Robot Staging Depots and Spare-Unit Planning, The Redundancy Arithmetic
At a glance: Everyone plans the fleet. Almost nobody plans where the fleet lives when it is not working, and that is where redundancy quietly fails. A spare robot sitting in a warehouse 40 minutes away is not redundancy, it is a promise. This guide covers depot siting, footprint and power, hot-swap time as a specification, and the arithmetic that tells you how many spare units you actually need rather than how many you were sold.
The Difference Between a Spare and a Swap
Three words get used interchangeably in fleet planning and they are not the same thing, which is why redundancy budgets get spent badly.
- Spare part. A component held in stock: a brush, a battery pack, a sensor module. It restores a robot after a repair, and the restoration takes however long the repair takes.
- Cold spare. A complete robot, powered off, sitting in a depot, not counted in the operational fleet. It restores coverage as fast as it can be commissioned and dispatched.
- Hot swap. A complete robot, charged, mapped, and ready to enter service immediately. It restores coverage in minutes, because nothing has to be commissioned when the failure happens.
Only the hot swap delivers continuity. A cold spare is a cost-effective way to cover a repair cycle measured in days. The mistake we see most often is buying a cold spare and expecting hot-swap behaviour from it, which is a planning error rather than a hardware one, and it is where the downtime cost model usually gets its unhappy surprise.
Where to Site the Depot, Four Siting Rules
A depot exists to shorten the distance between a failure and restored coverage. Its location matters more than its size.
- Inside the response radius, not near the site. Draw the radius from the response time you have committed to, not from convenience. If the service agreement promises a two-hour restore, and transport plus commissioning takes 45 minutes, the depot has to be inside the remaining window with margin.
- On the robot's own navigation map where possible. A depot in a room the robot already knows removes a commissioning step from every swap. If the depot is off-map, the swapped unit needs a map update before it can work, and that is the commissioning time you were trying to avoid.
- Adjacent to power and network, centrally within the coverage area. The depot should sit where travel to any point in the building is roughly equal, so no zone waits longer than another for a replacement.
- Secure but accessible outside core hours. A depot that needs a key-holder at 3 a.m. has a hidden labour cost in every swap. Access control that logs entry is better than a lock that needs a person.
For multi-building portfolios the siting logic repeats per site rather than centrally. A single depot serving five buildings is a hub for parts, not for swaps, and the two should not be confused. The portfolio-level view is covered in multi-site deployment strategy.
Footprint and Power, the Numbers
A staging depot is a small room with more infrastructure than people expect. Size it from the unit count you intend to stage, not from the room you happen to have.
| Element | Per staged unit | Notes |
|---|---|---|
| Floor area, parked | 1.2 to 2.0 m² | Depends on unit footprint plus clearance for a person to walk to the rear |
| Clearance aisle | 0.9 m minimum, 1.2 m preferred | Less than this and the swap becomes a two-person job |
| Charging circuit | Dedicated, per dock | Do not share with domestic sockets; charging is the load that trips a circuit |
| Power per dock | Typically 300 to 900 W continuous | Confirm from the dock datasheet, not the robot datasheet |
| Ventilation | Air change appropriate to battery chemistry | Lithium charging in a closed room needs a stated air change rate |
| Network | Coverage of −65 dBm at every dock | Utility corners are the worst-covered place; survey them explicitly |
| Shelf space for spares | 1 to 2 m² per 5 units | Brushes, pads, filters, and at least one spare battery pack |
The power figure is the one that catches people. A staging area with eight docks is a multi-kilowatt continuous load, and if it is on a lighting circuit it will trip the first time all eight recharge together. Plan a dedicated circuit and stagger the charge schedule, which is the same fix as the fleet charging load planning at the main site, applied to a smaller room.
Hot-Swap Time Is a Specification, Not a Hope
If the service agreement contains a restore-time commitment, then hot-swap time is a number that has to be designed and then measured. Break it into its components and each one can be attacked.
| Swap step | Target | What it depends on |
|---|---|---|
| Failure detected and confirmed | < 5 min | Telemetry and alerting, not a person noticing |
| Decision to swap made | < 5 min | A pre-agreed threshold, not a meeting |
| Spare powered, checked, dispatched | < 10 min | Spare held charged and mapped |
| Transit to zone | Site dependent | Depot siting, discussed above |
| Task handover and map sync | < 10 min | Spare on the same map and same fleet software version |
| Total, within a single building | 30 to 45 min | Realistic with a hot spare staged on site |
The steps that are usually neglected are the last two. A spare on a stale map, or on a fleet software version the server does not recognise, turns a 30-minute swap into a 90-minute commissioning job. Keeping the spare on the same software version as the fleet is an operational discipline, and it is the cheapest reliability improvement available.
The Cold-Spare Ratio, How Many You Actually Need
The tempting answer is one spare per site. The correct answer comes from the failure rate and the acceptable coverage gap, and it is usually lower than the vendor suggests and higher than the budget hopes.
A simple ratio model: if units fail at an average rate of f per unit per year, fleet size is N, and the target is that at least k units are always available, then the number of spares S should satisfy the expected simultaneous failures the site can tolerate.
| Fleet size | Typical annual failure rate | Expected failures per year | Practical spare policy |
|---|---|---|---|
| 1 to 3 units | 1.0 to 1.5 per unit | 1 to 4 | At least 1 hot spare on site, or accept planned gaps |
| 4 to 8 units | 0.8 to 1.2 per unit | 3 to 10 | 1 hot swap plus 1 cold spare, or 20% cold ratio |
| 9 to 20 units | 0.6 to 1.0 per unit | 5 to 20 | 1 hot swap plus 10 to 15% cold ratio |
| 20+ units | 0.5 to 0.8 per unit | 10 to 16+ | Pooled cold spares across sites, 8 to 12% ratio |
Two things move this table in practice. First, larger fleets have a lower per-unit failure rate, because shared infrastructure absorbs some failures and the failure statistics are diluted by well-run units. Second, a fleet with a strong preventive maintenance regime shifts failures from the unplanned column to the planned column, which reduces the number of spares you need to hold. That is the real financial argument for the maintenance programme covered in preventive maintenance scheduling: it is not just cheaper per repair, it is cheaper per spare.
Depot Cost Against the Cost of a Downtime Hour
A depot is not free, and the business case should be stated as a comparison rather than an assertion. The two sides of the ledger:
| Depot investment | Value delivered |
|---|---|
| Room, fit-out and racking | Reduced mean time to restore coverage |
| Dedicated power and charging docks | Spares held charged, ready to swap |
| Network extension into the depot | Spares stayed mapped and version-current |
| Cold spare unit capital | Insurance against a multi-day repair |
| Labour to operate the depot | Converted into the swap time saved per incident |
The decision rule is straightforward. If the value of the coverage restored across a year exceeds the annualised cost of the depot, build the depot. That value comes from the downtime cost per hour multiplied by the hours recovered, and if you have not measured the downtime cost per hour, that is the first number to obtain, because everything in this section depends on it.
The Takeaway
Redundancy that lives 40 minutes away is not redundancy. A staging depot turns a fleet's recovery capability from an aspiration into a distance, and the components of that distance are all engineerable: where the depot sits, how big it is, what power it has, how fast a swap can run, and how many spares the failure arithmetic justifies. Specify the restore time, site the depot inside it, keep the hot spare charged and on the current software, and size the cold spares from the failure rate rather than from habit. The costs are small and the answer to the question nobody plans is a bay-marked square on a utility room floor.
