A series of design and construction errors, including use of an improper expansion joint connection and the wrong type of steel, caused the collapse of a section of flooring at the David L. Lawrence Convention Center a year ago today, engineers have concluded.

In a presentation yesterday, officials with Wiss Janney Elstner Associates, a prominent national engineering firm, said the collapse of the 30-foot-by-60-foot section of loading dock floor at the convention center had its genesis in the type of expansion joint connection used and was aided by problems with materials, fabrication and construction.

"There were a number of design and construction defects that led to the collapse," said Andrew Osborn, senior principal for Wiss Janney.

All contributed to the mishap last Feb. 5, which occurred when the expansion joint connection locked up, preventing a support beam from contracting properly in the frigid temperatures that day.

That, in turn, put too much stress on the welds of the L-shaped connection plates, or angles, used to anchor the beam to the expansion joint girder, causing the welds to separate. When a tractor-trailer drove over the spot, it created enough weight to cause the beam and floor to drop.

Although no one was injured, the collapse forced the convention center to shut down for more than a month, causing postponement and rescheduling of the Pittsburgh International Auto Show, a delay in the opening of the Pittsburgh Home and Garden Show, and relocation of other events.

To date, the city-county Sports & Exhibition Authority has spent nearly $4 million to repair the damage and to put steel seats under other support beams along the expansion joint to prevent future problems.

Of that amount, $1.95 million has been reimbursed by the authority insurer, Travelers, but more than $2 million more is still outstanding.

Given the Wiss Janney findings, the authority will seek reimbursement for the rest from convention center designer Rafael Vinoly Architects and its engineers and the ADF Group, the lead structural steel contractor, and its subcontractors, Solicitor Mark Hornak said.

Debra Pickrel, director of communications and public relations for Rafael Vinoly Architects, said the firm had no comment on the findings. Representatives for ADF could not be reached for comment.

In the weeks after the incident, Wiss Janney, called in to investigate the collapse, had determined that the expansion joint connection locked up and the welds separated, causing the support beam and floor to give way.

What was missing until yesterday was why it happened.

Wiss Janney concluded a combination of factors was at work, starting with the expansion joint connection. The L-shaped connection angles were fitted with slotted bolt holes and welded to the expansion joint girder.

Bolts go through the slots and the support beam. The slots are supposed to allow the beam to move as it expands or contracts with changes in temperature.

But Wiss Janney said the use of slotted holes in a bearing type expansion joint "was improper due to [the] likelihood of the large frictional forces that would develop." Mr. Osborn said such connections should be used for light loads.

"In hindsight it's a bad detail. It should not have been used," he said.

In addition, Wiss Janney found that the wrong type of steel connection angles were used in the construction. The ones used were made of a higher strength steel than the design called for. The higher strength steel had less flexibility and put more strain on the welds that held the connection and beam to the expansion joint girder.

Among other findings:

• The slots in the connection angle had a "bump" in them, adding to the friction created as the bolts tried to slide along the slots in response to expansion and contraction.

• A drift pin, used to enlarge holes or align holes prior to bolting, was "tightly wedged" into one of the bolt holes instead of a bolt being there, further inhibiting beam movement.

• Washer plates were added to the connection angle even though shop drawings did not call for them. The plates caused the threaded portion of the bolts rather than the smooth shank to ride in the slots. That produced gouging and created more friction.

Since the collapse, the authority has installed steel seats under all 50 beam connections along the expansion joint. The seats provide support and allow the beams to contract and expand.