基于傅里叶变换近红外光谱实时分析1,3-丙二醇发酵过程生物量的在线监测方法

On-line monitoring of biomass in 1, 3-propanediol fermentation by Fourier-transformed near-infrared spectra analysis

摘要:

生物量是反映生物发酵过程进展的重要参数，对生物量进行实时监测可用于对发酵过程的调控优化。为克服目前主要采用的离线方法检测生物量时间滞后和人工测量误差较大等缺点，本研究针对1,3-丙二醇发酵过程设计了一个基于傅里叶变换近红外光谱实时分析技术的生物量在线监测实验平台，通过对实时采集光谱预处理以及敏感光谱段分析，应用偏最小二乘算法，建立了1,3-丙二醇发酵过程生物量变化的动态预测模型。以底物甘油浓度为60 g/L和40 g/L的发酵过程作为外部验证实验，分析得到模型的预测均方根误差分别为0.341 6和0.274 3，结果表明所建立的模型具有较好的实时预测能力，能够实现对1,3-丙二醇发酵过程中生物量的有效在线监测。

Abstract:

Biomass is an important parameter reflecting the fermentation dynamics. Real-time monitoring of biomass can be used to control and optimize a fermentation process. To overcome the deficiencies of measurement delay and manual errors from offline measurement, we designed an experimental platform for online monitoring the biomass during a 1,3-propanediol fermentation process, based on using the fourier-transformed near-infrared (FT-NIR) spectra analysis. By pre-processing the real-time sampled spectra and analyzing the sensitive spectra bands, a partial least-squares algorithm was proposed to establish a dynamic prediction model for the biomass change during a 1,3-propanediol fermentation process. The fermentation processes with substrate glycerol concentrations of 60 g/L and 40 g/L were used as the external validation experiments. The root mean square error of prediction (RMSEP) obtained by analyzing experimental data was 0.341 6 and 0.274 3, respectively. These results showed that the established model gave good prediction and could be effectively used for on-line monitoring the biomass during a 1,3-propanediol fermentation process.