当前位置: X-MOL 学术Anal. Methods › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Rapid identification of fermentation stages of bioethanol SSF using FT-NIR spectroscopy: Comparisons of linear and non-linear algorithms for multiple classification issues
Analytical Methods ( IF 2.7 ) Pub Date : 2017-09-14 00:00:00 , DOI: 10.1039/c7ay01861d
Hui Jiang 1, 2, 3, 4, 5 , Congli Mei 2, 3, 4, 5 , Quansheng Chen 2, 3, 4, 6
Affiliation  

Solid state fermentation (SSF) is a critical step in achieving bioethanol product, and an effective monitoring of SSF process is on the urgent need due to the rapid changes in the SSF industry, which demands fast tools providing real time information to ensure the quality of the final product. The aim of this study is to investigate FT-NIR spectroscopy technique associated with supervised pattern recognition methods, to monitor time-related molecular changes that occur during SSF of bioethanol. Principal component analysis as an exploratory tool was employed to uncover molecular modification 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, respectively. The parameters of the four models were optimized by leave-one-out cross-validation (LOOCV) in identification model calibration. Experimental results showed that the non-linear identification models showed strong classification performance to identify fermentation stages in SSF of bioethanol. Meanwhile, compared with BPNN and SVM models, the ELM model obtained slightly better generalization performance with the identification rate of 92.60% in the validation process. The overall results show that the ELM-FT-NIR methodology is efficient in accurately identifying the fermentation stages during the SSF of bioethanol, demonstrating potential for apply in in-situ monitoring and control of large-scale industrial processes.

中文翻译:

使用FT-NIR光谱法快速鉴定生物乙醇SSF的发酵阶段:针对多个分类问题的线性和非线性算法的比较

固态发酵(SSF)是获得生物乙醇产品的关键步骤,由于SSF行业的快速变化,迫切需要对SSF工艺进行有效监控,这需要快速的工具来提供实时信息以确保乙醇的质量。最终产品。这项研究的目的是研究与监督模式识别方法相关的FT-NIR光谱技术,以监测生物乙醇SSF期间发生的与时间有关的分子变化。主成分分析作为探索性工具被用来揭示SSF过程中光谱数据的分子修饰。此外,使用偏最小二乘判别分析(PLS-DA),反向传播神经网络(BPNN),支持向量机(SVM)和极限学习机(ELM)算法。通过识别模型校准中的留一法交叉验证(LOOCV)优化了四个模型的参数。实验结果表明,非线性识别模型具有很强的分类性能,可以识别生物乙醇中SSF的发酵阶段。同时,与BPNN和SVM模型相比,ELM模型在验证过程中获得了更好的泛化性能,识别率为92.60%。总体结果表明,ELM-FT-NIR方法能够有效地准确识别生物乙醇SSF期间的发酵阶段,证明了在大规模工业过程的现场监控中的应用潜力。通过识别模型校准中的留一法交叉验证(LOOCV)优化了四个模型的参数。实验结果表明,非线性识别模型具有很强的分类性能,可以识别生物乙醇中SSF的发酵阶段。同时,与BPNN和SVM模型相比,ELM模型在验证过程中获得了更好的泛化性能,识别率为92.60%。总体结果表明,ELM-FT-NIR方法可有效地准确识别生物乙醇SSF期间的发酵阶段,证明了在大规模工业过程的现场监控中的应用潜力。通过识别模型校准中的留一法交叉验证(LOOCV)优化了四个模型的参数。实验结果表明,非线性识别模型具有很强的分类性能,可以识别生物乙醇中SSF的发酵阶段。同时,与BPNN和SVM模型相比,ELM模型在验证过程中获得了更好的泛化性能,识别率为92.60%。总体结果表明,ELM-FT-NIR方法可有效地准确识别生物乙醇SSF期间的发酵阶段,证明了在大规模工业过程的现场监控中的应用潜力。实验结果表明,非线性识别模型具有很强的分类性能,可以识别生物乙醇中SSF的发酵阶段。同时,与BPNN和SVM模型相比,ELM模型在验证过程中获得了更好的泛化性能,识别率为92.60%。总体结果表明,ELM-FT-NIR方法可有效地准确识别生物乙醇SSF期间的发酵阶段,证明了在大规模工业过程的现场监控中的应用潜力。实验结果表明,非线性识别模型具有很强的分类性能,可以识别生物乙醇中SSF的发酵阶段。同时,与BPNN和SVM模型相比,ELM模型在验证过程中获得了更好的泛化性能,识别率为92.60%。总体结果表明,ELM-FT-NIR方法可有效地准确识别生物乙醇SSF期间的发酵阶段,证明了在大规模工业过程的现场监控中的应用潜力。ELM模型在验证过程中获得了更好的泛化性能,识别率为92.60%。总体结果表明,ELM-FT-NIR方法能够有效地准确识别生物乙醇SSF期间的发酵阶段,证明了在大规模工业过程的现场监控中的应用潜力。ELM模型在验证过程中获得了更好的泛化性能,识别率为92.60%。总体结果表明,ELM-FT-NIR方法可有效地准确识别生物乙醇SSF期间的发酵阶段,证明了在大规模工业过程的现场监控中的应用潜力。
更新日期:2017-09-14
down
wechat
bug