Models and Simulations of Design

Models and simulations of design integrate computational representations of products, processes and systems with numerical experiments to predict performance, guide decision‐making and accelerate innovation. Parametric models encode geometric, material and functional aspects of a design, while virtual prototypes enable interactive exploration of shape, topology and operational variables. Simulation techniques—ranging from finite element analysis and computational fluid dynamics to multiscale and agent‐based methods—allow designers to assess structural integrity, aerodynamic efficiency, thermal behaviour and manufacturability before physical prototyping. Surrogate models and optimisation algorithms further refine designs by approximating costly high‐fidelity simulations, balancing multiple performance objectives and constraints. Advances in model interoperability, uncertainty quantification and high‐performance computing are transforming design workflows into iterative, data‐driven loops that integrate machine learning and real‐time feedback. This convergence is paving the way for generative design, adaptive architectures and autonomous systems that can self‐configure according to performance criteria and environmental interactions.

Research from Nature Portfolio

Recent work has evaluated parallel infill sampling strategies for Gaussian process surrogates, comparing criteria such as expected improvement, minimal predicted improvement and probability of improvement. These studies demonstrate how appropriate selection of acquisition functions in a parallel Bayesian optimisation framework can markedly accelerate convergence on complex engineering objectives while controlling computational overhead. In another development, a hybrid Kriging–Grey Wolf optimiser has been applied to robust transmitter‐placement problems. By interpolating signal coverage and power consumption metrics via a surrogate, researchers produced Pareto‐optimal frontiers that balance energy efficiency and coverage reliability, achieving high‐quality multi‐objective solutions with substantially fewer expensive evaluations.

Models and Simulations of Design publication trend

The graph below shows the total number of articles in models and simulations of design across all publications each year (not limited to Nature Index journals).

Technical terms

Surrogate model: A computationally efficient approximation of a high‐fidelity simulation or experiment used to predict performance metrics.

Gaussian process (Kriging): A probabilistic regression method that models functions with uncertainty estimates, central to Bayesian optimisation.

Acquisition function: A criterion in Bayesian optimisation that balances exploration of uncertain regions and exploitation of promising designs.

Hybrid mesh: A mesh combining different element types (e.g. prisms, tetrahedra) to capture boundary layers and interior domains in simulations.

Delaunay triangulation: A mesh generation method that maximises minimum angles, facilitating high‐quality unstructured discretisations.

C2 continuity: A smoothness condition requiring continuity of a function and its first and second derivatives across element boundaries.

Simplex spline: A generalisation of B‐splines defined on triangulations, offering partition of unity and local control for surface modelling.

References

1. Comparison of parallel infill sampling criteria based on Kriging surrogate model. Scientific Reports (2022).
2. A data driven approach in less expensive robust transmitting coverage and power optimization. Scientific Reports (2022).
3. Hybrid meshing using constrained Delaunay triangulation for viscous flow simulations. International Journal for Numerical Methods in Engineering (2016).
4. Construction of C2 Cubic Splines on Arbitrary Triangulations. Foundations of Computational Mathematics (2022).
5. Method of matched sections in application to thin-walled and Mindlin rectangular plates. Mechanics and Advanced Technologies (2023).