Study On Near-infrared Spectroscopy Detection Method For Parameters And Status Of Ethanol Solid-state Fermentation

In order to realize the real-time detection of ethanol solid-state fermentation process,the quantitative detection of the parameters and characterization of ethanol solid-state fermentation process based on near-infrared spectroscopy were carried out.Specific research work is as follows:(1)A quantitative method for the determination of key parameters(ethanol and glucose)in the solid-state fermentation of ethanol based on near infrared spectroscopy was discussed.The contents of ethanol and glucose in the samples were obtained by physical and chemical analysis,and the near infrared spectra of the samples were collected.The experimental data were provided for the establishment of parameters and state detection models of solid-state fermentation subsequently.(2)Quantitative detection method of key parameters(ethanol and glucose)in ethanol solid-state fermentation process based on near-infrared spectroscopy was discussed.Synergy interval partial least squares algorithm was used to select the optimal joint subintervals about the ethanol and glucose from the spectra that were pretreated by standard normal variate transformation algorithm.The characteristic wavenumber variables of the ethanol and glucose were selected from the optimal joint subintervals by iteratively retaining informative variables algorithm and compared with the traditional method of genetic algorithms and competitive adaptive reweighted sampling.Last,partial least squares prediction models for ethanol and glucose content were established.The results show that the numbers of characteristic variables extracted by iteratively retaining informative variables method about ethanol and glucose are 45 and 43 respectively.In the results of partial least squares model in the validation set established by the characteristic variables,the root mean square error prediction and correlation coefficient about the ethanol content are 0.2485 and 0.9937 respectively,and about the ethanol are 0.1418 and 0.9949 respectively.The results show that the near-infrared spectroscopy technique could effectively and quickly detect the key parameters of ethanol solid-state fermentation process.(3)The qualitative detection method of solid-state fermentation process based on near-infrared spectroscopy was discussed.The optimal joint subintervals of the fermentation process were selected from the standard normal variate transformation algorithm preprocessing spectra by synergy interval partial least squares method.The characteristic wavenumber variables of the fermentation process were selected from the optimal joint subintervals by iteratively retaining informative variables method,competitive adaptive reweighted sampling method and genetic algorithms respectively.The principal component analysis model and extreme learning machine model of the fermentation process were established respectively.The results show that the cumulative contribution rate of the principal component analysis model of the first two principal components is 96.0236%.The model was established by characteristic wavenumber variables,which were selected by iteratively retaining informative variables method.The cumulative contribution rate is much higher than other principal component analysis models.In the results of extreme learning machine state recognition model established by the characteristic wavenumber variables which were selected by iteratively retaining informative variables method,the accurate rates in calibration and validation sets are 99.8182% and 97.2728% respectively.The accurate rates are much higher than the rates of other models.The results show that the near-infrared spectroscopy combined with chemometrics methods can effectively and quickly detect the state of solid-state fermentation of ethanol.This study provides a new idea for near-infrared spectroscopy on-line detection of ethanol solid-state fermentation process,laying a theoretical and technical basis for the development of portable near-infrared spectroscopy equipment for on-line detection of ethanol solid-state fermentation process.