The construction of the Péribonka dam involved deep compaction of its foundation using vibroflotation and dynamic compaction. Surface wave testing was used, in addition to classical tests (cone penetration tests (CPTs) and standard penetration tests (SPTs)) for the assessment of vibrocompaction. More than 900 shear wave velocity (Vs) and 1000 CPT profiles were obtained. This set of tests performed prior to and following vibrocompaction constitutes an important data bank, used in this study to establish a relationship between normalized shear wave velocity, Vs1, normalized tip resistance, qc1, and mean grain size, D50. Using the Péribonka project data obtained on fairly coarse sands in conjunction with the Canadian Liquefaction Experiment (CANLEX) project data obtained on fine sands has confirmed the significant effect of particle-size distribution on the relationship between Vs and qc. The paper proposes a correlation between Vs1, qc1, and D50 for uncemented and Holocene-age granular soils in continuity with the relation developed by Wride et al. from the CANLEX project.

A precise evaluation of shear wave velocity, Vs, is a crucial issue in the design of foundations subjected to dynamic loading, liquefaction evaluation, and soil improvement control. Laboratory techniques such as resonant column (RC) and bender element (BE) have been developed over the years to measure Vs. At low strain (γ < 10−3), techniques based on piezoelectric elements (e.g., BE) can be considered superior to RC, as they can be used in conventional geotechnical devices (e.g., triaxial, oedometer, direct simple shear, etc.). However, it is a difficult task to verify that the obtained Vs values are correct and accurate, as there are several difficulties associated with these methods, including the mixed radiation of both primary and shear waves, near-field effects, boundary effects, and uncertain detection of first arrivals. This paper presents the use of a new technique to measure Vs in granular material, called the piezoelectric ring-actuator technique (P-RAT), developed at the Université de Sherbrooke. The paper also provides a detailed description of a unique interpretation method of the signals produced from this technique to minimize the difficulties associated with other techniques. The P-RAT has been incorporated into the well-known oedometer cell to measure the Vs of Péribonka sand through a series of oedometric tests, and the obtained results have been detailed, analyzed, and discussed in light of the basic state of knowledge of Vs and factors affecting it. Particular emphasis is also placed on the validation of the accuracy of the P-RAT by means of reliable experimental measurements available in literature.

This study was conducted to assess the representativeness of laboratory sampling protocols for purposes of trace metal analysis in soil. Five laboratory protocols were compared, including conventional grab sampling, to assess the influence of sectorial splitting, sieving, and grinding on measured trace metal concentrations and their variability. It was concluded that grinding was the most important factor in controlling the variability of trace metal concentrations. Grinding increased the reproducibility of sample mass reduction by rotary sectorial splitting by up to two orders of magnitude. Combined with rotary sectorial splitting, grinding increased the reproducibility of trace metal concentrations by almost three orders of magnitude compared to grab sampling. Moreover, results showed that if grinding is used as part of a mass reduction protocol by sectorial splitting, the effect of sieving on reproducibility became insignificant. Gy's sampling theory and practice was also used to analyze the aforementioned sampling protocols. While the theoretical relative variances calculated for each sampling protocol qualitatively agreed with the experimental variances, their quantitative agreement was very poor. It was assumed that the parameters used in the calculation of theoretical sampling variances may not correctly estimate the constitutional heterogeneity of soils or soil-like materials. Finally, the results have highlighted the pitfalls of grab sampling, namely, the fact that it does not exert control over incorrect sampling errors and that it is strongly affected by distribution heterogeneity.

No abstract

No abstract