A museum glass roof is cleaned for the collection below, not for the view above. Robots are allowed when they run dry or with recovered water, avoid vibration near hanging works, and never introduce cleaning chemistry that can outgas. Scheduled early-morning runs, 03:00-07:00, remain the norm in 2026.
What makes a gallery skylight different?
Daylight control. Many museums diffuse skylight with fabric scrims or louvers. Clean glass changes light intensity, which conservators chart. A dirty pane dropping 8-12% of daylight is sometimes left alone on purpose. So the cleaning plan starts with a lux reading, not a mop.
Then dust. Museums are obsessive about particulates because they settle on textiles and paintings. A wet-scrub robot that splashes on a hot pane and dries leaves mineral spots that later flake down. Recovered-water machines with squeegees reduce that, but a dry microfibre pass is often preferred above galleries.
Can vibration damage hung works?
Usually the vibration from a tracked robot is below the threshold that concerns conservation, but long cables and loose frames on a wall 12 m below can resonate. Check with the registrar. Lingfeng S1 at low speed keeps contact force modest; a heavy tether-plumbed unit can transmit more through the roof structure. If in doubt, move the cleaning window to a closed day.
What chemistry is off-limits?
Anything with strong perfume, ammonia-heavy glass cleaners, or solvent blends that offgas. Museums commonly approve pH-neutral, surfactant-light solutions and deionised water only. Hard water is banned outright – calcium deposits are visually and chemically unwelcome near organic materials.
- Approved: deionised water, pH-neutral mild surfactant, dry microfibre.
- Case-by-case: recovered-water wet scrub with immediate squeegee.
- Avoid: ammonia glass sprays, solvent cleaners, tap water on limestone or textiles galleries.
Does a robot cost more than a rope team here?
Not usually. Museums are small-area, high-restriction jobs, so rope crews charge a premium for method statements and out-of-hours work. A 1,500 m ² gallery glass roof might cost USD 3,000-6,000 per rope visit. A Lingyun Y3 at USD 25,000-40,000 can serve a museum estate for a decade if staff are trained to run it.
Who should keep manual cleaning?
Very small historic skylights, glazing with leaded or stained components, and roofs above open display cases that cannot be covered. For those, a trained in-house technician with a pole system and dry microfibre is the safer answer than any robot.
How should a museum budget the work?
Treat it as a conservation task with a cleaning component, not a maintenance job. Budget for a conservator’s time on the lighting check, out-of-hours access, and a documented method. The cleaning itself is a small line. Museums that skip the lighting review often find the robot is blamed for a change in the collection environment it did not cause.
What about glazing with lead or stained glass?
Leave it to specialists. Leaded panels flex, and any point load or vibration can loosen came and putty. Stained glass adds conservation weight – a single cracked panel may be irreplaceable. For these, use non-contact or remote methods and accept slower, higher-cost cleaning.
Do you need to cover or protect displays below?
Not for a careful robot run, but cover anything directly under the path for the first pass. Dust loosened from a dry microfibre head can fall a long way. Walk the route, look what is underneath, and either move or cover it. This is the same precaution rope crews already take, and it is the kind of detail a registrar will ask about in the method statement.
What records should a museum keep?
The cleaning log plus the lighting readings before and after, and photographs of each glazing zone. If a conservator later sees a change in the collection environment, the record shows whether cleaning coincided with it. Museums that document this well keep the freedom to clean on their own schedule rather than being forced into infrequent, expensive interventions.
How do you phase a large museum estate?
Do the most sensitive roof first with the tightest method, then broaden out. Museums often clean one gallery at a time for conservation and logistics reasons, so plan the robot route around exhibition changeovers rather than a fixed calendar. That means some months are heavy and some are quiet, and the machine should be scheduled to match the building’s rhythm, not the other way round.
What does a realistic atria schedule look like?
Two to four cleans a year suits most glass roofs above collections. More frequent cleaning suits dusty urban sites; less suits sealed, filtered buildings. The right number is the one that keeps measured daylight stable and dust fall low, not a uniform target applied across every building in a portfolio.
Insurance and access are the last hurdle. Museums often require out-of-hours permits, a named technician and coverage for the collection. Confirm all three before the first clean so the run is not cancelled at midnight. Sites that prepare this once rarely have trouble again; sites that improvise tend to lose windows to paperwork rather than to machinery.
Timing also matters for the collection. Cleaning while an exhibition is being installed is often easier than during a quiet display period, because cases are covered and staff are already on site. Plan the run around the install calendar, not the visitor calendar, and you will find more windows than a fixed monthly slot ever gives you.
Key Takeaways
- Measure lux before and after; clean glass changes conservation lighting.
- Dry microfibre or recovered-water scrubbing suits gallery skylights best.
- Deionised water only; tap water and ammonia sprays are off-limits near collections.
- Keep cleaning to closed hours; check vibration risk for hanging works.
- Robots beat rope crews on cost for roofs above about 1,000 m ².
Send gallery drawings and conservation constraints via the contact page. Further notes on heritage glazing are in the news archive.