This paper introduces SuperCond-GNN, a graph neural network surrogate model designed to predict voltage distribution in high-temperature superconducting magnets by mapping lumped-element circuits to graph representations. The model achieves a mean MAPE of 4.3% on tape stacks and enables fast inference of current redistribution across various circuit configurations.
- HTS magnets are modeled as lumped-element equivalent circuits and mapped onto graphs for message passing GNNs.
- The surrogate is trained on data from circuit simulations, achieving a mean MAPE of 4.3% within the design space.
- Physics-informed regularization via Kirchhoff's current law is evaluated to enhance model performance.
- Generalizability is assessed through zero-shot inference and few-shot fine-tuning on unseen topologies.
- The framework is topology-agnostic and extensible to complex HTS cable and magnet configurations.
This approach offers a scalable alternative to conventional circuit solvers for downstream applications such as design space exploration, current sharing analysis, and real-time magnet monitoring.