More mining companies are investigating filtered tailings as part of their long-term sustainability plans and to reduce risks associated with tailings storage. FLSmidth (FLS) recommends that miners conduct tailings technology trade-off studies which includes, analysing different tailings solutions options, including different flow sheets and equipment to determine the optimal dewatering process. This analysis includes determining the most economic approach for each tailings dewatering option. The primary factors that affect filtration are the settling rate of the solids for thickening, feed density to the filters, target filter cake moisture, particle size distribution (PSD), mineralogy and particle shape. The impact of feed density on filtration will be the focus of this paper.
Thickeners are frequently used to increase feed solids concentration to filters, reducing the quantity of water that filters need to remove from the tailings slurry. Less water in the filter feed, increases filtration rates and can reduce the quantity of filters required decreasing CAPEX and OPEX. Surface disposal of tailings requires a filter to remove enough water to provide the geotechnical stability to stack. There are many types of filters used on mine tailings such as, rotary vacuum disc filters, horizontal vacuum belt filters and filter presses. Gold mines tend to have a finer grind than other metal tailings and therefore filter presses are usually specified for surface stacking.
This paper presents a case study for a greenfield gold site looking at filtering tailings for increased water recovery. The tailings dewater options considered in this project was: 1) direct filtration of the gold tailings using a filter press; and 2) thickening followed by filtration with a filter press(es). A comparison of the filtration rates and filter costs with and without thickening prior to filtration will be presented in this paper.

Coal processing plants generate a large volume of tailings containing fine and colloidal particles. The most conventional method for coal waste disposal is using tailing storage facilities (TSFs) (Khazaie et al, 2022). Tailings dam failures have always been a serious environmental threat to the contamination of surface and groundwater due to seepage, blockages, and insufficient capacity of spillway systems which leads to overtopping (Fourie, 2009; Spain and Tibbett, 2012). Moreover, mine tailings disposal imposes additional cost to mining companies. According to NSW Environment Protection Authority (EPA), the waste levy rate for Coal Washery Reject (CWR) in 2019–2020 is $15 per tonne (NSW Environmental Protection Authority (NSW EPA), 2020). Sustainable alternative proposals for coal tailings disposal in tailings dams include mechanical dewatering methods such as coagulation, flocculation, and sedimentation, followed by filtration. Efficient selection of flocculants is an important factor in this process.
In this study, mono flocculation experiments were conducted using various polyacrylamide (PAM) based flocculants with different molecular weight and charge type to evaluate the effect of molecular weight on settling rate and turbidity removal of coal tailings. Microstructural analysis of coal tailings including the X-Ray Diffraction (XRD) and Scanning Electron Microscopy and Energy Dispersive X-ray Spectroscopy (SEM-EDS) results, demonstrated that coal tailings were mostly composed of quartz, magnesium calcite, dawsonite, muscovite, and different clay minerals, including kaolinite, illite, and montmorillonite.

Standard practice to evaluate global stability of tailings storage facilities (TSFs) entails the use of limit equilibrium analyses, which considers either peak or residual undrained shear strength of the saturated area and drained shear strength for the unsaturated soils and tailings. This practice might still be non-conservative, as nearly saturated materials can develop excess pore pressures and behave in an undrained mode. Moreover, limit equilibrium analyses do not account for the work input required to drive the stress state past the peak and into the softening process. Numerical deformation models, on the other hand, can readily model the transition from saturated to unsaturated conditions and from peak to residual strength, and therefore are able to provide further insight of the static liquefaction vulnerability of a given dam.
This paper describes the application of finite element modelling in Plaxis 2D to the analysis of the vulnerability of a TSF to static liquefaction. The procedure employs off-the-shelf features like the Hardening Soil Model with Small Strain Stiffness (HSS) constitutive model and Van Genuchten Soil Water Characteristic Curve (SWCC) to simulate the shear-induced generation of pore pressure in the saturated and nearly saturated materials while providing a smooth and physics-friendly transition from the fully undrained to the fully drained behaviours of the various portions of the dam. It is demonstrated that, by providing a step-like SWCC, the procedure produces nearly identical results to those obtained by ignoring suction altogether and zoning the dam by hand. This, by itself, is useful, because the position of the phreatic surface can be varied within the model without the need of remeshing. On the other hand, if a realistic SWCC is employed, the vulnerability of the facility can be better understood and reported.

With several catastrophic tailings dam failures happening in the recent decade, claiming human lives and causing environmental damage, it has become crucial that the tailings dam break analysis (TDBA) be properly conducted to represent the credible failure modes and scenarios, and to assist with the assessment of the critical failure impact. This also urged the mining industry and regulatory authorities such as ANCOLD, GISTM and Dam Safety NSW etc to either develop new guidelines or revise the existing guidelines and regulations to more accurately define the realistic failure modes and Failure Impact Assessment (FIA) for the tailings storage facilities (TSF).
Most of the current guidelines and regulations categorise the dam failure and consequence category scenarios based on the initial hydrologic conditions, which include sunny day failure and flood failure. Nonetheless, the flood failure scenario among different guidelines is generally only defined as the storm-related failure events with concurrent downstream flood condition, without much details about the failure mode, and is usually only assessed for the incremental dam break impact over the natural flood in an embankment overtopping scenario. This may not necessarily represent the worst-case failure impact and may consequently underestimate the consequence category for the TSF.
Based on the authors’ experience in tailings dam FIA, there should generally be two sub-categories of the flood failure for the TSF, namely the incremental flood failure (IFF) and the post-flood failure (PFF), with the former assessing the incremental failure impact over the natural flood event, while the latter one assesses the direct dam break impact once the downstream flood has subsided. These two scenarios should both be assessed following the potential credible failure modes, and the more severe impact from these two scenarios should then be considered for the flood failure consequence category assessment.
Key differences between the IFF and PFF for FIA are discussed in the paper, with several case studies presented to demonstrate the validity and necessity to consider both scenarios in assessing the credible, worst-case failure impact.

Tailings storage facilities (TSF) are an essential part of mining. Understanding their risks is vitally important for regulators as well as for nearby communities, lenders and shareholders. The tailings dam breach analyses (DBA) are an important input to understanding these risks. They provide information for the consequence classification assessments, risk assessments, emergency response plans, etc. These assessments are the driving force behind safely managing existing TSF and safely designing new TSF.
Identifying the hazards associated with a TSF is a vitally important step in this process. For this, there are currently several industry guidelines around the world such as the Global Industry Standard on Tailings Management (GISTM, 2020), Australian National Committee on Large Dams (ANCOLD, 2019), and the Canadian Dam Association (CDA, 2021), along with industry-specific regulatory requirements. In recent years, detailed company requirements have also been introduced. Together these create high standards for the DBA, and now, in most cases, they supersede the previous minimum requirements. Today, as industry pushes for increased consideration of credibility in the DBA’s input assumptions, a high degree of technical detail must be assessed. This poses challenges for consultants completing the work. To undertake a reliable and credible DBA, consultants now must consider a range of requirements: the TSFs technical aspects, inputs from stakeholders and the Engineer of Record, while balancing all this with data gaps and the inherent uncertainties of our current methodologies, eg model uncertainties.
This paper aims to outline the main challenges encountered in studies carried out in Brazil, Canada and Australia. It also presents a framework, which includes suggested steps to overcoming these challenges. The key phases of this approach are a thorough desktop review and gap analysis; a workshop on credible failure modes with stakeholders; outflow volume definition considering embankment and tailings characteristic inputs; a rheology assessment on deposited tailings; definition of the downstream hydrological setting; breach and runout modelling; and sensitivity analysis on inherent uncertainties associated with the DBA.

As more scrutiny is given to upstream tailings facility’s liquefaction potential and resulting strengths around the world, saturation of the tailings becomes an important question for tailings practitioners, especially in conditions of near, but not fully saturated tailings. The difference in seismic behaviour between near, but not fully saturated and fully saturated tailings can have significant implications in the stability outcomes and design configurations for tailings storage facilities. Near fully saturated tailings have beneficial effects that include suction that can contribute to increased resistance against liquefaction. This may not be fully understood under the influence of cyclic loading and a complex tailings column comprising interbedded saturated and near saturated layers. Current in situ techniques for measuring Saturation Ratios (Sr) are still in development, leaving designers and owners with uncertainty around stability outcomes and conservative costly remediation in some instances.
This paper presents a unique and extensive site investigation performed at an Australian gold mine upstream tailings facility, in predominantly near (but not fully) saturated tailings. The paper presents the method and results obtained by direct measurement of Sr through the collection of 82 piston tube samples to full depth of the tailings column that were used to determine void ratio and density. The tube samples were supplemented with adjacent SCPTu dynamic Pore Water Pressures (PwP), and profiles of Compression wave velocity (Vp) and shear wave velocity (Vs) that will be presented for comparison and correlations. This data has been used to support and validate the instrumented columns along the tailings facility and improve the interpretation of the tailings saturation regime and resulting assessment of the susceptibility against liquefaction.

The use of critical state soil mechanics (CSSM) in tailings engineering has become well established. While CSSM in engineering practice includes many different procedures and analysis frameworks, the selection of an in situ state at which contractive or dilative behaviour is expected is one of the most important choices to be made. Over the past few decades, an evolving understanding of full scale flow liquefaction behaviour and state parameter Ψ interpretation methods has led to a value of Ψ = -0.05 being commonly applied as the transition value. However, while justifiable from case history data, the rationale for, and implications of, this contractive/dilative boundary are often not appreciated. For example, viewing CSSM in a ‘steady state’ interpretation framework, where shearing soil tends towards the critical state line (CSL) at large strains, is quite inconsistent in outlook from a contractive/dilative boundary of Ψ = -0.05. To illustrate the rationale for this boundary, a review is first carried out of relevant case history data, then of element test behaviour when sheared from initial states near to the CSL, with a particular emphasis on localisation and fabric anisotropy. Finally, the wisdom and implications of extending the Ψ = -0.05 boundary far from where it was originally justified are outlined.