The Rack That Leaned After the Pallet Jack Hit It

What happened

A plumbing and water heater shop kept its tank stock on four bays of selective pallet rack along the back wall: 12-foot frames, three beam levels per bay, boxed water heaters standing on pallets. On a Monday, a helper pulling a loaded manual pallet jack around the end of the aisle clipped the front upright of the end bay at floor height. There was a bang, the bay shook, and when it settled the helper thought the frame leaned "a little" toward the aisle. Two people looked at it, agreed it was not much, and went back to work. Nobody unloaded it and nobody told the shop manager.

Eight days later, during an ordinary morning, the aisle-side beam on the second level of that bay dropped at its end on the struck frame. Six boxed water heaters slid off the pallets and fell about 5 feet into the aisle. Nobody was in the aisle. Two heaters were scrapped, three had dented jackets the supplier would not take back, and one survived.

The shop manager's first question was whether the rack had been overloaded. It had not been, and working out what had actually happened took most of a day.

Before anyone looked for a cause

The lead who heard the crash did three things before anyone went near the bay, and these are the right three.

• Closed the aisle. Cones and a rope across both ends, and a person posted at the front until the shop manager arrived. A bay that has dropped one beam has shown that its frame or connections cannot be trusted, and the levels still loaded above and beside it can follow.
• Left the fallen heaters where they lay. Nobody walked into the aisle under the loaded third level to drag cartons out. A water heater in its box weighs on the order of 100 to 150 pounds, and the third level still held more of them.
• Unloaded from the top, by forklift, with the aisle empty. The shop manager had the trained forklift operator take the third-level pallets off the damaged bay first, then the remaining second-level load from the adjacent bay that shares the struck frame, with no person in the aisle or behind the rack while each pallet came out. Only then were the fallen cartons cleared. Had the frame looked any worse, the right call was to leave it loaded and call the rack supplier to unload it.

Candidate one: too much weight

The obvious explanation for a falling beam is overload, and it was the first thing checked.

The difficulty was that the shop could not answer it, because the bay had no load plaque. Nobody knew the rated capacity of the beams. The rack had been bought used and installed by the previous tenant.

The shop manager photographed the beam's manufacturer marking and its connector, measured the beam's face height and length, and sent them to a rack dealer, who identified the beam and gave its rated capacity per beam pair in writing. The six heaters on that level weighed roughly 800 pounds together, pallets included, from the supplier's shipping weights. That was a small fraction of the rated capacity of the pair.

So overload was out. It was also the first finding: a rack whose capacity is unknown cannot be loaded correctly by anyone, even when it happens to be. See related: Rack Load Ratings and the Label Every Bay Needs.

Candidate two: a failed beam

The dropped beam was next. If the beam had cracked, bent under load or had a weld fail at its connector, that would explain the drop at one end.

The beam was straight end to end and had no deformation at mid span. The welds where the connector bracket joins the beam were intact. The connector's hooks, the tabs that drop into slots punched in the upright, were undamaged but the connector had come completely out of its slots at the struck frame.

And there was no beam lock. A beam lock (also called a safety clip or safety pin) is a small device at each connector that stops the beam from lifting up out of the slots. Without one, a beam is held down only by gravity and the weight on it. The beam's other end, at the next frame, still had its lock. The end at the struck upright had none, and there was no sign one had ever sheared off.

The shop then checked every beam end in all four bays. Of 48 beam ends (4 bays, 3 levels, 2 beams per level, 2 ends per beam), 11 had no lock. The previous tenant's installer had either run short or the locks had been lost as beams were moved. The beam was not the cause, but the missing lock was a precondition: it was the reason a connector could come out at all.

Candidate three: the frame and the floor

That left the question of what lifted or shifted the connector at the struck upright after a week of standing still.

The struck upright had a visible dent at about 10 inches from the floor and a slight bow along its first few feet. Measured with a 4-foot level held against the upright and a plumb line from the top of the frame, the frame was out of plumb by about 1 1/4 inches over its 12-foot height, toward the aisle. The commonly used limit for pallet rack in service, from the Rack Manufacturers Institute's guidance, is half an inch for every 10 feet of height: for a 12-foot frame that is 0.6 inch. This frame was at roughly twice the limit.

Then the base plates. Each upright stands on a footplate meant to be anchored to the slab with bolts. Two footplates in that bay had no anchors: the struck front upright and the rear upright of the same frame. The holes in the plates were empty. The other uprights in the row were anchored.

With that, the sequence explained itself:

1. The hit pushed an unanchored frame. With no anchors, the jack's blow did not just dent the upright, it slid the base of the frame a fraction of an inch and twisted it, and the dent then let the upright bow. The frame leaned.
2. The lean pulled on the connectors at the struck upright. A frame leaning toward the aisle changes the angle between the beam and the upright slots, and the connector on the struck upright sat partly lifted rather than fully seated.
3. A week of picks walked it out. Every pallet placed on or taken off that bay, and on the adjacent bay sharing the frame, gave the connector a small jolt. With no lock to hold it down, that end of the second-level beam rose a little more with each one until the hooks cleared the slots.

The pallet jack started it. The missing anchors let the hit move the frame, the missing lock let the beam come out, and the eight days of normal use did the rest.

The inspection record the shop produced

The shop manager filled one line per bay before any of the four bays was reloaded.

RACK DAMAGE CHECK     after incident, all 4 bays
BAY  PLAQUE  PLUMB 12FT  ANCHORS  LOCKS   UPRIGHT
 1   none    0.2 in      4 of 4   9/12    ok
 2   none    0.3 in      4 of 4   10/12   ok
 3   none    0.25 in     4 of 4   10/12   ok
 4   none    1.25 in     2 of 4   8/12    dent, bow
LIMIT        0.6 in      all      all     no dent
BAY 4 AND SHARED FRAME: OUT OF SERVICE, REPLACE

Bay 4 is the struck end bay. Locks missing across the four bays add to 3, 2, 2 and 4, which is the 11 of 48 found.

What changed

• The damaged frame was replaced, not straightened. The rack dealer supplied a new frame of the same make; bent uprights are replaced with the manufacturer's part or repaired only with a repair kit the manufacturer or a rack engineer approves. Heating and bending a column back is not a repair.
• Every footplate anchored. All four bays were checked and the two empty plates anchored to the slab with the anchor type and size the rack manufacturer's installation instructions call for. See related: How to Anchor and Inspect Pallet Rack and Shelving.
• Beam locks on all 48 ends. The 11 missing locks were replaced with the manufacturer's locks, and a beam lock check became a line on the monthly safety walk.
• Column guards and an end-of-aisle guard. A steel column guard bolted to the floor at each upright on the aisle side, and a guard rail across the end of the aisle where the jack turns. The hit happened at the one spot every pallet movement passed.
• Load plaques on every bay, as ANSI MH16.1 calls for, with the dealer's written rating per beam pair, so the next overload question can be answered by reading.

The rule that would have prevented it

The one change that would have stopped this on its own was not hardware. It was a rule the shop now posts at the end of the aisle: any time anything hits the rack, the bay is reported to the shop manager the same day, and the struck bay and any bay sharing its frame are unloaded before anything else is put on them or picked from them, until the frame has been checked against the plumb limit and the dent criteria.

The two people who looked at the lean on Monday were not careless. They had no rule telling them what "a little" meant and no reason to think it would get worse while standing still. It got worse because a rack is not standing still in a working shop: it is loaded and unloaded dozens of times a week, and each of those is a small push on a connection that has already been disturbed.

Two other conditions would have changed the outcome. A bay with every lock in place and every footplate anchored would most likely have taken the same hit with a dent and no lean; the shop would still have needed to replace the upright, but nothing would have dropped. And had the struck bay held gas water heaters on its top level rather than its middle one, the fall would have been from about 9 feet onto the jack path rather than 5 feet into an empty aisle.

References

• ANSI MH16.1, the specification for the design, testing and utilization of industrial steel storage racks, which binds rack manufacturers and is the reference for anchoring and load plaques.
• Rack Manufacturers Institute guidance on rack inspection and damage, including the out-of-plumb limit used above.
• See related: Rack Load Ratings and the Label Every Bay Needs; How to Anchor and Inspect Pallet Rack and Shelving.
• See related: The Monthly Shop Safety Walk SOP.