FLO Mobility

Case study · Stargate Data Centre, UAE

Four workers pushed debris all day.
The robot runs that route on its own now.

Measured over 28 recorded trips at the Stargate data centre site: what manual material movement was costing the crew in strain and wasted capability, and what changed when an autonomous machine took the route off them.

  • Stargate Data Centre
  • ALEC, UAE
  • 28 trips recorded
  • Worker impact study
Flo Mobility material movement robot on the Stargate data centre site
of the route's physical work now carried by the machine
93%
workers taken off the debris route entirely
4
no longer pushed under load, per worker per shift
2.1 km
per trip, now carried by the machine instead of four people
400 kg
returned to trade work every month
858 hrs

Site footage

The same route, worked two ways

Manual

Two workers to one loaded trolley

Full body effort for the length of the route

Manual

A crew gathered round one load

Four hands minimum, every single trip

Autonomous

The robot runs the route alone

Nobody walking it, nobody pushing it

1What the job asked of the body

A full shift of strain that produced nothing

Debris movement from the building to the garbage yard ran on four workers and a pallet trolley. Four hundred kilos behind four people, over a route with sloped sections and stretches of broken ground. They averaged 76 minutes per trip and completed about seven trips in a ten hour shift, which is very nearly the whole working day. Not one minute of it advanced the build.

Load behind the crew, every trip

400 kg

Seven times a shift

The same 400 kg, trip after trip, all day

Distance walked under load

2.1 km

4.2 km walked in total

Loaded trolley outbound, empty trolley back

Push force each worker sustains

100–200 N

Over the limit on slopes

Indicative limit for sustained pushing is 100–150 N

Shift hours consumed by the route

8.9 hrs

Of a 10 hour shift

Four workers with no capacity left for anything else

Physical burden per worker, per ten hour shift

What the body absorbed on the manual method, and what remains once the robot runs the route

Distance pushed under load
Manual
2.1 km of loaded trolley
Robot
NoneFully removed
Unloading by hand
Manual
7 unload cycles of 400 kg
Robot
NoneAuto dump
Hours tied to the route (of a 10 hour shift)
Manual
8.9 hrs per worker
Robot
2.5 hrs, one personPart time

Human hours the route consumes per shift

35.5 man hours down to 2.5, a 93% reduction in human effort

Human hours the route consumes per shift
CategoryManual crewRobot
Manual crew of four35.5 hrsNo Robot value
Robot, part time coverNo Manual crew value2.5 hrs

Where the 76 minutes of a manual trip went

71% of every trip was hands on the material, loading and unloading by hand

Where the 76 minutes of a manual trip went
CategoryHands on the materialWalking the route
Loading by hand26 minNo Walking the route value
Pushing outNo Hands on the material value11 min
Unloading by hand28 minNo Walking the route value
ReturnNo Hands on the material value11 min

Most of the trip was hands on the material

Loading and unloading by hand accounted for 54 of every 76 minutes. The walking was the visible part of the job, but the lifting and lowering is where the bodies were spent, and auto dump removes half of that outright.

Nothing about the task was skilled

Four workers, a trolley and a 300 metre route. The work demanded stamina and nothing else, from people hired to build a data centre. That is the real waste in the old method, and it never shows up on a programme report.

2Fatigue, measured rather than assumed

The gap between people and machine is the fatigue curve

Head to head trials on a 300 metre loop at the site, run at a light trial load and again at the 400 kg the route actually demands, plus a continuous 60 minute endurance run. Light, the two are level. At 400 kg the difference is the human body reaching its limit.

Light trial load, 300 m loop (robot vs crew)

3:27 vs 3:20

Level, 7 seconds apart

A fit worker matches the robot when the load is light

At the route load of 400 kg (robot vs crew)

3:32 vs 5:37

Crew 2:05 slower per trip

The robot holds its pace. The crew cannot.

Robot pace decay over 60 minutes

None

Trip 19 matches trip 1

No rotation, no recovery breaks, no drop off

Workers needed to hold that pace

Two, alternating

Even at a light load

One person could not sustain the cycle for an hour

Time for one 300 m loop, by load

Level at the light trial load, then human pace falls away at 400 kg

Time for one 300 m loop, by load
CategoryHuman crewRobot
Light trial load3:203:27
Route load, 400 kg5:373:32

Sixty minutes of continuous cycling

19

Trips in 60 minutes, robot running unattended

No rotation and no rest cycle. The pace at the end of the hour is the pace at the start, and it would still be the pace at the end of a ten hour shift.

20

Trips in 60 minutes, two workers alternating

Two people were needed to match one robot at 90 kg, a fraction of the 400 kg the route demands, one working while the other recovered. A single worker could not hold the cycle.

3The exposures that came with doing it by hand

Three risks that only existed because people carried the load

Recorded across the debris route and the store clearance runs at the site. Each one is a consequence of the task being done with human bodies, and each one disappears when the task stops being done that way.

400 kg pushed over slopes and broken ground

The route was not flat concrete. On the sloped and rough sections the force each worker had to sustain rose above the recommended limit for pushing, seven times a shift. No technique or training closes that gap, because the gradient does not negotiate.

Overexertion risk removed at source

Hand unloading on every single trip

Twenty eight minutes of each trip was spent lowering debris out of the trolley by hand, and each load was wrapped before transport to stop spillage. The robot's auto dump does the unload, so both the lowering and the wrapping stop happening.

Repetitive stoop cycles eliminated

Scrap standing for days at a time

Clearing the stores meant pulling a four person crew off productive work, so it kept getting deferred and scrap sat for two to three days, turning working areas into trip and fire hazards. The robot clears on a 44 minute cycle, so the pile never forms.

Housekeeping risk designed out

Manual method, measuredWith the robot, measured

People tied to the route

4, all shift

Four workers on the pallet trolley for 8.9 of 10 hours

None

One part time operator on call

The machine runs the route autonomously

Load behind the crew

400 kg

Four workers to a trolley, over slopes and broken ground

Nothing

The 400 kg sits on the machine

Carries the full load up the slopes unassisted

Distance under load

300 m per trip

Pushed out loaded, then 300 m back empty, seven times a shift

Zero

Nobody walks the route

No pushing, no pulling, no carrying

Unloading

By hand, every trip

28 minutes of lowering debris out of the trolley

Auto dump

Step deleted, not shortened

No lowering by hand and no wrapping beforehand

Clearance interval

Every 2 to 3 days

Deferred whenever the crew was needed elsewhere

Continuous, 44 min

The pile never accumulates

Runs without taking anyone off their own work

Summer rules compress the work, the robot absorbs it

Under the UAE Occupational Heat Stress Prevention Policy, outdoor work in open areas stops between 12.30pm and 3pm from 15 June to 15 September, and daily hours are capped at eight in those months. A manual crew loses that window and has to make it up. A machine cycling at a fixed pace does not.

Removing the task beats managing the risk

Wrapped pallets, handling training and team lifts are all controls layered on top of a task that is inherently over the limit. Taking the task off people removes the need for the controls, which is the top of the hierarchy rather than the bottom.

4Where the four workers went

The point was never fewer people. It was better work.

Four workers came off the debris route and none of them left the site. One picked up the robot as a part time responsibility alongside other duties, stepping in when the machine needs a hand. The other three went to the work they were actually hired for.

Role beforeRole afterWhat changed for them
Trolley crew, 1 of 4Robot operator, part timeGrade upCovers the machine alongside other duties, stepping in when it needs an intervention. Roughly a quarter of a shift, and a skilled responsibility rather than a stamina one.
Trolley crew, 2 of 4Internal fitout supportMoved to the work face on drylining and partition installation. Work that puts progress on the programme and builds a trade skill.
Trolley crew, 3 of 4MEP first fix assistanceSupporting cable and conduit installation with the services teams rather than clearing what those teams leave behind.
Trolley crew, 4 of 4Finishes and flooring supportOn the floors with the finishing teams, on work that is measured by what gets installed rather than by what gets carried away.

Human effort redirected out of haulage

Workers tied to the debris route
4 down to none
Man hours the route consumed per shift
35.5 hours
Part time operator cover per shift
2.5 hours
Man hours returned to productive trades
858 hours per month
Share of the route's physical work now carried by the machine
93%

The same robot, pointed at work that helps rather than work that hurts

Once the debris route is automated, the machine is available for the material runs that currently eat skilled trade hours across the site.

Cable drums and conduit

Heavy drums and conduit bundles taken from the MEP store to the installation zone, instead of two electricians carrying them and arriving tired.

Ductwork and HVAC panels

Prefabricated duct sections and insulation panels moved to the floor, keeping fitout crews on installation rather than fetching.

Drylining and partitions

Stud frames and plasterboard in bulk to the work face. One robot run replaces several two person carries down a long corridor.

Fixtures and switchgear

Boxed lighting, distribution boards and switchgear staged at floor level ahead of the installation team arriving.

Flooring and tiling

Tile boxes, screed bags and adhesive buckets, which are the heaviest and most repetitive carries on any fitout floor.

Plumbing and pipework

Pipe bundles, valves and sanitary ware to shaft areas and bathroom zones, taken off the shoulders of the plumbing crew.

The strain is gone, not shifted

No pushing under load, no carrying distance, no hand unloading. The machine took the parts of the job that wore people down, and it does not hand them to somebody else.

Autonomous, not just assisted

The robot runs the route by itself. One worker covers it part time and steps in when needed, which is a supervision role rather than a haulage one.

The capability comes back

858 man hours a month move from hauling debris to building the data centre, and the crew that used to push a trolley is now on the work the programme is measured by.

See what your site could save

Material Movement Robot, Robotics as a Service. Model your own route with the ROI calculator, or talk to us about a site trial.

Method

Based on 28 trips recorded on site: 12 robot and 4 manual debris trips, 4 store clearance cycles, and 8 head to head load trials. Basis for derived figures: 300 m one way route including sloped sections and stretches of broken ground, 400 kg per loaded trolley trip, 4 trolley workers per trip, part time operator cover at a quarter of a shift, 10 hour shift, 26 working days. Handling limits from the NIOSH lifting equation, ISO 11228-1 and ISO 11228-2. Construction injury shares from the US CDC overexertion analysis of construction musculoskeletal disorders.