Soft-sensing method based on modified ANN inversion and its application in erythromycin fermentation

摘要

In this paper, we modify the soft-sensing method based on ANN inversion. By using the non-state variables or the so-called function variables besides the state variables, the possibility in constructing the soft-sensing model will increase and the derivative order in soft-sensing model will be low. The simulation results verify that the soft-sensing method based on the modified ANN inversion is more accurate than the unmodified one.