Line-Down Clock

Hours until each plant stops, per part number. Every number on the board can be recomputed by hand from the inputs, which is what makes it usable in a war room at 2am.

The mechanism

Consumption is always derived, never stored:

burn_uph = plant.line_rate_uph * part.per_vehicle

Time to first stop walks the arrival timeline forward. Between arrivals stock depletes at burn_uph; at each arrival stock steps up. The first interval in which stock would go negative gives the exact stop hour:

stock = on_hand; t = 0
for arrival in sorted(arrivals):
    gap = arrival.eta_h - t
    if stock - gap * burn < 0:
        return t + stock / burn      # stops before this truck lands
    stock -= gap * burn
    t = arrival.eta_h
    stock += arrival.qty
return t + stock / burn              # stops after the last truck

Downtime is the number that matters

First-stop time answers "when do we stop". It does not answer "how long are we down", and those come apart the moment a recovery truck lands after the stop. An expedite that arrives an hour late cannot prevent the stop, but it can still cut an eight hour outage to one. So the board also integrates total starved hours across the 72 hour horizon, and the re-planning demo makes every recovery decision against that number rather than against the first stop.

Status bands

Band	Hours to stop	Meaning
critical	under 4	Inside a single shift. Escalate now.
at risk	under 12	Recoverable today if a decision is made this shift.
watch	under 24	Monitor; normal replenishment may still cover it.
ok	covered to 72	No stop inside the horizon.

Reading the board

• Burn /h is line rate times per-vehicle usage. A Class 8 line at 18/hour running 8 drive tires per truck burns 144 tires an hour.
• Cover is on-hand divided by burn, ignoring anything in transit. This is the honest floor.
• Stops at folds in-transit arrivals into the timeline. A part with thin cover but a truck landing in two hours is not the emergency; a part with more cover and nothing scheduled often is.
• Downtime 72h is the total starved hours, which is why two parts that both stop at 4h can still deserve very different responses.
• JIS marks a sequenced part. Those are thin by design and are the ones a sequence break hurts.
• border marks a cross-border supplier, the population a customs hold hits.

Scenarios

Scenario	What it models
normal	Comfortable coverage. Most parts covered to the horizon.
tight	Lean inventory. Several parts inside 12 hours.
launch	New model ramp. Sequenced parts run thin and some parts go critical.
peak	Line rate up 25 percent, so the same stock covers proportionally less.