Transforming data to information: A parallel hybrid model for real‐time state estimation in lignocellulosic ethanol fermentation

Abstract

Operating lignocellulosic fermentation processes to produce fuels and chemicals is challenging due to the inherent complexity and variability of the fermentation media. Real‐time monitoring is necessary to compensate for these challenges, but the traditional process monitoring methods fail to deliver actionable information that can be used to implement advanced control strategies. In this study, a hybrid‐modeling approach is presented to monitor cellulose‐to‐ethanol (EtOH) fermentations in real‐time. The hybrid approach uses a continuous‐discrete extended Kalman filter to reconciliate the predictions of a data‐driven model and a kinetic model and to estimate the concentration of glucose (Glu), xylose (Xyl), and EtOH. The data‐driven model is based on partial least squares (PLS) regression and predicts in real‐time the concentration of Glu, Xyl, and EtOH from spectra collected with attenuated total reflectance mid‐infrared spectroscopy. The estimations made by the hybrid approach, the data‐driven models and the internal model were compared in two validation experiments showing that the hybrid model significantly outperformed the PLS and improved the predictions of the internal model. Furthermore, the hybrid model delivered consistent estimates even when disturbances in the measurements occurred, demonstrating the robustness of the method. The consistency of the proposed hybrid model opens the doors towards the implementation of advanced feedback control schemes.

Digital Twin in biomanufacturing: challenges and opportunities towards its implementation

“…State estimators such as Kalman filters are often considered a more robust approach than RPE because they merge process data with the model predictions without altering model parameters. Numerous examples of different state estimating algorithms applied to fermentation processes are described in the literature [14,[25][26][27][28][29][30]. Successful implementation of state estimation at bench scale (15 L) was developed by Krämer et al [14,29] using an extended Kalman filter (EKF) [14] and a sigma point Kalman filter (SPKF) [29].…”

“…The authors estimated the states of yeast fermentation by combining on-line and at-line measurements of the optical density, carbon dioxide, pH and spectroscopic measurements of the substrates and product concentrations. Lopez et al [30] also used an EKF to monitor the concentrations of glucose, xylose and ethanol from ATR-MIR spectroscopy in second-generation ethanol fermentation processes. Golabgir and Herwig, as well as Kager et al [27,28], used a particle filter to estimate the process states and kinetic rates in Penicillium chrysogenum fed-batch fermentation.…”

“…Cabaneros et al [4,30] used ATR-MIR spectroscopy to measure on-line the concentration of key state variables using a recirculation loop connected to a flow cell. From a practical point of view, the real-time measurement of glucose holds the key to ensure an optimised fermentation operation.…”

“…From a practical point of view, the real-time measurement of glucose holds the key to ensure an optimised fermentation operation. The state variables (glucose concentration in [4] and glucose, xylose and ethanol concentrations in [30]) were calculated from the collected spectra using a partial least squares (PLS) regression model. Even though the spectra were turned into actionable information, the monitoring platform alone does not constitute a Digital Shadow.…”

“…Therefore, incorporating the real-time measurements with the predictions made by the Digital Model is the basis for developing a full Digital Shadow. Cabaneros et al implemented two alternative approaches (a recursive parameter estimation [4] and a state observer [30]) to incorporate measurements into the Digital Model predictions. In [4], Cabaneros et al sequentially combined the PLS regression model with the Digital Model using a recursive parameter estimation approach.…”