The Design-to-Operation Pipeline
A software pipeline carries a fusion design from parametric study to fabrication data to live plant control.
A pipeline, not a handoff
In immature engineering, design and operations are separate worlds connected by paperwork. A modern fusion program treats them as one pipeline: parameters, geometry, and models flow forward, and measurements flow back, through shared data structures rather than re-keyed documents.
Stages of the pipeline
- Parametric design: sweep physics and engineering variables to find viable operating points.
- Detailed engineering: convert chosen points into geometry, materials, and tolerances.
- Fabrication data: generate the exact instructions and check them against the model.
- Commissioning: match as-built hardware back to the model.
- Operation: run control and analysis on top of the same model.
Single source of truth
Each stage reads from and writes to a common definition of the machine. When the on-axis field of the breeder Hyperion is stated as 8 T with a 16.84 T peak, that number lives in one place and propagates everywhere, so a change cannot silently disagree between a design deck and a control setpoint.
Traceability
Every artifact in the pipeline should carry its provenance: which model version produced it, from which inputs. This makes it possible to answer, months later, why a component has the dimensions it has.
Why it pays off
A connected pipeline shortens the loop between learning something in operation and improving the next design. Construction of Hyperion is planned for Q2 2027; the pipeline built now is what will carry operational lessons back into the fleet that follows.