Product Quality Sensors for ML-Guided Process Optimization and Control of Modularized Production Plants
Members: Iowa State University, Novonesis
Year launched: 2023
This project will create a generalizable machine learning framework for optimization and control of bioreactors to reduce resources needed to design new processes and improve product quality throughout production. This will include implementation of new product quality sensors capable of measuring secreted biocatalyst activity continuously and directly from the reactor. This approach is unique because current methods of process optimization rely on much manual intervention and indirect measures of product quality. The project will focus on glucoamylase from a non-proprietary strain as a demonstration, allowing for the ML framework and training data to be shared. This research will be used to create an open-source framework for ML-enabled process optimization and control, that is adaptable to new sensor inputs and process control variables. Upon completion, the enzyme activity sensors will be commercialized and expanded to characterize other hydrolases and enzyme classes (e.g., transferases, ligases, etc.). As more enzymes are used to achieve sustainability goals in water, energy, and carbon reduction in many industries, this efficient design and process control tool will be critical to more rapid market entry and better product quality.
Funding source: Schmidt Futures Bioreactor Innovation Project Call
Related news: