A skylight robot cleaning log should record date, operator, area cleaned in m ², water used, battery cycles, faults cleared and photos of each roof zone. In 2026, insurers and ISO 45001 auditors treat that export as evidence of planned maintenance, not as optional paperwork.
Why do auditors ask for roof cleaning records?
Falling-object risk is the reason. A facility that can show dated evidence of controlled glass cleaning has a weaker liability case against it after an incident than one with a single invoice saying ‘cleaning, roofs’. The robot gives you a time-stamped trace almost for free. Rope access teams often hand over one PDF a year.
ISO 45001 clause 7.5 wants documented information controlled and retrievable. ISO 9001 leans on the same idea for maintenance evidence. Neither names robots, but both reward a consistent digital trail.
What should the log capture at minimum?
- Timestamp and duration per cleaning run, down to the zone.
- Area in m ² and water volume, so you can prove coverage and detect drops.
- Operator ID and a short competence note, linking to training records.
- Fault codes and when they were cleared, plus any manual stop.
- Before/after images per zone, ideally geotagged by route segment.
How does this compare with rope-access documentation?
Rope access gives you anchor-point certificates, harness inspection dates and method statements – strong on safety, thin on routine evidence. A robot log flips that. It is weak on human rescue planning and strong on showing that the glass was actually cleaned on a schedule.
What breaks a robot log in an audit?
Three things. First, gaps – a two-month hole with no explanation reads as neglect. Second, mismatched areas: if the roof is 4,000 m ² and every run logs 4,000 m ² at 95% efficiency, the auditor suspects copy-paste. Third, no fault entries at all, which suggests nobody was recording when the machine stopped. A believable log has outliers.
Lingkong K3 and Lingyun Y3 both export run data; check the format before purchase. CSV that opens in a spreadsheet ages better than a locked app dashboard, especially if you change suppliers.
Does insurance care about the machine itself?
Yes. A robot operating above a public area is equipment with a liability profile. Tell your insurer it exists, confirm the CE marking and IP rating, and keep maintenance records for suction cups and tethers. Undisclosed machinery on a roof can void a claim faster than a missed cleaning.
Who should build this trail, and who can skip it?
Any building with public circulation below the glass – malls, airports, hospitals, hotels – needs it. Single-tenant industrial sheds with no public access can keep a lighter log. If you are a contractor bidding on facility contracts, the log is your differentiator; bring an example export to the pitch.
How do you keep the log trustworthy over years?
Assign one person to own the file and export it monthly to a shared drive, not just into the machine app. If the supplier changes or the app subscription lapses, a CSV survives and a dashboard does not. Pair each run with the cleaning work order number so finance, safety and operations all point at the same record.
What does a good log look like at year end?
A twelve-month export with a clean monthly pattern, a handful of documented faults each resolved, and area totals within a few percent of the measured roof. That is believable. A perfect twelve-month run with zero faults and identical numbers each time is what auditors push back on.
Does the log help with warranty claims?
It can. If a suction cup or motor fails inside warranty, a dated record of hours and fault codes supports the claim. Without it, the supplier cannot tell whether the machine ran 50 hours or 500, and warranty decisions default to the cautious answer. Keep the export in the same system as your other asset records so it is not orphaned.
What is the single most useful field?
Area per zone. It ties the log to the drawing, shows coverage over time, and exposes both over-claiming and under-cleaning. Water volume is a close second, because a sudden jump usually means a worn squeegee or a leak, which is a maintenance signal long before anything breaks visibly.
How does a robot log compare with a BMS or CMMS record?
It feeds one. The cleaning export is a source, not a destination. Push the run summary into the CMMS as a completed work order and the roof cleaning becomes part of the same asset history as the lifts, chillers and pumps. When something fails on the roof, the whole record sits in one place rather than in a machine app nobody else can open.
What is the biggest mistake first-time users make?
Recording cleaning but not conditions. Wind speed, glass temperature and rain tell you why a run was short or why a zone was skipped. Notes that only say ‘cleaned atrium’ are nearly useless six months later when someone asks why a section was missed. Two extra fields, wind and weather, turn a bare log into something you can defend.
One caution: do not let the log become a marketing document. Real logs have missed cleans, short windows and operator notes that read like a human wrote them. A tidy file with no anomalies invites scrutiny. The value is in the pattern over a year, including the messy parts, not in a polished summary nobody can verify.
Key Takeaways
- A robot log proves planned maintenance; an invoice proves almost nothing.
- Capture timestamp, m ², water, operator, faults and photos per zone.
- Gaps and perfect-looking area totals are the two fastest ways to fail an audit.
- Disclose the machine to your insurer and keep cup and tether records.
- Public-circulation buildings need the full trail; private sheds can keep it light.
To see sample log formats and machine data exports, reach us via the contact page. Related notes sit in the news archive.