The aerolog-scaling programme sizes a cargo VTOL and, in its hub milestones, asks what limits a delivery hub: pads, chargers or spare packs. Two earlier steps changed the spare-pack count one factor at a time: a constant-voltage charge taper (fitted to the Molicel INR-21700-P45B datasheet) and a daily demand shape (City of Chicago ride-hail trips per 15 minutes, used only for its shape; it is not delivery data).
This step runs both together at 6 flights per hour, with 8 pads and 8 chargers. Spares needed so that the 95th-percentile wait is at most 300 s:
| Charge window | Poisson | Wednesday shape | Saturday shape |
|---|---|---|---|
| 20 to 90% (43 min) | 9 | 9 | 9 |
| 30 to 100% (72 min) | 20 | 28 | 28 |
I predicted the full-charge case would need at least double the Poisson count, because the daily peak (10.6 flights/h) is above that window's charger limit (6.7/h). That was wrong: 1.4x. A pool of packs rides out a peak of a few hours. With the 90% charge ceiling the daily shape costs nothing at this rate.
Limits: the spares grid is coarse, the demand shape is ride-hail, and the exponential taper decay is an assumption. Results and code are in research/aerolog-scaling/experiments/hub_throughput/. AI-assisted research; preliminary.