That Half Hour Roping Up on the Roof Never Makes It Into Cleaning Efficiency

A friend works at a curtain wall maintenance company, and last month he told me they'd brought in a facade cleaning robot and asked me to come see whether it was any good. When I got to the site, sure enough, it was a machine actually working on the wall.

The machine was bigger than I expected—one person can't lift it. Two ropes go down from the rooftop, one load-bearing and one as a safety backup, and the machine walks down clinging to the glass surface. Adhesion is by negative pressure—simply put, a fan continuously pumps air out of the machine's interior, pressing it against the glass via atmospheric pressure, and then it walks across the facade. I'm familiar with this principle; wall-climbing robots basically all use it.

The operator stands on the rooftop holding a remote, and the screen shows the machine's posture and position. It moves in a Z pattern, sweeping row by row from top to bottom.

I watched for about twenty minutes and noticed the speed isn't constant. On flat glass it moves quite smoothly, but at mullions and operable window frames it clearly slows down, and the operator has to tap a few times manually to get it over. The materials say it supports obstacle crossing; from what I actually saw, it can cross obstacles, but not automatically and smoothly—there's human intervention in between.

I know this all too well. When we do AGVs crossing thresholds and speed bumps, "rated to cross" and "smoothly crossing every day" are two different things. The former is a demo; the latter is a product.

On efficiency, the official cleaning rate is over three hundred square meters per hour. The number looks impressive on its own, but I'm calculating a different account: from laying ropes on the rooftop, hanging the machine, debugging, cleaning, reeling in the ropes, then moving to the next facade—of the whole process, how much time is actually spent cleaning on the wall. My friend says that for buildings with regular facades, the daily throughput is clearly more than pure manual labor; but for buildings with lots of decorative lines, dense operable windows, and irregular corners, prep work eats up a big chunk of time.

Safety is this machine's most solid value. A spider-man hanging two hundred meters up, wind and sun—one accident is a big deal. With the machine on the wall, the person retreats to the rooftop; that account doesn't need careful calculation. But conversely, if the machine gets stuck high up, loses power, or has a rope problem, how to get it back—someone has to think through that contingency plan in advance. My friend says they have to stop when it's windy; there's a wind speed limit. I believe it, and that's the only way to do it.

There's one more thing I didn't see fully on site but the materials mention: besides cleaning, it also does exterior wall inspection, using a camera to record the curtain wall's condition. I think this part is underrated. Cleaning is an expense; inspection is an asset. Glass spontaneous breakage, sealant aging—if every cleaning leaves a set of images, over a few years you can see trends. The premise is that this data can actually be archived and retrieved, not just sit on a hard drive after shooting with no one looking.

A few downsides too. Prep time is long—laying ropes, reeling ropes, moving between facades—two people is the minimum configuration. Obstacle crossing still needs someone watching; put it up and forget about it—that's not possible yet. The rooftop needs a place to anchor the ropes; some old rooftops are equipment floors without that condition. And then there's the weather window—wind and rain mean stopping, same as manual labor, living at the mercy of the sky.

I wrote a post before saying that the boundary of full automation is how long you can go without checking on it and still not have an accident. This machine isn't there yet; it's more like a semi-automatic tool that takes over the high-risk actions.

So the conclusion is: it depends.

Suitable for property management or maintenance companies that own their buildings, have large curtain wall areas, and have a stable maintenance crew. The equipment cost can be amortized, people retreat from high altitude to the rooftop, and both the safety account and the labor account work out. Not suitable for a single small building, for per-job outsourced billing, or for projects with lots of irregular curtain walls. Fixed costs can't be spread out, and prep time takes too high a proportion—better to keep using manual labor.

My friend asked me whether it's worth adopting. I said first calculate three numbers: how many times a year you clean, how many square meters each time, and whether the rooftop can anchor ropes. Once those three numbers are clear, the answer comes out on its own.

To calculate efficiency, you have to start from the moment ropes are laid on the rooftop—just counting the moment the machine is on the wall isn't enough.