Unsaturated flow in tailings

The work presented in this document is two-fold. The major part of the work described herein, involved designing and building an apparatus to determine the unsaturated coefficient of permeability of mine tailings in a laboratory environment. The Modified Permeameter was built based on a permeameter used by Meerdink et al (1996) which was used with compacted clay samples. The modifications include using a large size sample; the ability to accommodate mobile sensors to alleviate the problems associated with the formation of water pockets and to use the same test specimen and determine the SWCC and unsaturated coefficient of permeability of the tailings over a large range of saturation. Mobility of the sensors was achieved with a system of counterweights, which also retained the water-tightness of the apparatus. Although Time Domain Reflectometry (TDR) sensors did not yield any results during this research program, the Modified Permeameter includes counterweight system to provide mobility for three TDR probes to measure volumetric water content. During the experimental work, water content was measured with sampling tools developed and built for use in the Modified Permeameter. Negative pore water pressures were measured with tensiometers. Two gradations of tailings were tested with the Modified Permeameter. The two gradations were the overflow (fine) and the underflow (coarse) samples. Prior to laboratory experimentation with the tailings, numerical predictions of soil-water characteristic curves (SWCC) and conductivity curves were computed. These predictions were obtained using the SoilVision software package. The predicted curves were compared with the laboratory values. The estimated SWCC for both tailings gradations and the laboratory results show good agreement. In regards to the conductivity curves, insufficient laboratory data was gathered to compute the unsaturated coefficient of permeability of the coarse underflow sample. The laboratory results for the overflow sample show important scatter around the predicted conductivity curve. Due to the differences observed between the laboratory results and the predicted conductivity curve, the numerical modeling portion of this research, which constitutes the second portion of this research, was conducted using typical permeability values. The modeling portion consisted of producing a generic model of an upstream tailings dam on which individual factors were analysed. The factors considered were as follows: (1) Effect of beach length; (2) Effect of pervious foundation; (3) Effect of blanket drain; (4) Effect of segregation along the beach, (5) Effect of anisotropy; (6) Combined effect of segregation along the beach and blanket drain. The numerical analysis showed that maintaining a good beach length throughout the construction period of the upstream dam helps to prevent saturation of the fine tailings underlying the shell. The addition of a blanket drain to a good beach length provided the desired effect, which was to lower the phreatic surface away from the fines closest to the dam's shell