Digital Twin Calibration for Organic System-of-Techniques: Cell Tradition Manufacturing Course of
Authors: Fuqiang Cheng, Wei Xie, Hua Zheng
Summary: Biomanufacturing innovation depends on an environment friendly design of experiments (DoE) to optimize processes and product high quality. Conventional DoE strategies, ignoring the underlying bioprocessing mechanisms, usually undergo from a scarcity of interpretability and pattern effectivity. This limitation motivates us to create a brand new optimum studying method that may information a sequential DoEs for digital twin mannequin calibration. On this examine, we contemplate a multi-scale mechanistic mannequin for cell tradition course of, often known as Organic Techniques-of-Techniques (Bio-SoS), as our digital twin. This mannequin with modular design, composed of sub-models, permits us to combine knowledge throughout numerous manufacturing processes. To calibrate the Bio-SoS digital twin, we consider the imply squared error of mannequin prediction and develop a computational method to quantify the affect of parameter estimation error of particular person sub-models on the prediction accuracy of digital twin, which may information sample-efficient and interpretable DoEs.
