You are on page 1of 15

Comparison of predictions

of beach slopes using


alternative models
J. Covacevich Hatch Associates, Chile
J. Wates Fraser Alexander, South Africa
S. Dal Pozzo Codelco Andina, Chile
Agenda
1. Introduction
2. Methods Overview
3. Tailings characterization
4. Results
5. Conclusions
Introduction
•Chilean context:
• Largest copper mine operations
• Mainly in the high Andes of Northern
Chile
• Tailings disposal constitutes big
proportion of Capital costs
• Increasing costs of energy
• P&TT experience derived from relatively
small operations

•Designs are based on Pilot-scale tests


•Pilot Plant involve thickening, transport
and deposition tests:
• A variety of methods exist
• Comparison of most often cited
methodologies
Methods Overview
McPhail Method:
•Incorporates entropy and stream power.
•Based on:
• Natural phenomena can be described using entropy minimization.
• The stream power is the energy per unit of time consumed by a particle moving through the beach.

•Information required:
• Beach elevation.
• Beach length.
• Slurry density.
• Flow rate.
• Estimate stream power.
• Estimate initial beach slope.

•Using equilibrium equations calculate velocity (stream power and flow rate).
•Plot shear stress vs shear rate.
•Use tailings and operation characteristics to estimate initial stream power (normal hydraulic
equations).
•Calculate flow rate at the outlet of the plunge pool and initial slope.
•Beach length is iterated until shear stress and shear rate matches actual rheogram.
Methods Overview (continued)
•8 were curried out
•Solids concentration between 65% and
68% w/w
•Rheological parameters were
monitored during trails
•Surveys were curried out after each
test
Methods Overview (continued)
Fitton Method:
•Beach slope is defined by an equilibrium between deposition and erosion in
a channel.
•For given flow rate, mean velocity V depends on an assumed channel depth
with assumed channel shape.
•Critical velocity is a function of rheology, tailings density and the hydraulic
radius for the given shape of the channel.
•Channel depth is iterated until V=Vc.
•Friction slope is calculated assuming uniform flow and defines the beach
slope.
Methods Overview (continued)
•Rheological characterization
• Yield Stress
• Viscosity
• Herschel-Bulkley fitting parameters

•Particle size distribution


•Solids specific gravity
Tailings characterization
•Bingham Plastic
•Solids Specific gravity: 2.63 ton/m3
•Particle Size Distribution: D50 = 86 um
D90 = 326 um
-38 um = 31.5%
Results
Cw = 65%

Cw = 68%
Results (continued)

•Non-segregating slurries should not produce


concavity
•We see it almost in every thickened tailings deposit
•Only segregation of heaviest particles can cause
concavity at fixed flow rates and solids concentrations
Results (continued)
Concavity on tailings beach:
•Variations of UF properties
•Variations on tailings flow rate

Assumptions:
•Data obtained from five mines
•Coefficient of variation of Solids
Concentration: 5%
•Coefficient of variation of flow
throughput: 10%
Results (continued)
Cw =62% 65% 68%

Cw =65% 68% 71%


Conclusions
•About the Methods:
• Are based on fundamentally different models
• Both give similar results for average slope and for concavity
• Both methods requires comparatively same amount of information
• Results should be used as a guide

•About the Phenomenon:


• Concavity is explained best by variations in the slurry yield stress
• At the end of the run out the tailings stalls
• Layer thickness plays a big part
Conclusions (continued)
•About the Trails:
• Actual slopes may differ from those predicted by applying theory to pilot scale trails
• Distribution systems must be design in order to allow good deposition conditions
• Variability of rheological properties from sample
• Practicity for Greenfield projects
• Representativeness of the samples to be trailed
Thank you!

You might also like