Post mining hazards: ground failures and surface consequences
In former underground mine districts, prior mining activity caused irreversible effects on the rock mass equilibrium. Hence, the stability of the ground, and surface and subsurface infrastructure can be affected by different forms of mine collapses and the risk of ground instability remains long after the mine closure.
To consider the potential impacts of underground mining on overlying structures in the study area of this thesis, the next sections describe surface and subsurface phenomena that can occur in the post-mining period, factors governing them as well as the causes and the mechanisms behind them.
Factors that affect rock mass stability in post-mining conditions
The nature of potential disturbances which can be expected during the post-mining period are related to several factors known to affect stability and govern underground and surface deformation and their specific configurations.
Particularly, the mining method has a strong impact on underground and surface deformation. In terms of expected failures around mining works, mining methods are typically classified in two categories: complete extraction methods, ensuring integral exploitation of the ore (i.e. longwall, sublevel stoping with backfilling…), and partial extraction methods, allowing the persistence of ore pillars after closure (i.e. abandoned room and pillar, dissolution cavities). Each technique can affect the surrounding rock in a different way. It has been observed that long-term impact on the ground surface is generally more significant in partial extraction mines with residual pillars leftover after mine closure than in complete extraction mines such as longwall, provided their voids are treated after extraction, for example with backfilling. This is especially the case for the room-and-pillar method, which is considered to be potentially the riskiest, from the perspective of long-term consequences (Schuchová & Lenart 2020). Impact of mining technique on the surrounding rock is governed by a combination of factors such as cavity geometry (e.g. height to width ratio of room, e.g., Bell & Genske 2001), the total volume of extracted material and/or pillar layout (F.T. Lee and F. Abel Jr. 1983, Didier et al. 2008). Over time, pillars may be affected by deterioration and weakening. For example, environmental factors such as flooding of mining works can induce degradation of mechanical properties of rock material (e.g., rock strength, rock stiffness) (Y. Yu et al. 2018). Furthermore, the stability of pillars depends considerably on their design (size and shape), which is based on the nature of the extracted material and its depth (Karfakis 1993). Decreased stability of pillars that are left underground to support the overburden impacts considerably the stability of the entire mine. Depth on mining excavation may represent a significant factor of impact on surface and subsurface deformation as well, especially in room-and-pillar mines (F.T. Lee and F. Abel Jr. 1983). The greater the depth of extraction, the longer it takes for rock deformation to reach the surface, which means that the shallower coal mines experience surface deformation earlier than deeper mines.
Therefore, old abandoned coal mines exploited by rooms and pillars often present a potential hazard as pillars collapse or voids migration to the surface, especially for mines at shallow depth.
The next section provides a description of typical failures such as subsidence or mining collapse of underground structures which can be expected in abandoned mines, with particular focus on mines exploited by the technique of rooms and pillars generalized collapse c) progressive subsidence. Adapted from Salmon et al.
A localized collapse (Figure 1.1a) is characterized by the sudden appearance of a collapsing crater at the surface. The horizontal extension of this phenomena generally varies between a few meters to a few tens of meters in diameter, while the crater depth depends mainly on the depth and dimensions of mine voids at its origin. The dimensions of the disorder and the brutality of its manifestation on the surface makes it potentially dangerous if it appears in the proximity of urbanized areas.
Depending on the initiating mechanism of the disorder they may take the forms of:
It refers to disorder created by the rupture of the roof of a gallery and the gradual ascent of a chimney that reaches the surface, following an initial cave-in within an underground excavation (i.e room), Appearance or this disorder is very frequent and observed in shallow mining works up to 50 m of depth;
• Rupture of the isolated pillar (s)
It refers to the collapse of one (or a few) pillar (s) that can lead to a collapse at the surface when the depth of the work and the thickness and stiffness of the overburden are up to 50 m.
Generalized collapses (Figure 1.1b), are manifested by the rupture, often dynamic (within a few seconds), of all or part of the underground mining works, thus affecting the stability of surface terrain. It can extend over areas up to several hectares. The depth of collapsed surface affecting the central part can reach several meters, or even several tens of meters in the case of a cavity collapse of saline dissolution (Franck et al. 2019). This is a rare phenomenon with consequences potentially very damaging to people and properties located on the surface. They may also be accompanied by earthquakes and blast effects which can project the materials through galleries and open wells to great distances. The main mechanisms or initiating scenarios are following:
• The abrupt collapse of abandoned pillars resulting from a ruptured roof
In underground mines with high extraction rates, large voids, undersized pillars or multi-level operations, where overburden consists of one (or more) stiff horizon(s) that may break suddenly. This creates a sudden overload on the pillars which break simultaneously. It is characterized by a sudden rupture of the overburden, in only a few seconds.
• Cascade-like failure of pillars
Pillars of an underground mine having reached the limit strength, affected by modification or development of a trigger. Overburden collapses following the underground cave-in front. It is a less brutal phenomenon than the previous one (a few minutes to a few hours), but potentially remains dangerous and can endanger the safety of people. Pillars of an underground mine having reached the limit strength, affected by modification or development of a trigger.
Both of these phenomena can be initiated several years or decades after the closure of a mine. The required configurations cause this phenomenon to be rather rare.
• Progressive subsidence
Progressive or continuous subsidence (Figure 1.1c) refers to the collapse of underground cavities that manifests itself on the surface with a “bowl-shaped” topographical depression without major failure, and it is characterized by a usually slow (progressively over several days or months, or even years), smooth and flexible readjustment of the surface.
Predisposition for this type of event concerns large-scale exploitations carried out at depths of several hundred meters and with significant horizontal extensions as well as a significant amount of residuals voids left after extraction. In the case of partial-exploitation type of mines (but not limited to), the main mechanism is described as delayed rupture of one or more supporting elements of the underground mining works (pillars, interlayers, roof, wall). The magnitude of phenomena is directly proportional to the opening of underground works, depending in particular on the depth of the work and the nature of methods of operation and treatment of voids (caving-in, backfilling, etc.). In the majority of cases, the maximum deflection observed at the centre of the “bowl”, during or after an operation, are decimetric to metric. Surface infrastructures are not so much affected by vertical movements but rather by the deformations of the ground (horizontal differential displacements, flexions, toppling…). The effects on the surface are weaker as exploitation is deeper. The effect on the surface can extend to larger areas than for mining works collapses. This phenomenon can be initiated several years or decades after the closure of the mine and is observed very often, depending on the size of the opening of the underground mining works, the exploitation technique, the extraction rate, the depth and the width of the exploited layer, the nature of overburden, the dip of layers, the surface topography, presence of faults, etc.
• Brittle subsidence
Brittle or discontinuous subsidence refers to a phenomenon that requires the following conditions: partially exploited mining works (by rooms and pillars in particular) located at depths of a few hundred meters, the existence of significant residual voids in the mining works and overburden mostly rigid and brittle throughout. It involves roof failure due to shearing along the support pillars of certain partial exploitations in very specific conditions. The pillars collapse due to overloading, followed by movement of the roof as well as the entire overburden all the way to the surface.
Characteristics of the brittle subsidence which differ from progressive subsidence include the possible development of a network of crevices along the periphery of the concerned panel, which can present structural risks for buildings located within the subsidence zone and promote much faster kinetics of the phenomenon.
Although the beginning of disorder underground generally lasts several days (fracturing and crushing of the pillars, preliminary fracturing of the roof), the manifestation on the surface can be sudden (rupture of a fragile roof) and accompanied by one or more seismic tremors.
The role and challenges of microseismic monitoring in post-mining period
As described in the previous Chapter, managing the abandoned mining sites often involves mitigation of the long-term risk of potential ground movements and instabilities.
As the mining works in post-mining period are often inaccessible due to flooding or bad conditions, surveillance of potential disorders is no longer possible with standard deformation measurements or seismic monitoring installed within the mine works (as is the case in active mining). Therefore, seismic monitoring devices are installed on the surface or in the sub-surface (Bennani et al. 2003). Even though microseismic monitoring has been used as a tool in active mining sites for almost a century, one of the first tests to adapt it to abandoned mining sites with risk of ground failures was done in the Lorraine iron basin by Ineris (Bennani et al. 2003 and references within).
Since then, microseismic monitoring plays an important role in post-mining risk mitigation. Nowadays, real-time microseismic monitoring is often applied to highly hazardous zones where remediation is not possible and risk cannot be reduced forever with other options such as backfilling due to its high cost (Couffin et al. 2003, Contrucci et al. 2008, 2011, 2019, Didier et al. 2008). In these cases, monitoring is used for the detection of precursory seismic events, e.g., failure initiation in rocks and expected ground instabilities. Additionally, seismic monitoring is one of the key tools in surveying the long-term hazard evolution in post-mining environments, which depends on several complexly interrelated factors (e.g., evolving meteorological and climatic conditions, mechanical degradation of the mine workings) that are difficult to take into account in geomechanical forecast approaches.
However, reliable hazard assessment and mapping often face challenges as mining areas can be large, which makes it difficult or in some cases impossible to forecast the location of potential failures, especially above older mines for which even the mine layout is not precisely known anymore (e.g., in Lorraine region in France, Bennani et al. 2003, Contrucci et al. 2019). For this reason, the determination of zones that should be prioritized for seismic monitoring can be difficult.
For economic reasons, seismic networks are generally rather sparse and very often consist of only a single antenna, such as borehole sensors, positioned at a place identified as having the highest risk. These single station networks do not have high performance in terms of seismic event location accuracy and are clearly focused on event detection. When seismicity appears in any part of these large mining areas, complementary temporal mobile seismic monitoring networks (such as surface stations) may be installed to improve location accuracy and better understand the origin of seismic sources. Due to the generally limited station number, the implementation of automatic processing routines (e.g., event location and classification) remains challenging. Locating of events is therefore mostly done manually, which requires longer processing time at the cost of real-time monitoring capacity.
Apart from a well-designed monitoring network setup, the performance of its data processing system makes another important difference, as mislocated events may lead to severe interpretation problems and communication of results toward the public.
In the case of seismic monitoring networks of limited station coverage, especially the ones composed of one-component geophones, an automatic detection and location method would lead to improvement of the detection performance of weak seismic signals, more accurate locations, magnitudes, and source parameters. This would be of great value especially in emergency situations where real-time operational monitoring is required.
Seismicity in post-mining districts
Earthquakes were for the first time associated with human activity in 1894 in Johannesburg (McDonald 1982) and were attributed to the Witwatersrand gold production mining production processes (Simpson 1986, McGarr et al. 2002 and references therein). Ever since, seismicity resulting from human activity (i.e., anthropogenic seismic activity) has been monitored and observed in numerous other mining settings, as well as in many other industrial operations.
Table of contents :
PART I GENERAL INTRODUCTION: POST-MINING SEISMICITY AND GARDANNE MINE CASE STUDY
1 State of the art regarding seismicity in post-mining conditions
1.1 Post mining hazards: ground failures and surface consequences
1.2 The role and challenges of microseismic monitoring in post-mining period
1.3 Seismicity in post-mining districts
2 Case study: Gardanne mine
2.1 A brief history of the mine and general context
2.2 Mining disorders, rockbursts, induced and natural seismicity
2.3 Groundwater management during mining period
2.4 Post-mining period
2.5 Post mining seismicity – the seismic swarm of Fuveau in Regagnas sector
2.6 Discussion on the origin on seismicity and seismic hazard
3 Thesis motivation and objectives
3.2 Methodology and objectives
PART II IMPROVEMENT IN DETECTION AND LOCATION OF MICROSEISMICITY WITH A SPARSE NETWORK
4 State of the art regarding detection and location of seismic events
4.1 Advantages of full-waveform and array coherency-based methods
4.2 Waveform-based detection and location methods: common basic principles
4.3 BackTrackBB (BTBB) method overview
5 New development for detection and location of events of sparse networks in the case of Gardanne
5.1 Testing BTBB parametrization and limitations
5.2 Step 1: Detection and first noise removal criteria using STA/LTA approach
5.3 Step 2: Location and second noise removal criteria using the amplitude-based approach
5.4 BTBB location
5.5 Local magnitude determination
5.6 Event classification and location quality assessment
6 Results – new catalogue of 2014-2017
6.1 Application of the new processing scheme to the 2014-2017 dataset
6.2 Classification scheme
6.3 𝒃−value estimation and magnitude of completeness
PART III CLUSTER AND MULTIPLET ANALYSIS
8 State of the art regarding clustering and multiplets
8.1 Clustering and underlying physical phenomena and mechanisms
8.2 Multiplets and repeaters
8.3 Repeat/Recurrence time
9 Spatial cluster analysis
9.1 K-means clustering method
9.2 Clustering of events of new catalogue 2014-2017
10 Multiplets and cluster activity 2010-2017
10.1 Cross-correlation technique
10.2 Identification of multiplet families using Fuveau station
10.3 The spatial resolution of multiplet analysis
10.4 Spatio-temporal distribution of multiplet families
10.5 Main multiplet families in the Fuveau swarm (study area)
10.6 Reconstruction of cluster activity before 2014
11 Detailed multiplet analysis 2014-2017
11.1 Multiplet families in new catalogue 2014-2017
11.2 Spatio temporal distribution of multiplets
12 Repeaters or multiplets?
12.1 Source parameters determination
12.2 First repeater identification method: seismic source overlap
12.3 Second repeater identification method: coherency analysis
13 Summary of main observations
PART IV SEISMICITY ORIGIN AND TRIGGERING MECHANISM
14 Origin of seismicity
14.1 The repetitive long term seismic activity
14.2 Depth and source mechanism
14.3 Conclusion about source origin
15 Seismicity-hydrology connection
15.1 Influence of the rainfall on the seismicity triggering
15.2 Influence of the pumping on seismicity migration
15.3 Conclusion about the hydro-seismic connection
16 Mechanics behind triggering and clustering of earthquakes
16.1 Triggering mechanism of the seismic activity
16.2 The secondary, driving mechanism behind clustering and multiplet families’ sequences
16.3 Interpretations of mechanism for each cluster
16.4 Conclusion about triggering
GENERAL CONCLUSION AND PERSPECTIVES