Hybrid modeling for Nosiheptide fermentation process

A hybrid modeling scheme was developed for fermentation process, which combines the first principle equation based on mass balance and data based support vector machine model. First principle describes the rough framework of fermentation process. And support vector machine model is used to estimate the unmeasured process parameters that are difficult to model with first principle. Also, an on-line correction strategy was proposed to improve the model's practicality. Applying the proposed modeling scheme to Nosiheptide fermentation process demonstrates that the hybrid modeling has good estimation performance.