A13

An experimental program was conducted to determine whether or not the Cone Penetration Test can be used to detect layers of gas hydrates in the soil below the seafloor of the Gulf of Mexico. To simplify the experimental procedure, gas hydrate layers were modeled by ice layers.

In a first step, Cone Penetration Tests were run without confinement using a 10 cm2 penetrometer on homogeneous sand samples which were 90 cm high and 60 cm in diameter.

Then additional tests were run, both without confinement and with a 0.7 bar confining pressure, using a 1.67 cm2 penetrometer. The samples were 60 cm high and 30 cm in diameter. Homogeneous sand samples, homogeneous crushed ice samples and layers samples were successively tested. The layers samples consisted of a 1 or 3 cm thick frozen saturated sand layer located in an homogenous dry sand medium.

The penetration tests run without confinement did not enable a proper identification of the sand and crushed ice studied.

The application of a confining pressure of 0.7 bar enabled a proper classification of the laboratory sand and of the crushed ice. The crushed ice was classified as a sandy gravel to gravelly sand.

Tests run with confinement showed that a 1 cm thick frozen layer can be detected but not properly identified with a 1.67 cm2 penetrometer. But, with a confining pressure of 0.7 bar, a 3 cm thick frozen layer can be detected and classified as heavily cemented soil. Since a layer of heavily cemented soil is not likely to be encountered in oceanic subsoils, layers of massive hydrate may be successfully identified by the means of a CPT run with a 1.67 cm2 penetrometer.

In addition, the scale effect was analyzed mathematically and a computer program was developed to correct the recorded profiles.. However the conclusions on the 1.67 cm2 penetrometer cannot be extended to the 10 cm2 penetrometer because the computer program could not be checked experimentally. Recommendations for future work are included.