I have walked into plants where a six-year-old double-wall system still looked like it was commissioned last month, and others where the same brand of duct was already shedding liner fibers into the airstream. The difference was almost never the duct. It was what happened after handover: the moisture events nobody reported, the cleaning crews that used the wrong tools, the inspections that never happened. Once you understand what the liner is actually doing inside that outer skin, the maintenance logic becomes obvious.
Double-wall duct is essentially two concentric shells with the liner doing three jobs between them. It insulates, so the supply air does not sweat on the outer wall in a humid ceiling plenum. It absorbs sound, which matters enormously in an office tower or a hospital ward where the duct runs exposed. And it protects the inner wall from mechanical abuse during installation. A spiral tubeformer producing 100 to 150 mm thick double-wall sections can hold concentricity within a millimeter or two, but none of that engineering survives if the liner gets soaked, crushed, or loaded with dust.
The failure modes are boring, which is exactly why they get missed. Nobody calls you when the liner is slowly compacting. They call you when the office below the duct run is two degrees warmer than the drawings promised, or when the acoustician re-measures and finds NC levels eight points over spec.
In my experience the single biggest killer of duct liner is water, and it almost never comes from the duct itself. It comes from roof leaks dripping onto the outer jacket, from condensate drain pans overflowing above the duct run, or from a construction phase where the system ran without filters and humid outdoor air condensed on cold inner walls overnight. Fiberglass liner that gets wet loses maybe a third of its thermal resistance, and worse, it can stay damp inside the outer shell for weeks because nothing in that cavity promotes drying. By the time anyone notices a smell, you are dealing with a microbial problem, not an insulation problem.
I tell every contractor I work with: during construction, temporary caps on duct ends are not optional. Open double-wall spirals sitting in a rain-exposed riser shaft will drink liters of water before the system is ever balanced. One project in a coastal Southeast Asian plant lost forty sections this way and the replacement cost landed at roughly fifteen percent of the duct contract value.
Liner exposed to unfiltered air loads up with dust at the leading edge, typically the first two or three meters downstream of any AHU or branch takeoff. The fibers act like a filter media, which nobody designed them to be. A millimeter-thick dust cake on the airstream surface adds real airflow resistance and, if the liner has a coating, that coating gradually gets abraded away where the velocity is highest. In systems running above 12 m/s, I have seen coated liner surfaces erode enough to expose raw fiber within three years.
Every hanger point is a compression point. If the outer shell is clamped hard against a strut with no spacer, the liner between the shells compresses locally and its R-value drops to nearly nothing in that ring. The same thing happens where cable trays are strapped directly across a duct. None of this is visible from outside, which is why a borescope inspection matters more than people expect.
Cut ends of liner, especially where sections butt together at a joint collar, are where fibers start to migrate. A poor quality factory joint or a field-cut section without edge treatment will shed fibers into the airstream at every pressure pulse. The trend toward closed-cell foam liners or factory-sealed edges exists because of this exact problem.
Here is where I watch people make expensive mistakes. The instinct is to blast the duct with compressed air or run a rotating brush, and both of those can do more harm than the dust they remove.
For routine dust loading, contact vacuuming with a HEPA-filtered machine through access doors is the right first move. Soft-bristle agitation only where the vacuum head cannot reach. Compressed air just pushes dust deeper into the fiber and, at high pressure, lifts the liner coating off in patches. If the system has never been cleaned, expect the first pass near the unit to pull out a surprising amount of material; on one hospital retrofit we filled three HEPA vacuum canisters from the first thirty meters of supply duct alone.
If there is microbial growth or sticky residue, you can use a low-moisture foam or a very lightly applied sanitizer, but the rule is simple: the liner must be dry again within 24 hours, and you need to know why it got dirty in the first place. Cleaning a liner without fixing the moisture source is a subscription service, not a repair. Fans running during and after cleaning help enormously with drying, and a moisture reading at the liner surface before closing everything up is cheap insurance. Below roughly 16 percent surface moisture content on fiberglass, you are safe; above that, wait.
None of this works if you cannot get into the duct. When we set up a double-wall line for a client, I push them to specify cleanout doors at every branch takeoff and at least every fifteen meters on trunk runs. The doors cost almost nothing at fabrication time and save you cutting into the outer shell later, which is a job nobody wants because the repair patch becomes the new weak point for moisture ingress.
You do not need to inspect everything every year. What follows is the schedule I actually give clients, adjusted by environment: coastal or high-humidity sites tighten every interval by about a third.
| Interval | Action | Tool / method | Pass criteria |
|---|---|---|---|
| Every 6 months | Visual at access doors, first 3 m downstream of AHU | Torch and mirror | No dust cake over 1 mm, coating intact, no standing moisture |
| Every 12 months | Borescope through hanger points and clamp rings on a sample of 10% of the run | Video borescope, 2–3 m reach | Liner thickness at clamp within 20% of nominal; no compression rings |
| Every 12 months | Outer jacket walk-down: seams, hangers, condensate lines above | Visual plus photo log | No staining, no fastener corrosion, hangers with spacers intact |
| Every 24 months | Thermal scan of exposed runs under load | Infrared camera, ΔT between supply air and ambient at least 8°C | No warm bands indicating compressed or wet liner sections |
| Every 36 months | Full internal vacuum clean of trunk duct, edge reseal where fibers are exposed | HEPA vacuum, soft agitation, mastic-compatible edge sealant | Before/after photos filed; pressure drop across section within 5% of baseline |
The thermal scan deserves a note. It is the cheapest early warning you will ever get for water damage, because wet liner changes the surface temperature pattern of the outer jacket in a way that is obvious on a scan but invisible to the eye. A €1,500 camera has caught problems on projects that would otherwise have shown up as mold complaints two winters later.
Localized damage has a genuine repair path. A compressed section at a hanger can be relieved by re-hanging with a spacer, and the liner will mostly recover if it was not crushed past its elastic limit. Small areas of fiber exposure can be re-coated with the same polymer coating the factory used, as long as the surface is dry and vacuumed first. I keep the repair threshold conservative: if more than about one square meter in any ten-meter section shows coating loss or fiber exposure, replace the section. Patch economics look good on paper and then the next inspection finds the same problem three sections downstream, because the coating was failing everywhere and you only treated what you could see.
Water-soaked liner is not a repair candidate, full stop. Once fiberglass has been saturated and stayed damp long enough to smell, you are removing the section. The outer spiral shell is usually salvageable, which softens the cost: a crew can strip the damaged liner, inspect the inner and outer walls, and a fabricator can supply a re-lined section or liner roll to refit on site. On one retrofit we recovered roughly 70% of the affected outer shells this way and the bill came in under half of full replacement.
Most of what I have written above becomes unnecessary work if the original spec was right. A few choices matter more than the rest.
First, coated airstream surfaces. The difference between a mat-faced liner and one with a factory coating is the difference between a surface that holds dust and one you can vacuum clean for a decade. Second, edge treatment at joints: closed edges or factory collars stop fiber migration, which is both an indoor air quality issue and a maintenance cost. Third, density: a 48 kg/m³ liner resists compression and airflow erosion far better than the 24 kg/m³ product that wins a tender on price. And fourth, ask how the fabricator's auto duct production line handles liner: a proper double-wall line with controlled gap and bonded liner produces consistent sections, while a shop wrapping insulation by hand between shells produces variable compaction you will spend years paying for. When I compare suppliers, I ask to see their double-wall output, not their brochure.
It also pays to think about the whole production chain. Shops that run their own spiral and fitting equipment, including a beading machine stage for stiffening the outer jacket, tend to deliver cleaner section geometry, and clean geometry means hangers sit right and liner compression is rare. This is one of those quiet connections between fabrication quality and fifteen years of maintenance cost that never shows up in a quotation comparison.
Whatever schedule you adopt, photograph everything and file it by section number. The single most useful document in a duct dispute is a dated photo of the liner at an access door from two years before the complaint. Plants that keep these records settle warranty questions in days; plants that do not pay for lawyer time instead.
Can we vacuum clean liner without opening the duct at all? Only where the system was built with proper access doors or where the connection collars give usable openings. Roto-brush trucks marketed for sheet metal duct will damage an unprotected fiberglass liner surface; if you hire a cleaning contractor, require soft-bristle, HEPA-contained methods and a written method statement before they start.
How do we tell water damage from normal discoloration? Discoloration from dust is uniform along the airflow path and wipes or vacuums off. Water damage shows as tide lines, local darkening that follows gravity, and often a smell that returns after rain. If in doubt, take a surface moisture reading; anything persistently above 16 percent on fiberglass means the section needs opening up.
Is foam liner easier to maintain than fiberglass? Closed-cell foam resists water better and does not shed fibers, which removes two failure modes. It costs more per meter and its acoustic performance at low frequencies is weaker than dense fiberglass, so in noise-critical projects we still specify fiberglass with coated faces and sealed edges, then rely on the inspection schedule above.
What should we demand from a new double-wall supplier on this topic? Three things: the liner specification in writing (density, facing, edge treatment), access door locations shown on the fabrication drawings, and a recommended maintenance schedule stamped by the supplier. A manufacturer who will not put a maintenance recommendation in the offer is telling you they have never stood next to their duct five years later.