A recurring problem in mining projects is the collection and interpretation of data which is applicable to backfilling operations in general. The basis for design is therefore compromised by external factors, so the design expectations must be tempered to give confidence to Mine Management that the design is robust.

From an underground geotechnical engineering perspective, the long-term aim of backfill is to promote mine stability through rockmass confinement, while optimising recovery of the ore. Empirical assessments are a good starting-point for a design, but 3D numerical modelling is key to understanding how the mining sequence influences rockmass conditions and therefore the conditions into which backfill will be placed.

Wiles (2006) proposed that the combined interaction of modelling input parameters can be expressed as a coefficient of variation and characterised according to the following formulae:

Clearly the coefficient values vary by project and the stage of study, but some have suggested that COutput  values of 40% are possible with high quality data, but in practice the variation will be closer to 60%.

How does this influence backfill placement? The calculation demonstrates that numerical modelling alone cannot always predict rockmass conditions to a high level of confidence. Many sites will therefore use numerical modelling to indicate likely conditions based on a comparison between modelling scenarios to demonstrate the most favourable option for the mine.

From a backfill perspective, geometric confidence could relate to a ‘vertical’ stope overhanging an adjacent footwall stope by 5°-10° in areas. In these cases, the design assumption that the backfill is free standing and vertical is misleading and could lead to unplanned failure of the fill when undercut and exposed.

The modelling resolution applied at the design stages of a mine tends to be low, with the aim of developing a generalised understanding of the deposit at mine-scale. Unless development is already in place or detailed drilling has been undertaken, there will be limited understanding of localised structure and geological variance. This might manifest itself as localised instability and overbreak; influencing the backfill volume requirements and potentially subjecting a backfill column to hanging wall failure during filling (in the case of paste fill, potential for backfill liquefaction and/or barricade rupture).

There are several influences to be considered regarding stress including the virgin stress regime, mining-induced stresses, brow stresses and the influence on the regional stress field caused by regional mining. If the backfill is designed to withstand certain forces but is subjected to unforeseen conditions, excess strain, lack of support or backfill failure may occur. A backfill designed to yield under certain conditions may yield excessively (or not at all), creating conditions for which the mine has not planned.

Depending on the backfill and rockmass properties, the numerical model may not actually ‘see’ the backfill due to a large disparity between parameters. In these cases, numerical modelling tells the engineer nothing about the influence of the backfill on confinement. This is a modelling limitation which may be overcome by ignoring the fill altogether and treating the fill as a void (conservative design method). At the other extreme, the modeller may artificially increase the backfill parameters to force a reaction within the model. Whether either of the two approaches are taken or another method applied, the backfill engineer needs to be aware of how the backfill has been modelled so they can mitigate against any unplanned outcomes when filling the voids.

The age old saying in numerical modelling is ‘rubbish in, rubbish out’. The definition of backfill properties is relatively straightforward compared to, stress, geological and geotechnical properties but poor inputs in other areas will influence modelling behaviour. The engineer must also be aware of the modelling calibration process. What parameters were adjusted to make the model run? Are the adjustments reasonable and do they reflect expected values?

Finally, the suitability of the modelling code should be considered in respect to backfill. There are many different methods so the resultant outputs should be considered in the context of the modelling applied.
So why bother modelling at all? A good model will indicate likely conditions and by running multiple scenarios, will help the engineer explore the envelope of possibilities and determine the most favourable scenarios.

Regarding the under-cutting of backfill, this action represents a challenging method of mining which cannot be adequately assessed using empirical methods. The mine requires confidence in the analysis before exposing personnel or machinery to development (or stopes) with exposed backfill backs. This analysis should be considered separately to the overall mine model and targeted towards determining the strength requirements in these areas.

Is there a place for empirical modelling? Absolutely. Mitchell, Terzaghi and others have developed empirical modelling and these methods have been applied in mines across the globe. The level of certainty is appropriate for many backfill applications (especially in the early stages of design) and as with numerical modelling, have factors of safety applied during interpretation to account for unforeseen variables.

P&C’s approach is to select the assessment appropriate to the level of study and appropriate to the data available at the time. The modelling process can be re-evaluated as more information becomes available and in order to increase confidence in the backfill design. Modelling complexity is only escalated where necessary and only where it can be supported by the input data.

About The Author
Andy Beveridge
BSc. (Hons), ACSM, Applied Geology, MAusIMM(CP)

Andy has worked on a variety of civil and mining projects and undertaken various roles including Paste, Geotechnical, Project and Service Engineering. He has significant site experience and specialises in the design and application of backfill for mining applications. Works have included engineering studies ranging from scoping to feasibility level, through to the construction, retrofitting and maintenance of fixed and mobile plants. Additionally, Andy has operated as a Geotechnical Engineer in large open pit and underground operations where he was responsible for a team of engineers and the stability of the mine workings.