Artificial Neural Network Based Soft Sensor for Fermentation of Recombinant Pichia Pastoris

The lack of reliable online sensors, which can accurately detect the important state variables, is one of the major challenges of controlling bioprocess accurately, automatically and optimally in biochemical industries. In this paper Artificial Neural Network (ANN) based soft sensors were developed to predict cell concentration and product activity of recombinant pichia pastoris fed batch fermentation. Two types of ANN based soft sensor namely static and dynamic neural network were developed for prediction of inter sample values of low frequency cell concentration and product activity measurement using high frequency online measurements such as pH, Dissolved Oxygen (DO), temperature and stirring speed. Results indicate that a dynamic neural network based soft sensor shows a better performance compared to static neural network based soft sensor.