Large quantities of oil-contaminated sands resulted from exploded oil wells, burning oil fires, the destruction of oil storage tanks, and the formation of oil lakes in Kuwait at the end of the Gulf War. An extensive laboratory testing program was carried out to determine the geotechnical characteristics of this material. Testing included basic properties, compaction and permeability tests, and triaxial and consolidation tests on clean and contaminated sand at the same relative density. Contaminated specimens were prepared by mixing the sand with oil in the amount of 6% by weight or less to match field conditions. The influence of the type of oil, and relative density was also investigated by direct shear tests. The results indicated a small reduction in strength and permeability and an increase in compressibility due to contamination. The preferred method of disposal of this material is to use it as a stabilizing material for other projects such as road construction.

This paper presents the results of full-scale repeated-loading tests performed on three prestressed and six passive anchors grouted to various lengths in a rock mass. Different levels of prestress load, ranging from 22% to 73% of the ultimate static load, were applied to the prestressed anchors. Different levels and amplitudes of repeated loads, ranging from 10% to 170% of the initial prestress load and from 10% to 73% of the ultimate static load, were applied to prestressed and passive anchors, respectively. The results have shown that the anchors subjected to repeated loadings have the same service life as those only subjected to static loading, provided that the applied repeated loads are less than the initial prestress loads for prestressed anchors, or less than 30% of the ultimate static loads for passive anchors. Based on this research, it is suggested that grouted anchors subjected to repeated loadings should be prestressed to loads greater than any likely load fluctuation, and the design capacity of the anchors should be 70% of the static load holding capacity.

The bearing capacity of strip footings over a two-layer foundation soil is considered. The kinematic approach of limit analysis is used to calculate the average limit pressure under footings. The method is applicable to any combination of parameters of the two layers, but the results are presented only for a specific case when a footing is placed on a layer of granular soil resting on clay. The depth of the collapse mechanism is found to be very much dependent on the strength of the clay. Very weak clay can “attract” the mechanism even at great depths. The results are presented as limit pressures rather than traditional bearing-capacity coefficients. The latter are strongly dependent not only on the internal friction angle of the sand, but also on the thickness of the sand layer, cohesion of the clay, and surcharge pressure. Results are presented in the form of dimensionless charts for different internal friction angles of sand. It was found that linear interpolation within 5° increments is acceptable in the range of φ from 30° to 45°.

An attempt has been made to implement artificial neural networks (ANNs) for modeling the stress-strain relationship of sands with varying grain size distribution and stress history. A series of undrained triaxial compression tests for eight different sands was performed under controlled conditions to develop the database and was used for neural network training and testing. The investigation confirmed that a sequential ANN with feedback is more effective than a conventional ANN without feedback, to simulate the soil stress-strain relationship. The study shows that there is potential to develop a general ANN model that accounts for particle size distribution and stress history effects. The work presented in this paper also demonstrates the ability of neural networks to simulate unload-reload loops of the soil stress-strain characteristics. It is concluded from this study that artificial-neural-network-based soil models can be developed by proper training and learning algorithms based on a comprehensive data set, and that useful inferences can be made from such models.

Centrifuge tests were conducted on single and 3 × 3 pile groups at three-diameter (3D) and five-diameter (5D) spacings. In all of the tests, the piles were driven and laterally loaded in flight without stopping the centrifuge. The piles simulated 430 mm diameter by 13 m long pipe piles founded in medium loose ( Dr = 33%) and medium dense ( Dr = 55%) sands. Results of the tests showed that the ratio of lateral resistance of a group to a single pile, i.e. efficiency, was independent of soil density. The group efficiency at 3D spacing was 0.74, whereas at 5D spacing the group efficiency was 0.94. Drained triaxial compression tests were performed on the sand at multiple densities to obtain the necessary p - y strength data. Reese et al.'s p - y curves of 1974 resulted in excellent predictions of the prototype single-pile response for both relative densities. p - y multipliers were back-computed using the single-pile response and the measured total and individual row contributions within the group. It was found that the multipliers did not vary significantly at 5D spacing but did vary somewhat at 3D spacing. The latter was due to the variation in row distributions measured at the different soil densities. All tests were on free-headed piles.

Fractal analysis was used to evaluate the roughness of four commercially available geomembranes used in waste-containment systems. Fractal analysis uses the concept of the fractal dimension D as a way to calculate the roughness of simple and complex profiles. In the present study, it was determined that the fractal dimension D of the geomembrane profiles increased as the roughness of their profiles increased. For example, the smoothest of the four geomembrane profiles had a fractal dimension D equal to 1.001; the roughest of the geomembrane profiles had a fractal dimension D equal to 1.1345. The difference in the fractal dimension D for the four geomembrane profiles was found to be relatively small. Thus, when the fractal dimension D is used to evaluate the roughness of geomembranes, at least four-digit precision after the decimal point is advisable. The sensitivity of the fractal dimension to inputs such as the direction of measurement and cut as well as the size of the segment length used to cover the geomembrane profiles were also analyzed.

One aspect of the original Cam-clay model is that the derived undrained stress-strain curve can be expressed in closed form and used to determine a finite value of stiffness at ε= 0 that is equivalent to an undrained plastic modulus (or E _u ) for normally consolidated clays. With respect to soil behavior, this predicted modulus in no way represents the initial modulus E_max that is fundamental and obtained from resonant column tests at small-strain levels (ε< 10 ⁻⁵ ), but interestingly corresponds to the apparent initial undrained stiffness obtained from a hyperbolic fit of isotropically consolidated undrained triaxial stress-strain data over the range of intermediate to large strains (0.001 <ε< 0.10). A review of triaxial and consolidation data from 17 clays supports this observation.