Study On On-line Detection Of Bacillus Subtilis Fermentation Process

Fermentation Engineering is one of the high-tech science and technology in the 21st century, As the earliest studies in the microbiological field, they has made tremendous contributions for the development of the production of human life and society in the last few hundred years. They are playing an increasingly important role in agriculture, medical industry, food industry, new energy and fine chemical about economy and people's life. With the development of science and technology and the continuous improvement of people's needs, the requirements of fermentation product quality are increasing, therefore, to achieve the optimal operation and optimized control of the fermentation process become a hot topic.To achieve the fermentation processes'optimization and control, we first need to develop new type of sensors applied to biological processes and on-line detection technology. However, there are few detection equipments which can be used for large-scale industrial production, and has advantages of simple operation and maintenance, stable performance, low price. In the industry, only pH, DO and the partial pressure of fermentation tank gas can be measured. Such as microbial biomass concentration, substrate concentration and product concentration which can accurately reflect the microbial state can't be detected online, because detection equipment are expensive and complex in operation and maintenance.The equipment can't be used in the industry.For the above problems of online testing in fermentation engineering,This paper proposes the use of an mid-infrared spectrometer to detect the concentration of the ingredients in the fermentation broth,and we do fermentation experiments to achieve spectrum data about Bacillus subtilis producting ?-polyglutamic acid.Then according to the selection rules of the calibration set to model these spectrum data. The models are not only able to predict the concentration of glutamate, glucose in fermentation broth but also predict concentration of biomass. By comparing prediction value of the model and reference value which are measured offline by other instruments, the accuracy of prediction models come to the online measurement requirements.This measurement method solves the shortcomings of offline measurement such as lag, detection of a single component and others, and also addresses some of the drawbacks of the use of advanced sensor.