USE OF VALID ANALYTICAL METHODS: Oil Tank in Norway.


Bjerrum and O ¨ verland (1957) studied the failure of an oil tank in Norway.

Soil borings were made at the site of the tank. Properties of the soil were obtained by the use of the in situ vane and by the unconfined compression test. The upper layer of soil was a silty clay to a depth of 7 m with a soft marine clay below. The undrained shear strength was reconstituted to obtain values for the construction time of the tanks. The strength of the soil was almost constant to a depth of 10 m with a value of about 3 tons/m2, but the shear strength was over 4 tons/m2 at a depth of 15 m. The diameter of the tank was 25 m.

If a failure of the entire tank was assumed, the average shear strength along the failure surface yielded a factor of safety of 1.72. On the other hand, if failure was assumed at one edge of the tank, the average shear
strength along the shallower failure surface was less and the factor of safety was computed to be 1.05.

The heave of the soil was near one edge of the tank and indicated a local failure in which the weaker soil was mobilized. The error in the original design was due to the selection of a model for general failure where the
average shear strength along an assumed failure surface was significantly larger than the average shear strength of the near-surface soil related to local failure. Thus, the careful selection of a model for a failure surface that will yield the lowest factor of safety is essential.

0 comments:

Post a Comment