Future fusion power plants will operate in regimes far beyond current experiment regimes, requiring physics-based predictive modeling to assess performance, feasibility, and risk. Fusion plant design is inherently a multi-physics, multi-scale, and multi-fidelity problem, coupling plasma, material response, and engineering analyses within iterative optimization workflows. The Fusion REactor Design and Assessment (FREDA) project is developing a flexible component-based integrated modeling framework to enable self-consistent, end-to-end modeling of fusion devices. FREDA combines a hierarchy of core-to-edge plasma models spanning analytic scaling relations, reduced transport models, and high-fidelity simulations, coupled to engineering analyses of thermal, structural, and nuclear performance. These workflows enable parameter scans, sensitivity studies, surrogate model construction, and geometry optimization within a unified framework. Recent developments include coupled core-edge-SOL (CESOL) modeling to map plasma heat and particle flux to plasma-facing components. However, a central challenge remains: how to use integrated modeling to make reliable design decisions. In practice, model forms are often chosen ad hoc, with limited ability to quantify which model predictions can be trusted for a given application. Ideal workflow would include 1) model selection and credibility assessment across fidelity levels, 2) uncertainty propagation through tightly coupled multi-physics workflows, and 3) quantification of extrapolation error from present-day validation regimes to fusion plant conditions. This talk will present example tokamak design workflows from FREDA and highlight challenges as opportunities for advances in multi-fidelity methods, uncertainty quantification, and optimization for fusion device design.
* Supported by the FREDA SciDAC and US DOE DE-AC05-00OR22725