The Wrong Nut on a 90-Foot Truss: What Pittsburgh’s Convention Center Collapse Still Teaches About Fastener Selection

Summary

On a February morning in 2002, a 90-foot steel bow truss on Pittsburgh’s new riverfront convention center let go. One ironworker died. Two others were hurt. Later reporting and an ASCE forensic case study pointed to a connection error that still shows up on jobsites: the wrong nut type was asked to do the work of a structural anchor nut.

What Happened on Truss 13

The David L. Lawrence Convention Center was the showcase project of its moment: a vast LEED-certified hall along the Allegheny River, framed by a family of long north–south bow trusses. On February 12, 2002, during Phase 3 erection, truss 13 (one of fifteen principal supports) collapsed. Ironworker Paul Corsi Jr., harnessed to the steel, was killed. Two coworkers were injured. Phase 1 still opened on schedule days later, but the industry got a hard lesson about hardware that looks “close enough.”

According to a detailed Penn State failure case summary drawing on contemporary Pittsburgh Post-Gazette and Tribune-Review reporting, the truss connections used large rods supplied for the steel work. The fastener package from Williams Form Engineering included more than one nut style on site:

• Black, heat-treated anchor nuts about 2 inches thick, meant to take the truss load and clamp the joint.
• Silver locking / jam nuts, about ½ inch or 1 inch thick, meant only to back up the anchor nuts so they would not back off.

On truss 13, crews used the silver locking nuts as the primary nuts on the truss bolts. The thick black anchors never did the job they were designed for. When the joint could no longer hold the erection loads, the undersized nuts gave way and the truss fell. Coverage at the time quoted the supplier’s account that the wrong nuts had been used; later technical write-ups, including Lavon and Goodrich’s “Convention Center Steel Truss Collapse” in Forensic Engineering 2009 (ASCE), frame the event as a failed bow-truss connection during construction.

A design change from a compression strut support scheme to a tension strut arrangement made the erection sequence more sensitive. Shop drawings did not give clear, stage-by-stage hardware instructions. Inspection was thin: a foreman hand-checked that bolts “felt” secure and did not verify nut type, thickness, or color against the installation package. Multiple parties later said checking every nut “was not their job.” The result was a critical connection built with hardware that was never rated to be the main nut.

What Went Wrong With the Fastener Selection

This was not a mystery alloy or a hidden corrosion cell. It was a wrong-type / wrong-duty nut error:

1. Wrong role. Jam nuts and lock nuts exist to retain a primary nut. They are shorter. They present less thread engagement and a smaller bearing face. They are not a substitute for a full-height structural nut under tension.
2. Wrong strength path. The black nuts were specified as heat-treated, hardened anchors. The silver pieces were weaker locking hardware. Swapping them swapped the joint’s capacity.
3. Visual cues ignored. The correct and incorrect parts differed in color and thickness. That is rare luck on a jobsite. Still, without a check against the bill of materials, “a nut is a nut” won.
4. Process failure around the hardware. Unclear erection notes, mixed inventory at the workface, and no independent bolt-up inspection meant the error survived until the steel taught the lesson.

If you only remember one mechanical fact: thread engagement and nut height are part of the design. A short nut on a large rod can strip, jump threads, or simply run out of engaged length long before the rod yields. Structural steel practice pairs high-strength bolts and rods with matching nuts (for example ASTM A563 nut grades matched to the bolt specification, or project-specific high-strength systems from the rod manufacturer). The nut is half the connection.

The Correct Choice for That Kind of Joint

For a large truss or tension-rod connection like Pittsburgh’s:

• Use the manufacturer’s primary anchor nut (here, the thick heat-treated black nut) as the load-bearing nut. Seat it fully. Develop the intended clamp and thread engagement.
• Use the jam / lock nut only as a secondary retainer, snugged against the primary nut after the joint is tensioned per the erection procedure. Never invert the stack.
• Match the nut grade to the rod or bolt. Do not mix hardware from “whatever is in the bucket.” If the rod is a proprietary high-strength system, stay inside that system’s nut, washer, and tensioning instructions.
• Control the kit at the point of work. Stage only the hardware for the current connection. Segregate jam nuts from anchor nuts by bin, tag, and color code if the manufacturer already color-codes them.
• Inspect by type, not by touch. A spin-tight check confirms the threads turn. It does not confirm you installed a 2-inch structural nut instead of a 1-inch lock nut. Use the drawing, the packing list, and a go/no-go height or marking check.
• When the support scheme changes (compression to tension, temporary brace removed, load path reversed), re-issue erection notes. Hardware that was “fine” for temporary conditions may be wrong for the new permanent load path.

The same logic applies far outside convention halls. Fleet shops that put a thin locknut alone on a suspension stud, maintenance crews that reuse a short nylon-insert nut as a primary wheel nut, and plant turnarounds that grab “any hex” for a high-pressure flange all repeat Pittsburgh’s mistake in miniature: the secondary fastener becomes the primary load path.

Match the Fastener to Load, Environment, and Application

Every serious joint answers three questions:

Grade marks, finish, and locking style matter. So does which part of the assembly is allowed to carry tension. A jam nut is a tool. An anchor nut is a structural member. Confuse them and the structure votes.

Closing

Paul Corsi Jr. did not die because steel is unpredictable. He died on a connection where the wrong nut type was asked to do a job it was never built for, while the right nuts sat in the same project package unused on that joint. The supplier’s later account, the newspaper record, and the ASCE forensic write-up all circle the same practical rule.

Pick the fastener for the load path you actually have. Use primary nuts for primary loads. Use lock nuts to lock. Write the erection sequence so a stranger on night shift can tell the difference. Then inspect for the difference, not for the feeling that the threads “seem tight.”

The right fastener for the job is not the one that fits the wrench. It is the one the joint was designed to trust.

Sources

• Lavon, B., & Goodrich, A. (2009). “Convention Center Steel Truss Collapse.” In Forensic Engineering 2009 (ASCE). ASCE Library
• Penn State Architectural Engineering failure case study: David L. Lawrence Convention Center Truss Collapse (summarizing Pittsburgh Post-Gazette / Tribune-Review reporting, including Tom Barnes, “Fasteners Blamed in Collapse…,” June 26, 2002)
• Contemporary trade coverage cited in the case study (e.g., Engineering News-Record reports on the inquest and fastener findings, 2002)