About this Research Topic
In the last decades, the rapid advancement of digital technologies has provided new tools to complement and enhance traditional scientific methodology and engineering design procedures. In particular, great progress has been made in the fields of numerical simulation and data processing, establishing these techniques as standard scientific or engineering tools alongside with more traditional approaches such as theoretical derivations and experiments. A more recently emerging trend would be the development of techniques relying on both data and models to provide more accurate representations of engineered systems or physical problems, thus blurring the boundaries between the aforementioned fields.
In engineering and computational mechanics, where the use of both models and experimental data is almost a necessity, several instances of this new trend can already be found:
• In the solution of inverse or identification problems and model updating, numerical models are modified based on measurement data to more accurately describe the current state of structures or systems.
• The development of reduced order techniques creates the potential for real time simulations that run alongside actual problems and often provide the possibility of interaction.
• Methods for uncertainty quantification/propagation employ models to assess the effect of uncertainty in certain parameters on the response of engineering systems.
• Advanced discretization techniques, such as cut finite element methods (CutFEM) or fictitious domain methods, facilitate the creation of models directly from data.
The current Research Topic will aim at further advancing the current state of the art by presenting methods and techniques dealing with different aspects of model-data fusion. Some possible topics for potential contributions would include:
• Interface between multi-scale methods and statistical methods for material modelling, in particular uncertainty quantification, propagation and multi-scale inverse problems.
• Uncertainty propagation through partial differential equations as well as “full uncertainty cycles” starting from data to identification to propagation through PDEs or other mathematical models.
• Data-driven inverse methods and data assimilation, in particular Bayesian inference and regularization, regression, projection and extrapolation, real-time assimilation, data fusion.
• Acceleration methods for large-scale (industrial) applications (model order reduction).
• Statistical approaches to build relevant parametric distributions.
• Generation of numerical models directly from image data.
Keywords: Virtualization, Digital Twins, Uncertainty Quantification, Inverse Problems, Model Order Reduction
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.