Development of new algorithms to increase the accuracy, robustness and reproducibility of intelligent instruments in response to the requirements of the industrial biotechnology industry

Real-time monitoring of bioprocesses can increase performance and reduce losses. The main difficulty comes from the limited amount of information available in real time and the complex and highly non-linear phenomena involved. Thus, conventional statistical algorithms do not allow to interpret adequately and with the accuracy required in industrial biotechnologies the signals measured by spectral probes.

The objective of the project is to develop new complex non-linear algorithms that will allow the transformation of complex raw spectral signals into accurate, robust and reproducible performance indicators.

This project will first develop an algorithm that improves the performance of real-time estimation of product concentrations in fermentations and cell cultures. It will also allow the development of a second algorithm that will allow the real-time detection of anomalies within a bioprocess.