The 99-ha Keele Valley Landfill is located in a former sand and gravel pit at Maple, Ontario. The base and sides of the pit are lined with a minimum of 1.2 m of excavated clayey silt till recompacted to achieve a design hydraulic conductivity of 1 × 10−8 cm/s or less. Extensive construction controls and monitoring programs have been implemented to determine the hydraulic conductivity and advective performance of the liner. A total of 267 postcompaction laboratory hydraulic conductivity (k) tests indicated that the first two stages of the liner had a geometric mean k of 7.7 × 10−9 cm/s. Calculations of in situ hydraulic conductivity based on lysimeter effluent collection rates show decreases in k to field values close to the laboratory values. In situ electrical conductivity sensors and lysimeter effluent chemistry measurements have monitored the advance of leachate-derived chemicals into the liner. Concurrent field verification by liner exhumation and chemical analysis has confirmed the importance of diffusion as the dominant migration mechanism through this low-k liner. Similar concentration trends for major ions have been observed in the field lysimeter effluents, effluents from laboratory liner–leachate compatibility tests, and pore water extracted from core samples of sections of exhumed liner exposed to leachate. The multicomponent field and laboratory testing and monitoring programs have shown good cross-agreement, and the actual performance of the liner has been close to preconstruction predictions. Key words : landfill, clayey liner, field hydraulic conductivity, field diffusion, municipal solid waste leachate, field lysimeter test, laboratory hydraulic conductivity, liner–leachate compatibility.

Tests were conducted over the complete range σ′2=σ′3 to σ′2=σ′1 in a comparatively simple form of generalized three-dimensional shear testing apparatus (σ′1≥σ′2≥σ′3) with two pairs of rigid platens. These tests were compared with the results from triaxial compression and triaxial extension tests. Good agreement was found with the former but important strength differences were found between ISC tests near b=1·0 (μ=+1·0) and triaxial extension tests, particularly as the initial porosity of the samples increased. Special tests conducted with varying stress paths to failure demonstrated that this was not responsible for the difference in strength. In addition, errors of sufficient magnitude to account for this difference could not be invoked from examination of possible errors from platen friction, sample sheath strength, average stress level, non-uniform deformation, anisotropy or the assumption that the applied lateral cell pressure was increased by platen friction restraint. It is concluded that the method of loading was in some way responsible for the differences between ISC tests near b=1·0 and triaxial extension tests. Des essais ont été réalisés sur l'intervalle complet de σ′2=σ′3 jusqu'à σ′2=σ′1 l'aide d'un appareil relativement simple de cisaillement généralisé à trois dimensions (σ′1≥σ′2≥σ′3) comportant deux paires de platines rigides. Ces essais furent comparés aux résultats obtenus à partir d'essaisde compression triaxiale et d'essais d'extension triaxiale. Un bon accord fut trouvé avec les premiers, mais des différences de résistance importantes furent trouvées entre les essais ISC près de b=1·0 (p=1·0) et les essais d'extension triaxiale, particulièrement lorsque la porosité initiale des échantillons augmentait. Des essais spéciaux, conduites à la rupture avec des chemins decontrainte divers, ont démontré que ceci n'était pas lacause des différences de résistance. De plus, des erreurs de grandeur suffisante pour expliquer cette différence ne pouvaient pas être invoquées à partir de l'examen d'erreurs possibles de frottement de platine, de résistance d'étui d'échantillon, de niveau moyen de contrainte, de déformation non uniforme, d'anisotropie, ou en supposant que la pression appliquée des cellules latérales soit augmentée par la contrainte du frottement de platine. On en conclut que la méthode de chargement a été en quelque façon responsable pour les différences entre les essais ISC près de b=1·0 et les essais d'extension triaxiale.

The results of laboratory and in situ testing and test blasting, the observations made on a test embankment, and a description of actual construction practice associated with engineering studies for the management of uranium mill tailings at Elliot Lake, Ontario are presented. Relative density values inferred from standard penetration tests and cone penetrometer tests are shown to be inconsistent with relative density values determined from maximum and minimum void ratios. Some of the data contradicts existing correlations.The compressibility of in situ saturated tailings is presented in graphical form in terms of void ratio, vertical effective stress, and mean grain size. Hydraulic conductivity is shown to range over many orders of magnitude, depending on the void ratio. The observations on an instrumented test embankment are used to explain the appropriate selection of geotechnical parameters that gave good agreement between back-calculated and observed settlements. One-dimensional consolidation theory was found to be valid for the embankment case. It is necessary to account for changes in soil properties that occur during the consolidation process in order to obtain a good fit between back-calculated and observed settlements. The successful use of tailings sand for embankment construction is described. On the basis of normalized standard penetration resistance values, it is concluded that localized zones of saturated tailings may be prone to liquefaction under predicted earthquake loadings. Key words: uranium tailings, geotechnical parameters, relative density, test embankment, liquefaction.

No abstract