Current Advances in Tools Improving Bioreactor Performance

Improving the performance of their bioreactors is a continuous distinguished task of process engineers for the entire lifetime of their processes. In this paper new developments which help to improve the performance of bioreactors for recombinant protein formation are reviewed. In order to judge the performance of a cultivation process an important prerequisite is accurately and easily monitoring the key quantities of their processes. Recently, the well-established monitoring techniques, got competition from new, so-called kernel methods, which are more precise as they need a smaller number of free model parameters. In the field of dynamic estimators, “Extended Kalman Filters” are being replaced by “Unscented Kalman Filters” which simplify this sophisticated technique as they do not need Jacobian matrices. They use the full process model instead of linear approximations. Process supervision, optimization and control examples are given from microbial fermentations as well as from animal cell cultures. Optimization the operation of the fermentations, an extremely prominent task, is discussed in case of E. coli cultivations where the product appears in soluble form as well as in the form of inclusion bodies. In order to keep the process on its optimal path at a significant batch-to-batch reproducibility, open loop control along robust trajectories should be used for the initial biomass growth phase. Later, during the product formation phase, closed loop feed forward feedback control is the technique of choice to keep the process on its optimal trajectory. All procedures mentioned can be implemented in modern industrial automation systems. Advances in such bioreactor control systems develop towards virtual plants, which simultaneously allow simulating the fermentation process. They allow for improved controller developments, online process supervision, and can also be used for a more realistic training of the personnel similarly to flight simulators for pilot training.