B. Brouard

Mechanical Behavior of Salt Caverns: Closed-Form Solutions vs Numerical Computations

Creep closure and structural stability of a cylindrical elongated cavern leached out from a salt formation are discussed. The Norton-Hoff creep law, or “power law”, is used to capture the main features of salt rheological behavior. Two failure criteria are considered: (1) shear stresses must not be larger than a certain fraction of the mean stress (dilation criterion); and (2) the effective stress at the cavern wall (actual stress plus cavern fluid pressure) must not be tensile. The case of a brine-filled cavern whose pressure is kept constant is discussed first. It is proved that creep closure reaches a steady state such that stresses in the rock mass remain constant. However, decades are needed to reach such a state. During the transient phase that results from the slow redistribution of stresses in the rock mass, deviatoric stresses decrease at the vicinity of the cavern wall, and onset of dilation is less and less likely. At this point, the case of a rapid brine pressure increase, typical of a tightness test, is considered. It is proved that during such a swift pressure increase, cavern behavior is almost perfectly elastic; there is no risk of dilation onset. However, even when cavern pressure remains significantly smaller than geostatic, the effective stress at cavern wall can become tensile. These results, obtained through numerical computations, are confirmed by closed-form solutions obtained in the case of an idealized perfectly cylindrical cavern; these solutions provide a better insight into the main structural features of the behavior of the cavern.

A Very Slow Creep Test on an Avery Island Salt Sample

The applied deviatoric stress during most creep tests performed on salt samples is in the 3.5–20xa0MPa range. However, the stresses actually experienced in the vicinity of a salt cavern are much smaller. Any extrapolation is difficult to vindicate, as the dominant micro-mechanisms are strongly suspected to be very different in the low-stress and medium-stress domains. To answer this concern, a very slow creep test was performed on an Avery Island salt sample. To minimize the influence of even the smallest of temperature deviations during the test, the testing apparatus was placed in a remote gallery of the Varangéville salt mine, taking advantage of the very stable temperature conditions offered in an underground environment. The test was performed in multiple stages and lasted 42xa0months. The successive loads of 0.1, 0.2, 0.3, and 0.5xa0MPa were applied. Measured steady-state strain rates were of the order of 10−12 s−1, which are significantly faster than that extrapolated from creep tests performed at loads ranging between 3.5 and 20xa0MPa.