Tritium Plant FMEA

Failure modes of breeding, extraction, and accountancy for the ~4 kg/yr tritium class and ~1.97 kg/yr helium-3 outputs.

The fuel cycle is a plant

The breeder (Hyperion) is valued for its products: a ~4 kg/yr tritium class output, ~1.97 kg/yr helium-3, and 14 MeV neutrons. That makes the breeding blanket, tritium extraction, and isotope accountancy a plant in their own right, with their own failure modes. The tritium breeding ratio is a design lever across 1.1/1.5/1.8, so the fuel-cycle FMEA is sensitive to which point is chosen.

Accountancy is safety-critical

Tritium is regulated as byproduct material, so mass-balance accountancy is both an operational and a compliance function. An undetected accountancy drift is high-detection-difficulty because it is silent - it does not trip anything - which raises its RPN even at moderate severity. Sensor fusion across neutron yield, extraction flow, and inventory closes the balance.

def mass_balance(bred, extracted, inventory_delta, tol):
    # bred - extracted should equal change in held inventory within tolerance
    residual = bred - extracted - inventory_delta
    return abs(residual) <= tol, residual   # flag if out of balance

TBR lever and resiliency

Choosing a higher TBR (toward 1.8) buys margin against breeding-chain faults but stresses the blanket; a lower TBR (1.1) is leaner but less fault-tolerant. Resiliency treats the lever as an input to the availability model, not a fixed number. Slow-degradation modes feed predictive maintenance. Pre-FOAK the scores derive from blanket simulation, not an operating fuel cycle.