基于近红外光谱技术的乙醇固态发酵过程参数定量检测

发布时间：2018-08-28 18:19

【摘要】：为了提高乙醇固态发酵过程在线监测的精度,开展了基于傅里叶近红外光谱(FT-NIRS)分析技术的乙醇固态发酵过程参数快速定量检测研究。采用联合区间偏最小二乘法(siPLS)对标准正态变量变换(SNV)预处理后的光谱进行特征波长区间优选;引入遗传算法(GA)、竞争自适应重加权采样(CARS)法和迭代保留信息变量(IRIV)法从优选后波长区间中进一步筛选特征波长变量;最后,建立不同变量筛选方法所得特征波长的乙醇固态发酵过程参数(乙醇和还原糖含量)的偏最小二乘(PLS)预测模型。实验结果显示,与GA和CARS方法相比,IRIV方法所得的波长变量数最少;其中,与乙醇和还原糖相关的特征变量个数分别为43和40;在验证集中,PLS预测模型乙醇含量的验证集均方根误差(RMSEP)和预测相关系数Rp分别为0.2511和0.9934,还原糖含量的RMSEP和Rp分别为0.1730和0.9926,其预测精度亦高于其他方法所得结果。实验结果表明,利用近红外光谱分析技术实现乙醇固态发酵过程关键参数的在线检测是可行的;并且IRIV方法是一种有效近红外光谱特征波长优选方法,可提高预测模型精度。
[Abstract]:In order to improve the accuracy of on-line monitoring of ethanol solid-state fermentation process, a rapid quantitative analysis of ethanol solid-state fermentation process parameters based on Fourier near infrared spectroscopy (FT-NIRS) was carried out. Combined interval partial least square method (siPLS) is used to select the spectrum of standard normal variable transform (SNV) pretreated with characteristic wavelength interval. The genetic algorithm (GA),) competition adaptive rescheduled sampling (CARS) method and iterative reserved information variable (IRIV) method are introduced to further screen the characteristic wavelength variables from the optimized wavelength interval. A partial least-squares (PLS) prediction model of ethanol solid-state fermentation process parameters (ethanol and reducing sugar content) was established by different variable selection methods. The experimental results show that, compared with the GA and CARS methods, the GA method has the least number of wavelength variables. The number of characteristic variables related to ethanol and reducing sugar were 43 and 40 respectively. The root mean square error (RMSEP) and predictive correlation coefficient (Rp) of the validation set were 0.2511 and 0.9934, respectively, and the RMSEP and Rp scores of reducing sugar content were 0.2511 and 0.9934 respectively. The prediction accuracy is higher than that of other methods. The experimental results show that it is feasible to use near infrared spectroscopy to detect the key parameters of ethanol solid-state fermentation process on line, and IRIV method is an effective near infrared spectrum characteristic wavelength optimization method, which can improve the precision of prediction model.
【作者单位】：江苏大学电气信息工程学院;
【基金】：国家自然科学基金(31271875) 江苏省自然科学基金(BK20140538,BK20130531,BK20151345) 江苏省高校自然科学研究面上项目(16KJB210003) 江苏省博士后科学基金(1601038C,1601130B)
【分类号】：TQ920.6;TQ223.122;O657.33

【相似文献】