A Soft Sensor Modelling of Biomass Concentration during Fermentation using Accurate Incremental Online v-Support Vector Regression Learning Algorithm

Abstract

In order to model real fermentation process, a soft sensor modelling of biomass concentration during fermentation using accurate incremental online ν-Support Vector Regression (ν-SVR) learning algorithm was proposed. Firstly, an accurate incremental online ν-SVR learning algorithm was proposed. This algorithm solved the two complications introduced in the dual problem based on the equivalent formulation of ν-SVR. Moreover, it addressed the infeasible updating path problem during the adiabatic incremental process by relaxed adiabatic incremental adjustments and accurate incremental adjustments. Then, the proposed algorithm is used to predict the biomass concentration of glutamic acid fed-batch fermentation process online. The results of simulation experiment showed that the soft sensor modelling of biomass concentration during fermentation using the proposed algorithm was of better generalization ability and cost less training time than that of ν-SVR.