Towards a digital twin: a hybrid data‐driven and mechanistic digital shadow to forecast the evolution of lignocellulosic fermentation

The high substrate variability and complexity of fermentation media derived from lignocellulosic feedstock affects the concentration profiles and the length of the fermentation. Failing to account for such variability raises operational and scheduling issues and affects the overall performance of these processes. In this work, a hybrid soft sensor was developed to monitor and forecast the evolution of cellulose‐to‐ethanol fermentation. The soft sensor consisted of two modules (a data‐driven model and a kinetic model) connected sequentially. The data‐driven module used a partial‐least‐squares model to estimate the current state of glucose from spectroscopic data. The kinetic model was recursively fitted to known concentrations of glucose to update the long‐horizon predictions of glucose, xylose, and ethanol. This combination of real‐time data update from an actual fermentation process to a high‐fidelity kinetic model constitutes the basis of the digital twin concept and allows for a better real‐time understanding of complex inhibition phenomena caused by different inhibitors commonly found in lignocellulosic feedstocks. The soft sensor was experimentally validated with three different cellulose‐to‐ethanol fermentations and the results suggested that this method is suitable for monitoring and forecasting fermentation when the measurements provide reasonably good estimates of the real state of the system. These results would allow the flexibility of the operation of cellulosic processes to be increased, and would permit the scheduling to be adapted to the inherent variability of such substrates © 2020 Society of Chemical Industry and John Wiley & Sons, Ltd.