the prediction correlation coefficients（ Rp） of the total acid and alcohol were 0.9177 and 0.9803,respectively,and the root mean square errors for prediction（ RMSEP） were 0.2132 and 0.7816 respectively.Comparing with the results showed that NIRS combined with si PLSGA can achieve good performance in detection of the content of total acid and alcohol in the fermentation process of Chinese Rice wine.
出处
《食品工业科技》
CAS
CSCD
北大核心
2015年第17期290-294,299,共6页
Science and Technology of Food Industry
基金
国家自然科学基金(61134007)
