In computational mechanics, the impact of uncertain phenomena is commonly evaluated by probabilistic techniques inducing iterative procedures. However, the repetitive evaluation of nonlinear Finite Element analysis poses high computational costs. This thesis proposes a multi-fidelity scheme exploiting projection-based Model Order Reduction techniques for structural analysis to reduce the computational burden. The key feature is the intertwining of intrusive and non-intrusive reduction methods tailored to analysis and optimisation under uncertainty.
«
In computational mechanics, the impact of uncertain phenomena is commonly evaluated by probabilistic techniques inducing iterative procedures. However, the repetitive evaluation of nonlinear Finite Element analysis poses high computational costs. This thesis proposes a multi-fidelity scheme exploiting projection-based Model Order Reduction techniques for structural analysis to reduce the computational burden. The key feature is the intertwining of intrusive and non-intrusive reduction methods ta...
»