On-line Estimation of Biomass Concentration and Specific Growth Rate in the Fermentation Process

The biomass concentration, the same as concentrations of substrates or products, a very important state variable of almost every bioprocess, cannot often be measured on-line, due to the lack of reliable and cheap sensors, which makes this set of variables limited for control purposes. In this paper, a type of soft sensing model for the on-line estimation of biomass concentration has been developed based on the least squares support vector machines approach. The soft sensing model was trained on data coming from nosiheptide fermentation process, and evaluated the production. The results show that this method is potentially useful for estimation biomass on-line in bioprocesses. In addition, an on-line computing model based on the biomass estimator was developed for the specific growth rate. This model consists of three sub-models. They describe the growth characteristic of different phases very well and facilitate its on-line monitoring