Congli Mei

Identification of solid state fermentation degree with FT-NIR spectroscopy: Comparison of wavelength variable selection methods of CARS and SCARS.

The use of wavelength variable selection before partial least squares discriminant analysis (PLS-DA) for qualitative identification of solid state fermentation degree by FT-NIR spectroscopy technique was investigated in this study. Two wavelength variable selection methods including competitive adaptive reweighted sampling (CARS) and stability competitive adaptive reweighted sampling (SCARS) were employed to select the important wavelengths. PLS-DA was applied to calibrate identified model using selected wavelength variables by CARS and SCARS for identification of solid state fermentation degree. Experimental results showed that the number of selected wavelength variables by CARS and SCARS were 58 and 47, respectively, from the 1557 original wavelength variables. Compared with the results of full-spectrum PLS-DA, the two wavelength variable selection methods both could enhance the performance of identified models. Meanwhile, compared with CARS-PLS-DA model, the SCARS-PLS-DA model achieved better results with the identification rate of 91.43% in the validation process. The overall results sufficiently demonstrate the PLS-DA model constructed using selected wavelength variables by a proper wavelength variable method can be more accurate identification of solid state fermentation degree.

Qualitative and quantitative analysis in solid-state fermentation of protein feed by FT-NIR spectroscopy integrated with multivariate data analysis

The potential of Fourier-transform near-infrared (FT-NIR) spectroscopy for qualitative and quantitative analysis in solid-state fermentation (SSF) of protein feed was verified based on FT-NIR spectroscopy combined with multivariate data analysis. The raw spectra were processed and analyzed by multivariate analyses, which integrated the approaches of discrete wavelet transform (DWT), principal component analysis and extreme learning machine (ELM) modeling. The noise of raw spectra was filtered and latent information was extracted by DWT, and then the characteristic information obtained by DWT was visualized in principal component space, in which the structures with the time course of the SSF were explored. Thereafter, some parameters of the calibration models were optimized by cross-validation. The results of the final models were achieved as follows: root mean square error of prediction (RMSEP) = 0.0987/Rp2 = 0.9322 for pH model, RMSEP = 0.0092 w/w/Rp2 = 0.8991 for moisture content model, and an identification rate of 91.43% for the discrimination model of the fermentation phase in the validation set. Finally, compared with partial least squares (PLS)/PLS-discriminant analysis and back propagation artificial neural network models, the ELM model showed excellent performance for prediction and generalization. This study demonstrates that FT-NIR spectroscopy coupled with appropriate chemometrics approaches could be utilized to monitor the SSF, and ELM reveals its superiority in model calibration.

Classification of Chinese Soybean Paste by Fourier Transform Near-Infrared (FT-NIR) Spectroscopy and Different Supervised Pattern Recognition

Fourier transformation near-infrared (FT-NIR) spectroscopy as an analytical tool combined with supervised pattern recognition was attempted to classify four different brands of Chinese soybean paste in this work. Three supervised pattern recognition methods, which were K-nearest neighbors (KNN), error back-propagation neural network (BP-NN), and support vector machines (SVM), were used to develop the identification models based on principal component analysis. Some parameters of the algorithms and also the number of principal components (PCs) were optimized by cross-validation in developing models. The performances of three identification models were compared. Experimental results showed that the performance of SVM model was superior to KNN and BP-NN models. The optimal SVM model achieved when the 5 PCs were included, and the identification rates both were 100% in the training and validation sets. This work demonstrated that FT-NIR spectroscopy technique coupled with SVM algorithm could be successfully used to discriminate different brands of Chinese soybean paste.

Recent advances in electronic nose techniques for monitoring of fermentation process

AbstractMicrobial fermentation process is often sensitive to even slight changes of conditions that may result in unacceptable end-product quality. Thus, the monitoring of the process is critical for discovering unfavorable deviations as early as possible and taking the appropriate measures. However, the use of traditional analytical techniques is often time-consuming and labor-intensive. In this sense, the most effective way of developing rapid, accurate and relatively economical method for quality assurance in microbial fermentation process is the use of novel chemical sensor systems. Electronic nose techniques have particular advantages in non-invasive monitoring of microbial fermentation process. Therefore, in this review, we present an overview of the most important contributions dealing with the quality control in microbial fermentation process using the electronic nose techniques. After a brief description of the fundamentals of the sensor techniques, some examples of potential applications of electronic nose techniques monitoring are provided, including the implementation of control strategies and the combination with other monitoring tools (i.e. sensor fusion). Finally, on the basis of the review, the electronic nose techniques are critically commented, and its strengths and weaknesses being highlighted. In addition, on the basis of the observed trends, we also propose the technical challenges and future outlook for the electronic nose techniques.

Monitoring the wheat straw fermentation process using an electronic nose with pattern recognition methods

To monitor the wheat straw solid-state fermentation process in real time, an electronic nose (e-nose) was attempted in this study. The e-nose was designed to detect gas changes in the fermentation process and was equipped with a sensor array composed of eleven selected commercially available metal oxide semiconductor (MOS) gas sensors. Using the e-nose data, an appropriate monitoring model can be constructed to determine process states. Therefore, selecting the optimal pattern recognition method was crucial. For the simplicity of monitoring models, principal component analysis was used to extract features (i.e. principal components or latent variables) of the e-nose data as inputs of monitoring models. For comparison, three representative methods (i.e. Gaussian process, support vector machine and back propagation neural networks) were assessed. The results sufficiently demonstrated excellent promise for the e-nose technique and the Gaussian process performed better than the other two pattern recognition methods.

A review of learning algorithm for radius basis function neural network

In this paper, Radius Basis Function Neural Networks basic learning algorithms are reviewed from the aspects of convergence, training speed, network structure, generalization, etc. Advantages and disadvantages of each learning algorithm are pointed out. And the prospect of dynamic neural network is considered.

Quantitative analysis of yeast growth process based on FT-NIR spectroscopy integrated with Gaussian mixture regression

To improve the yield of industrial fermentation, herein, we report a method based on Fourier-transform near-infrared spectroscopy (FT-NIR) to predict the growth of yeast. First, the spectra were obtained using an FT-NIR spectrometer during the process of yeast cultivation. Each spectrum was acquired over the range from 10 000 to 4000 cm−1, which resulted in spectra with 1557 variables. Moreover, the optical density (OD) value of each fermentation sample was determined via photoelectric turbidity method. Then, using a method based on competitive adaptive reweighted sampling (CARS), characteristic wavelength variables were selected from the preprocessed spectral data. Gaussian mixture regression (GMR) algorithm was employed to develop the prediction model for the determination of OD. The results of the model based on GMR were achieved as follows: only 13 characteristic wavelength variables were selected by CRAS, the coefficient of determination Rp2 was 0.98842, and the root mean square error of prediction (RMSEP) was 0.07262 in the validation set. Finally, compared to kernel partial least squares regression (KPLS), support vector machine (SVM), and extreme learning machine (ELM) models, GMR model showed excellent performance for prediction and generalization. This study demonstrated that FT-NIR spectroscopy analysis technology integrated with appropriate chemometric approaches could be utilized to monitor the growth process of yeast, and GMR revealed its superiority in model calibration.

Improved particle swarm optimization algorithm and its global convergence analysis

This paper proposed an novel improved particle swarm optimizer (PSO) algorithm with global convergence performance. The global optimum position is unpredictable, so a random solution is introduced to the improved PSO as the best solution(Pg) in the end of every generation. The novel search strategy enables the improved PSO to make use of the uncertain information, in addition to experience, to achieve better quality solutions. Theoretical proof shows the novel random search strategy enables the improved PSO to own the performance of global convergence. Five of well-known benchmarks used in evolutionary optimization methods are used to evaluate the performance of the improved PSO. From experiments, we observe that the improved PSO significantly improves the PSOs performance and performs better than the basic PSO and other recent variants of PSO.

A review of decoupling control based on multiple models

In this article, we review and discuss methods for decoupling control which include conventional decoupling and adaptive decoupling. Decoupling methods which is used to solve the multivariable coupled in nonlinear control are always a hot issue. The major algorithms such as neural network and fuzzy control are described in details and critically reviewed in this work. The advantages and disadvantages of these algorithms are analyzed. The review reveals the tremendous prospect of decoupling algorithms in nonlinear control. It is important to seek effective and simple decoupling methods.

Rapid identification of fermentation stages of bioethanol solid-state fermentation (SSF) using FT-NIR spectroscopy: comparisons of linear and non-linear algorithms for multiple classification issues

Solid-state fermentation (SSF) is a critical step in bioethanol production, and a means for the effective monitoring of the SSF process is urgently needed due to the rapid changes in the SSF industry, which demands fast tools that could provide real time information to ensure the quality of the final product. The aim of the present study was to investigate the FT-NIR spectroscopy technique associated with supervised pattern recognition methods in order to develop a means to monitor the time-related molecular changes that occur during the SSF of bioethanol. Principal component analysis as an exploratory tool was employed to uncover details on the molecular modifications of the spectral data during the SSF process. Furthermore, identification models were constructed using partial least squares discriminant analysis (PLS-DA), back propagation neural network (BPNN), support vector machine (SVM), and extreme learning machine (ELM) algorithms. The parameters of the four algorithms were optimized by leave-one-out cross-validation (LOOCV) for the calibration of the identification models. The experimental results showed that the nonlinear identification models achieved strong classification performance to identify the fermentation stages in the SSF of bioethanol. Moreover, compared with the BPNN and SVM models, the ELM model achieved a slightly better generalization performance with an identification rate of 92.60% in the validation process. The overall results show that the ELM-FT-NIR methodology was efficient in accurately identifying the fermentation stages during the SSF of bioethanol, thus demonstrating its potential for application in the in situ monitoring and control of large-scale industrial processes.