Filtration performance of geotextiles in cyclic flow conditions : a laboratory study

Geotextile filters are often used as a replacement for, or in combination with, traditional granular filters in many engineering works. Conventional design criteria, which are largely empirical, are generally sufficient for applications where flow is unidirectional and the soil is internally stable. However, for conditions including reversing flow regimes and potentially internally unstable soils, these criteria may not be adequate and performance tests may be necessary. The gradient ratio test is a performance test that assesses soil-geotextile compatibility under an applied hydraulic gradient. The gradient ratio device developed at UBC is a modified version of the ASTM apparatus, which allows the application of both unidirectional and reversing flow to soil-geotextile systems at varying hydraulic gradients and confining pressures. In this research work, three soils were tested in combination with seven geotextiles, using the modified gradient ratio device. Two of the geotextiles were nonwoven materials and 5 were woven, with AOS values ranging from 0.212 mm to 0.600 mm. The soils were a Fraser River sand, a copper mine-waste tailings and a Port Coquitlam silty sand. The mine-waste tailings and Port Coquitlam silty sand were recognized as potentially 'problematic' from a filtration standpoint. The soils had a relatively narrow range of D85 (from 0.330 mm to 0.215 mm), and a moderate range of coefficient of uniformity, Cu (from 1.8 to 5.8). The tests therefore provided results for AOS/D85 values ranging from 0.6 to 2.8. The intent was to gain insight to (i) the influence of geotextile type (woven versus nonwoven), (ii) the influence of flow regime (unidirectional versus cyclic), and (iii) the validity of existing design guidance for the range of soil and geotextile combinations used in testing. Based on the very limited comparison of three geotextiles of the same opening size, it appears there is little difference in behaviour of these woven and nonwoven geotextiles. All tests were relatively stable, with insignificant quantities of soil passing through the geotextiles. Results for the Port Coquitlam silty sand, which yielded the most soil passing through the geotextile, showed a small difference in the grain size distribution of the passing soils. It appeared that more of the finer material passed through the woven geotextile, than the corresponding nonwoven. The influence of flow regime was studied from tests in unidirectional flow, and cyclic flow with and without confining stress. No significant influence of the frequency was found in testing for the flow reversal at frequencies of 0.02 Hz and 0.1 Hz. The Fraser River sand is stable in all tests and therefore, the influence of flow regime does not appear to be significant. The mine waste tailings are stable in all unidirectional flow, and generate a very subtle trend towards piping instability in cyclic flow as the AOS/D85 approaches 2.0. The GRASTM and GRMOD values correspondingly are less than unity thus indicating the onset of piping. The Port Coquitlam silty sand behaved slightly differently than the other two soils, in that it is stable in both unidirectional flow and confined cyclic flow, but experienced significant piping and collapse of the soil structure with unconfined cyclic flow. The soil yielded catastrophic piping during sample preparation when the AOS/D85 was 2.8. The results were used to evaluate the design criteria of CGS (1992) and Luettich et al. (1992) in unidirectional flow, and CGS (1992), Luettich et al. (1992) and Holtz et al. (1997) in cyclic flow. The CGS (1992) and Luettich et al. (1992) guidance were found to be slightly conservative for soil-geotextile filtration compatibility in unidirectional flow. For cyclic flow, all three criteria were again found to be reasonable, but somewhat overly conservative.Applied Science, Faculty ofCivil Engineering, Department ofGraduat