Adaptive Optimization of Microbial Growth Processes

The problem of maximizing the biomass productivity in continuous microbial processes is addressed. An adaptive steady-state optimization algorithm is developed. The algorithm is based on the on-line identification of a dynamic model, and its steady-state version is used to calculate the optimum value of the dilution rate at each step. The algorithm\´s ability to attain the optimum and handle disturbances is demonstrated, employing a continuous chemostat model as the "true plant". Both linear and bilinear identified models are shown to succeed, the success not influenced by the chosen model order.