Self-updating digital twin of a hydrogen-powered furnace using data assimilation

Data-driven models and digital twins for sustainable combustion technologies

“…The latter have gained momentum in the combustion community to update numerical simulations based on high-fidelity measurements, 110 to calibrate physics-based ROMs, 109 , 111 and to update DTs based on few experimental measurements. 112 ROMs based on dimensionality reduction and regression methods are particularly appealing in this framework for the possibility of updating the regression coefficients (of a GPR regression or a neural network) based on the available data stream. ROMs based on CRNs can also be very appealing for their gray-box nature: data-driven algorithms can be designed to develop generalized CRNs consisting of combinations of ideal reactors for which the governing equations are known.…”