Fleet Availability Aggregation

Several independent units, a grid tie, and storage combine so the plant's availability far exceeds any single burner's 0.86-0.995 — the mechanism that reaches Tier III.

The whole beats the parts

No single burner reaches Tier III, but a plant of several units plus independent layers can, because the plant is down only when enough members are down simultaneously. Aggregation across independent members is the mathematical engine that turns 0.86-0.995 units into a Tier III plant.

# Plant availability: load needs N units; plant has N+spares, plus grid.
from itertools import product

def unit_up_prob(a=0.95):     # single burner availability
    return a

def plant_meets_load(n_needed, n_units, a_unit, a_grid):
    # brute force over up/down states of units + grid
    ok = 0.0
    for state in product([0,1], repeat=n_units+1):
        units_up = sum(state[:-1]); grid_up = state[-1]
        p = 1.0
        for s in state[:-1]:
            p *= a_unit if s else (1-a_unit)
        p *= a_grid if grid_up else (1-a_grid)
        # grid can supply the shortfall up to the load
        if units_up >= n_needed or (grid_up and units_up >= n_needed-1):
            ok += p
    return ok

print(round(plant_meets_load(2, 4, 0.95, 0.999), 5))  # >> 0.95

The model is illustrative, not a specification, but it makes the point: a plant sized with spare units and backed by an independent grid reaches a combined availability well above any single unit's. This is exactly the nines arithmetic applied to a real fleet, and it is how the availability gate is answered in practice rather than in words.

The honest caveats carry over. Real members are not perfectly independent — a shared bus, common controls, or a regional grid event can correlate outages — so the true aggregate is below the idealized number, and the design must minimize those common-mode paths. Fuel is another shared dependency: all units draw helium-3, so a fuel shortfall is a correlated risk the electrical redundancy cannot fix.

Aggregation is therefore powerful but not automatic: it delivers Tier III only when the members are genuinely independent and the shared dependencies — bus, controls, fuel — are managed as first-class risks.

• Plant is down only when enough members fail together
• Spare units + grid raise combined availability above any unit
• Real independence is imperfect; minimize common-mode paths
• Shared fuel (He-3) is a correlated risk redundancy cannot fix