I coordinate emergency steel fabrication and erection for industrial projects. I've handled 60+ rush orders in 8 years, including same-day turnarounds for warehouse contractors. The call I got on March 15, 2024, at 6:40 a.m. was one of the close ones.
A general contractor was 36 hours from a final inspection on a 48,000-square-foot steel structure warehouse in northern Indiana. The building was mostly up: a rigid frame structure with 80-foot clear spans, a smaller space frame entrance canopy, and roof purlins already staged. The project had a $50,000 penalty clause if the inspection slipped past Monday at 7:00 a.m. Normal punch-list and welding repair turnaround on a job like that is five to seven days. We had a weekend.
The red flag was not the weld. It was the drawing.
When I got to the site at 8:15, the superintendent walked me to the first problem. The local inspector had flagged 12 beam-to-column moment connections on the rigid frame. Some welds were undersized. A few had lack of fusion. That alone was bad enough. Structural welding on primary steel is not a place for guessing. Under AWS D1.1/D1.1M, the welders have to be qualified for the positions and thicknesses they're running, and the procedure has to match the approved WPS.
But then I looked at the purlins. The roof drawings called for Z purlins with laps at the interior supports. The yard had a mix of C purlins and Z purlins. The space frame canopy called for a different node detail than what was on the shop drawings. That was the real red flag.
If you've ever managed a steel package, you know that types of purlins are not interchangeable just because they both look like folded steel. C purlins work well for simple spans. Z purlins can nest and lapped to create continuous runs, which changes deflection and load paths. Sigma purlins have their place too, especially where you want deeper section efficiency. Pick the wrong profile or lap detail, and the roof diaphragm may not behave the way the engineer modeled it.
We pulled the revision log. The fabricator had used a 2023 set that predated a 2024 revision. The revision changed the space frame connection and increased the rigid frame structure's moment connection demand near the ridge. The old drawings were still in the field trailer. Someone had stapled the new cover sheet on top of the old package. Classic. Frustrating. Expensive.
Risk weighing at 10:30 a.m.
The upside of a quick field fix was clear: keep the inspection date, avoid the $50,000 penalty, and keep the client's trust. The risk was equally clear: if we welded over defects without engineering approval, the inspector could fail the whole frame. Then we'd be cutting out welds and maybe replacing members. Worst case: complete redo at about $18,000 in steel and labor, plus the penalty, plus the general contractor never calling us again. Best case: a $7,000 to $10,000 repair weekend and everyone moves on.
I kept asking myself: is saving a few thousand dollars worth potentially losing a $1.2 million contract? The expected value said fix it right. The downside felt catastrophic.
Normally I'd get three quotes and a second engineering opinion. We had two hours to decide before the mobile welder had to commit to another job. So we made the call with incomplete information. We stopped all purlin installation. We called the engineer of record. We called our certified welding partner. We called an NDT tech who could do ultrasonic testing on a Saturday night.
In hindsight, I should have pushed back harder on the fabricator when they said the drawings were 'close enough.' But with the superintendent waiting and the penalty clock ticking, I did the best I could with what we had.
The fix: grind, weld, test, repeat
The engineer of record issued a repair procedure by 1:00 p.m. It was fairly specific:
- Remove zinc coating at the weld zone before welding. Galvanized steel can be welded, but you need ventilation and a repair coating after.
- Grind out the defective welds to sound metal.
- Weld per the approved WPS with qualified welders using low-hydrogen electrodes.
- Ultrasonic test the repaired moment connections.
- Re-coat with zinc-rich paint.
For the purlins, the fix was not glamorous. We swapped the wrong C purlins out of the rigid frame roof bays and installed the specified Z purlins with the correct laps. On the wall girts, where the design used simple spans, the C purlins were acceptable. At the space frame canopy, we had to cut the wrong node plates and bolt in the revised connection. That was the part that almost broke us.
At 2:10 a.m. Sunday, the NDT tech called me over. One of the repaired welds still showed an indication. Ugh. Again. The welder had been at it for 11 hours. The engineer wanted a stiffener plate added to that connection. We did not have one fabricated. We found a pre-cut stiffener in the fabricator's yard—thankfully, because that was a one-click-away-from-disaster moment. We fitted it, welded it, tested it, and passed at 4:40 a.m.
Bottom line: we finished at 6:10 a.m. Monday. The inspector arrived at 7:00. He walked the frame, checked the weld maps, looked at the NDT reports, and signed off. The general contractor avoided the $50,000 penalty. We paid about $11,800 in rush labor, $6,500 in engineering and NDT, and maybe $3,200 in re-coating and consumables. Not cheap. But we saved the project.
So glad we stopped work when we did. Almost let the crew keep installing purlins over a connection that was going to be cut open. That would have turned a weekend repair into a two-week teardown.
What I learned after 150+ rush orders
It took me three years and about 150 rush steel orders to understand that document control matters more than welding speed. A fast welder cannot fix a wrong revision. A cheap purlin package is not cheap if it is the wrong profile. In steel structure warehouse work, the sequence is unforgiving: design, fabrication, welding, erection, inspection. If one link uses old information, the whole chain pays.
Five years ago—back in 2019—most of our warehouse projects were straightforward rigid frame structures. Today we see more space frames, hybrid canopies, and mixed rigid frame and space frame systems. The fundamentals have not changed. Steel still moves with heat. Welds still need qualified procedures. Purlins still need the right profile and lap detail. But the execution has transformed. Models get updated faster. Fabricators cut from files, not from paper. That should reduce errors, but only if the field is actually using the current revision.
According to AWS D1.1/D1.1M, structural welding requires qualified welders and approved procedures. AISC 360 covers the design side. The MBMA Metal Building Systems Manual is still a solid reference for metal building systems. None of that helps if the crew is building from a superseded drawing.
Here's what you need to know if you are buying or managing a steel structure warehouse:
- Check the revision number on every drawing before fabrication and again before erection. Do not trust the cover sheet. Look at the title block.
- Confirm the purlin type in writing. C, Z, sigma—each has a different load path and connection detail.
- Require the WPS and welder qualifications before the first arc. Structural welding is not a generic skill.
- Inspect galvanized repairs. Grinding zinc off is normal; leaving it uncoated is a corrosion problem later.
- Build in buffer time. The rush fee is usually worth it for deadline-critical inspections, but it is not a substitute for correct documents.
We kept the specified Owens Corning Thermafiber insulation under tarps during all of this. Mineral wool does not like getting soaked before the envelope is closed. That part of the spec was fine. The steel was the problem.
Trust me on this one: if you are 36 hours from an inspection and someone says the old drawings are close enough, that is not a small issue. It is a deal-breaker. Slow down for two hours, get the engineer on the phone, and fix the revision. It is a no-brainer compared with explaining to your client why the roof is coming down.