9 / 352
5
Today, the reliability of engineered products and structures is at an all-time high, but
this reliability often comes with a high cost. In fact, in the nuclear industry, compliance
with regulations intended to maximize safety may be so costly as to warrant the taking of
a reactor out of service. It is also important for manufacturers to be aware of the state of
the art as well as the latest standards. The number of manufacturers of small planes has
dwindled because of product liability losses incurred when it was shown that their manu-
facturing procedures did not meet the current state-of-the-art safety standards. To guard
against product failures, a number of firms now are organized in such a way that failure
analysis is a line function rather than a staff function, and a member of the failure analysis
group has to sign off on all new designs before they enter the manufacturing stage.
Two typical examples are selected to illustrate the significance of failure case studies
and the benefits which can be gained for structure design and service improvement.
2. THE ALEXANDER KIELLAND PLAT FORM DISASTER
On the evening of 27
th
March, 1980, the Alexander Kielland, a drilling rig converted
into an accommodation platform and located in the North Sea, started to capsize and
within 20 minutes had overturned killing 123 of 212 people on board /3/. The reason for
the failure was later traced to a small 6 mm fillet weld which joined a non load-bearing
flange plate to one of the main bracings. The purpose of the flange plate was to hold a
sonar device used in connection with drilling operations; ironically, the platform was
never actually employed as a drilling rig. This case study is concerned with the possible
factors that contributed to the failure of the weld.
The Alexander L. Kielland was a mobile platform of the pentagon type and was
designed and built in France at the Dunkirk Shipping Yards. The rig was ordered by the
Norwegians in 1973 and delivered in 1976. It was originally built as a drilling rig, but
during its entire operation it was utilized instead as an accommodation platform. Initially,
its capacity was 80 beds, and in April, 1978, this had been increased to 348. Altogether,
eleven platforms of this type have been built, nine are still in service in the North Sea.
The characteristic form of a pentagon platform design is shown in Fig. 2. The main
concern herein is the D column and the bracing D-6. A detail of this part of the rig is
given in Fig. 3.a, which shows the location of the sonar flange plate, welded to the main
bracing (D-6), Fig. 3.b. The D-6 bracing is 24 m long, circular, hollow beam of diameter
2.6 m and thickness 26 mm. It is left open to the sea and allowed to contain sea water in
order to increase the rig's stability. For that, the bracing contains an elongated opening
(300 x 800 mm) on the bottom of the bracing next to the sonar flange plate. Both the air
hole and elongated opening were fitted with flanges to reduce the stress concentrations.
The production schedule was such that the assembly work was divided between two
teams, one team being responsible for the main welding and fitting operations and the
other taking care of the welding and fitting of ancillary equipment. In this respect, for
example, the welding of the flanges to the elongated opening and air hole was included
among the duties of the main installation team, while the welding of the non load-bearing
sonar flange plate was the responsibility of the other team. Furthermore, it was not
considered necessary in the design work to carry out any stress analysis of the sonar
flange plate fitting, although a stress analysis of, e.g., the oval hole flange plate was
carried out. This turned out to be a vital omission. The main braces were of a welded
construction and made from Nb -microalloyed fine-grained steel.