Application of ANN-Inversion Soft-Sensing Method in Biochemical Fermentation

Based on Artificial Neural Network inversion (ANN-inversion), a novel soft-sensing method is presented to estimate some direct-unmeasurable biochemical variables, such as mycelia concentration, sugar concentration and chemical potency from other direct-measurable variables such as dissolved oxygen concentration, pH, and volume in erythromycin fermentation. The proposed ANN-inversion soft-sensor is composed of a static ANN and several differentiators. Experimental results show that the soft-sensing values are almost identical with the actual ones, or the designed ANN-inversion soft-sensor possesses good approximate ability to the direct-unmeasurable variables.