We provide extended analyses and interpretations of the parametrized-background data-weak (PBDW) formulation, a real-time in situ data assimilation framework for physical systems modeled by parametrized partial differential equations. The new contributions are threefold. First, we conduct an a priori error analysis for imperfect observations: we provide a bound for the variance of the state error and identify distinct contributions to the noise-induced error. Second, we illustrate the elements of the PBDW formulation for a physical system, a raised-box acoustic resonator, and provide detailed interpretations of the data assimilation results in particular related to model and data contributions. Third, we present and demonstrate an adaptive PBDW formulation in which we incorporate unmodeled physics identified through data assimilation of a select few configurations.
5 August 2016 in Archives of Computational Methods in Engineering
We present a model-order-reduction approach to simulation-based classification, with particular application to structural health monitoring. The approach exploits (1) synthetic results obtained by repeated solution of a parametrized mathematical model for different values of the parameters, (2) machine-learning algorithms to generate a classifier that monitors the damage state of the system, and (3) a reduced basis method to reduce the computational burden associated with the model evaluations. Furthermore, we propose a mathematical formulation which integrates the partial differential equation model within the classification framework and clarifies the influence of model error on classification performance. We illustrate our approach and we demonstrate its effectiveness through the vehicle of a particular physical companion experiment, a harmonically excited microtruss.