Someone walks the aisle and finds it: an upright leg bowed out where a forklift caught it, or a beam that no longer looks straight. The rack is loaded, the warehouse is running, and the question arrives immediately — is this something we fix, something we replace, or something we can live with?

It is a real decision with a real cost either way, and it is usually made badly, in one of two directions. Either the damage is photographed, discussed and left in place because the rack "looks fine", or the whole run is condemned when a single frame needed changing. Here is how the decision is actually made.

First, before anything else: unload it

A damaged upright has lost capacity, and nobody standing in the aisle can say how much by looking. The correct immediate action is to unload the affected bay and the bays either side of it, take it out of service, and mark it so nothing goes back in while the decision is being made.

That sounds obvious and it is routinely not done, because unloading two or three bays in a full warehouse is genuinely disruptive. But the load path in a selective rack runs down the uprights, and a compromised leg carrying pallets is the condition under which racks come down. Every collapse investigation reads the same way: the damage was known about.

Why a bent upright is not like a bent beam

These are different problems and they deserve different reactions.

A beam is designed to bend. It carries a uniformly distributed load and deflects under it, and the rating assumes that. What matters is whether it returns to straight when unloaded. A beam that stays deformed after the pallets come off has yielded — it is finished, and it is also a cheap and quick component to change.

An upright is a compression member. Its capacity depends on staying straight and on being properly restrained by its bracing at intervals. A kink or a bow does not reduce capacity in proportion to how bad it looks — it changes how the member buckles, which is a step change rather than a gradual one. This is why a leg with modest-looking damage can be a genuine problem while a beam with a visible sag is a Tuesday-morning replacement.

The other thing to check on an upright is the base plate and its anchors. A frame struck hard enough to bow a leg has often lifted or distorted the base plate too, and a base plate that is no longer flat on the slab is not transferring load the way the design assumed.

The three outcomes, and what decides between them

Rack inspection under SS EN 15635 sorts damage into escalating levels of action rather than a simple pass or fail: damage light enough to be monitored and re-checked, damage serious enough that the bay must be unloaded and the component replaced within a defined period, and damage requiring immediate offloading and isolation. The point of the system is that "keep an eye on it" is a legitimate outcome only for the lowest category, and only with a date attached and a record kept.

In practice the decision falls out as follows.

• Replace the component. This is the normal answer for a damaged upright leg or a permanently deformed beam. Components are replaced like for like — same make, same section, same connector — which is why knowing what you have matters and why an unidentified system is harder and slower to put right.
• Repair. There are engineered repair systems for uprights, and where the manufacturer supports one for that profile it can be a sound answer, particularly where a frame is loaded and awkward to change out. What is not a repair is a fabricator welding a plate over a kink. Field welding on a cold-formed upright changes the steel's properties locally and voids the load rating it was tested to — it makes the rack look mended without making it safe.
• De-rate. Reducing the rated load on a run is a legitimate engineering answer in specific circumstances, but it only counts if the new figure comes from someone competent to calculate it, the load notice is changed to match, and the people loading the rack are told. A de-rating that lives in an email is not a de-rating.

Replacing a leg in a live warehouse

Changing an upright is not the same as changing a beam. The frame is carrying load through the bracing to the leg being removed, so the run has to be unloaded around it and, depending on the configuration, temporarily supported while the component is swapped. It is skilled work with a sequence, and it is done with the bays empty.

Where the same run has taken damage in several places, replacing the whole frame is often quicker and cheaper than three separate component swaps, and it gives you one new component specification rather than a patchwork. Our published per-bay ranges give a sense of the scale involved — selective racking runs S$380 to S$1,200 per bay depending on the load class, with V3000 at S$380–S$650, V5000 at S$650–S$1,000 and V8000 at S$900–S$1,400 — but a repair or replacement in a running warehouse is quoted after a survey, because the access and the sequencing drive it as much as the steel.

Fix the cause or you will do this again in six months

Damage clusters. It appears in the bottom metre of uprights, at the aisle ends, at the entry bay of every run, and next to dock doors — because that is where forklifts turn, reverse, and misjudge. If the same three positions keep getting hit, the answer is not a stronger upright.

Things that actually reduce recurrence: column guards and end-of-run barriers at the positions that get hit rather than everywhere, an aisle width that suits the truck being used today rather than the one specified years ago, lighting at the aisle ends, and drivers who report an impact without expecting to be blamed for it. That last one is worth more than the hardware. An unreported strike is discovered by an inspection weeks later, or by nobody. We wrote about the patterns forklift damage falls into and why they are usually a training and layout problem rather than a driver problem.

Write it down, every time

Damage found and acted on should leave a trail: where it was, what it looked like with a dated photograph, what category it was assigned, what was decided, who did the work, and what the load notice says now. Two reasons. The first is that a pattern is only visible in the record — one damaged upright is an accident, the same bay three times is a layout problem. The second is that after an incident, the record is the difference between a system that was managed and a system that was ignored.

If your racking has no load notices and no inspection arrangement to hang this on, that is the more urgent job. Our note on what an installer should hand over covers what should exist, and it is all reconstructible for an inherited system.

The short version

Unload first and decide afterwards. A deformed beam that does not spring back is replaced. A bowed or kinked upright is a compression member that has changed behaviour, not just appearance — replace the component, or use an engineered repair the manufacturer supports, and never a welded plate. De-rating is valid only when it is calculated, signed, and reflected on the load notice at the rack. Then fix the thing that caused it, and write all of it down.

Related reading: what to check before buying used racking, why safety pins and locking clips decide whether a rack survives a knock, and the handover documents that tell you what your rack can carry.

Talk to us

Send a photograph of the damage and tell us what the system is, and we will tell you whether it is a component change, an engineered repair or a run that needs re-rating. Message us on WhatsApp, or see racking systems and our services, including SS EN 15635 inspection.