Electric Power Plant
Pacific Island Paradise
The fasteners holding the side pressure plates on a new type of
turbine generator were breaking. When the fasteners broke, for safety
purposes, pressure sensors would automatically shut down the plant, and
all electric power to the island was lost.
There were no physical injuries at the power plant. However, the
for the new plant was that when Hurricane Iniki hit the island, there
were sustained winds averaging 150 miles per hour (241 km/hr) and gusts
in excess of 200 miles per hour (322 km/hr). This phenomenally destructive
force ravaged the island with damage. It caused a number of deaths, and
it stripped the land of every blade of grass, turning a lush tropical
landscape into a desolate smudge of brown. The storm also took away their
power plant, and each time the new one went out because of fasteners,
the mending fears of the residents were dealt with a relapse of bad memories.
There was no damage to the plant, only to the fasteners.
The pressure plates were rectangular steel about 2' (200cm)
by 3' (300cm)
with fasteners spaced about every 3" (75cm) apart around the periphery,
with a thin gasket. All fasteners broke in the thread runout area adjacent
to the extrusion angle leading to the full body diameter. A flat washer
was used under the head of the screws, and they were of plain finish.
The plates and the surrounding metal structure appeared to be painted
with a type of hi-temp black paint.
This plant was the first of its type in service, and
it ran without
difficulties for a few days after being built. The manufacturer was contacted
when the blackouts began. However, there was no interest
in coming a great
distance over small bolts breaking. They said to follow the maintenance
guide, apply anti-seize compound to the threads and torque the bolts to
the manual. Further, because of liability, the manufacturer prohibited
changing the bolt from the specified product. The fasteners in question
were 1/4"-20 x 1 1/4" Grade 8 hex cap screws
through-hardened SAE flat washers. The failure mode was obvious fatigue.
The torque with lube was reasonable, and there were ample threads available
within the tapped holes. However, the screws were nearly bottomed out
into the extrusion angle because of the manufacturer's length selection.
While the installation was acceptable, the screws
too long, and this focused service loads to the thread runout area. The
loads needed to be spread out.
A good practice, where viable, is to distribute loads over as many
threads within the clamped joint as possible. The plant didn't dare
change the screw, but the application was made successful by placing thick
spacers under the heads. This effectively put a "spring"
the assembly, and was enough of a change for the island to keep its
As appeared in Fastener Technology