Monitoring and Numerical Modelling of Induced and

Triggered Seismicity in a Deep Sublevel-Stoping Mine Francesca De Santis (1,2), Yann Gunzburger (2), Vincent Renaud (1)

Isabelle Contrucci(1) , Pascal Bernard(3)

(1) Institut National de l’Environnement Industriel et des Risques (Ineris); (2) GeoRessources, Université de Lorraine; (3) IPGP

Objective of the study

2

Advantages Limits

Global in site

measurements

Precision of detection

and location

Real time monitoring No detection of

aseismic deformation

Insight into fractures

dynamic Local measures

Advantages Limits

Long-term prediction Needs lots of geo-

mechanical information

Simulation of different

excavation sequences

High computational

requirements

Global view of stress

changes

Why compare seismic and model data? Microseismic Monitoring

Real-time monitoring for short-term prevention

Numerical modelling Optimization of future excavations

Better understanding of interactions between stress changes and seismic activity generation

Gaining additional insights into failure processes

Improving long-term and short-term prevention

Finding possible links

Geographic and Historical context

Garpenberg mine

Sweden’s oldest mining area still in operation

Deposits first mined in in the 13th century

Acquired by Boliden in 1957 (300.000 tonnes/year)

Discovery of Lappberget orebody in 1998

Production moved in Garpenberg North in 2004 1200 Z

200 Z

800 Z

600 Z

400 Z

0 Z

1000 Z

North South

Schematic vertical profile of Garpenberg mine

Mined out areas

3

South

North

South

North

SWEDEN

zinc

silver

copper lead

gold

Extracted ore in 2016 2 622 000 t

Ineris monitoring network

Lappberget monitoring network (De Santis et al., 2017; Tonnellier et al., 2016)

3C

1C

Z [m

]

Lappberget

~ 3 km

Dammsjön

Kyrkan-Tyskgården

Gransjön

Kaspersbo

Blo

ck 1

25

0 L

app

ber

get

Production & Injuries

Amount of produced ore [kt] (1957-2014) Number of injuries (1996-2014)

0

500

1 000

1 500

2 000

2 500

1957 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 0

2

4

6

8

10

12

14

16

18

1996 1999 2002 2005 2008 2011 2014

Garpenberg mine

4

Improve automation and digitalization Mass mining methods

Microseismic monitoring

Mining method and sequence

5

Lappberget orebody

Sublevel stoping method with backfilling

Mining sequence

Z

Y

X

3D view between 2 levels Y

X

Primary drift

Secondary drift

Orebody profile

Plan view

Drift excavation: Development Blasts X

1107

1132

1157

1182

1207

1232

1257

13 12 11 10 9 8 7 6 5 22 21 20 19 18 17 16 15 14 23

2014

2015 - 2016

25

m

Stope excavation: Production Blasts Vertical section Z

November 2015 December 2016 Stope 13 Stope 13

Lappberget orebody

Lack of data:

Characteristics of fractures, joints or faults

Mechanical properties of rock types

No 3D information except for the orebody

Local geology

X

Z Y

Orebody

6

Lithology Dolomite

Limestone

Orebody

X

Y

X

Y

Very weak zones

Weak zones

Mechanical properties

Materials Young

modulus E [MPa]

Poisson’s ratio

n

Density r

[kg/m3]

Ore 66000 0.2

3030 Limestone 57000 0.18

Weak 20000 0.3

Very weak 2000 0.4

Weak rock masses

7

Analysis of Lappberget Microseismic Activity

between 2015 and 2016

Mining Production & Microseismic Activity

8

1000-3000m3 3000-4000 m3 4000-5000 m3 5000-6000m3 ≥ 6000 m3

Extracted rock mass volume:

Cu

mu

lati

ve M

SE n

um

ber

MSE

nu

mb

er

Time [year-month]

1 3 2 4 5

Microseismic swarms Mined Stopes

1

2

3 4 5

Stope 13

Space-time distribution

Mining Production & Microseismic Activity Cluster of MSE

9

Z [m

]

Central Cluster

Right Cluster

Mic

rose

ism

ic a

ctiv

ity

Min

e p

rod

uct

ion

1107

1132

1157

1182

1207

1232

1257

13 12 11 10 9 8 7 6 5 22 21 20 19 18 17 16 15 14 23

Central Cluster Right Cluster

Exploited Stopes 2014 - 2016

Period 2015 - 2016

X [m]

1e3

Central Cluster Right Cluster

MSE number & stopes volumes on a sliding window along X direction

Good correlation

Bad correlation

Central Cluster

Right Cluster

Bad correlation

Mining Production & Microseismic Activity Cluster of MSE

Z [m

]

Central Cluster

Right Cluster

Second period

Z [m

]

Right Cluster

Central Cluster

First period

First period Second period

1 2 3 4 5

Time [year-month]

Cu

m. M

SE n

um

ber

Cum. CC events

Cum. RC events

10

1

2

3 4 5

Stope 13

Seismic source characteristics Fracture size and Apparent Stress

11

Source radius ⇒ from 0.3 m up to 32 m

Median RC radius = 0.9 m

Median CC radius = 3 m

Fracture dimension Moment magnitude

Seismic moment [J]

Seis

mic

en

ergy

[J]

(Wyss and Brune, 1968)

Apparent stress ⇒ measure of stress release at a seismic source

Bigger stresses in the Right Cluster area

Seismic source characteristics Apparent stress

12

Level 1157 Level 1157

Contour maps of apparent stress

February 2015 - May 2016 June 2016 - December 2016

log10(meanAppStress) log10(meanAppStress)

First period Second period

Conclusions on seismic data analysis

13

Central Cluster Right Cluster

In agreement with production

in the area

Linked with remote blasting

in stope 13

Active since the beginning of

the monitoring

Activated at final stages of

stope 13 exploitation

+

Influenced by local geology

Bigger fractures Smaller fractures

Lower apparent stress values High apparent stress values

14

Numerical Modelling

Simulation of the exact mine sequence between 2015 and 2016

15

Numerical Model Geometry & Meshing

1. Galleries & Stopes

2. Weak & Very weak zones

3. Orebody

16

Numerical Model

Z Z = - 578

Z = - 1088

Upper levels mesh = 5 m

Z Z = 0

Z = - 1104

Z = - 1260

Z = - 1500

Upper levels Ore extension

Limits of the model

Box mesh = 20 m

Geometry & Meshing

Z = 0

Z = - 1500

Numerical Model Boundary Conditions & Mechanical Parameters

Initials & Boundary Conditions

Elastic parameters

Hoek-Brown parameters

Materials Young

modulus E [MPa]

Poisson’s ratio

n

m s sC

[MPa]

Density r

[kg/m3]

Ore 66000 0.2 10 0.112 188

3030 Limestone 57000 0.18 10 0.112 110

Weak 20000 0.3 1 0.001 30

Very weak 2000 0.4 0.63 0.00024 10

Paste 500 0.2 - - - 2000

17

Hoek-Brown failure criteria

⇒

Brittle failure criteria parameters

Materials mr s3b-d sr ξr bm b1

Ore 2 sC* (s)1/2 s*1e-5 0.0025 tan(15°) 750

Failure criteria

Brittle failure criteria

0

50

100

150

200

250

300

350

400

-0,002 0 0,002 0,004 0,006

Deviatoric stress

s1 - s3 (MPa)

Axial strain (-)Lateral strain (-)

s3 = 0.01 MPa

s3 = 1 MPa

s3 = 5 MPa

s3 = 10 MPa

s3 = 25 MPa

s3 = 50 MPa

s3 = 75 MPa

s3 = 60 MPa

(Souley et al., 2018)

18

Numerical Model Model Results

Deviatoric stress* [MPa]

X

Y

X

Y

X

Y

Step 15 – 2015/03/15

X

Y

Step 41 – 2016/08/03

Level 1182

Level 1182 Level 1182

Level 1182 Step 25 – 2016/02/09

Step 52 – 2016/12/28

*

Stope 13 Stope 13

Stope 13 Stope 13

19

Step 15 – 2015/03/15

Step 41 – 2016/08/03

Level 1182

Level 1182 Level 1182

Level 1182 Step 25 – 2016/02/09

Step 52 – 2016/12/28

X

Z

X

Z

X

Z

X

Z

Numerical Model Model Results

Deviatoric stress* [MPa]

20

Seismic Data and Model Data comparison

21

Seismic Data & Model Data 1st approach of comparison: individual measures

Seismic parameters Punctual measures at seismic

events locations

Model parameters Measures within the whole space

⇔

Model parameters calculated inside meshes at the position of MSE location

Seismic parameters

E – seismic energy

M0 – seismic moment

MW – moment magnitude

fc – corner frequency

sapp – apparent stress

Vapp – apparent volume

Δσ – stress drop

Model parameters

Stress variables Strain variables

s1 – max. principal stress

e1 – max. principal strain

s2 – inter. principal stress

e2 – inter. principal strain

s3 – min. principal stress

e3 – min. principal strain

sij – stress tensor eij – strain tensor

sq – Von Mises stress

eq – Von Mises strain

22

Seismic Data & Model Data Qualitative Analysis

Model parameter Seismic parameter

Mean deviatoric stress Mean apparent stress

sq [MPa] log10(meanAppStress)

Stope 13 Stope 13

23

Seismic Data & Model Data Quantitative Analysis: Principal Component Analysis

q

r = 0.0296

r = 0.0287

r = 0.0512

r = -0.0196

r = 0.0441 fc

E

M0

E

M0

E

M0

fc

E

M0

fc

q2 1/q

sqrt(q) fc

24

Seismic Data & Model Data 2nd approach of comparison: cumulated measures

Cumulated model parameters calculated inside the spheres of 10 m radius around each seismic events location

Seismic parameters

E – seismic energy

M0 – seismic moment

Va – max. principal stress

As – Source area

Model parameters

Elastic Plastic

Volumetric elastic energy

Volumetric plastic energy

Shear elastic energy

Shear plastic energy

Total elastic energy

Total plastic energy

Plastic volume

Mean Von Mises stress and strain

Mean of max. shear strain

Volumetric strain

Shear strain

Seismic parameters Punctual measures at seismic

events locations

Model parameters Measures within the whole space

⇔

25

Seismic Data & Model Data

Apparent volume ⇒

Plastic volume ⇒ volume of plastic zones in the spheres

Time [year-month]

Cu

mu

lati

ve p

last

ic v

olu

me

[m3]

Cu

mu

lati

ve a

pp

aren

t vo

lum

e [m

3]

1e4 1e6

Seismic parameter

Model parameter

CC-Va

CC-Vp

RC-Va RC-Vp

CC ⇒ Central Cluster RC ⇒ Right Cluster

Va

Vp

Time [year-month]

1e4

Cu

mu

lati

ve p

last

ic v

olu

me

[m3]

Cu

mu

lati

ve a

pp

aren

t vo

lum

e [m

3]

Seismic parameter

Model parameter

Qualitative Analysis

26

Seismic Data & Model Data

Source area ⇒

Volumetric strain ⇒

Time [year-month]

Time [year-month]

Cu

mu

lati

ve a

pp

aren

t vo

lum

e [m

3]

Cu

mu

lati

ve V

on

Mis

es s

trai

n [

mm

/m]

1e4

Cu

mu

lati

ve s

ou

rce

area

[m

2]

Cu

mu

lati

ve v

olu

met

ric

stra

in [

mm

/m]

As

εV

Seismic parameter

Model parameter

Seismic parameter

Model parameter

Apparent volume ⇒

Von Mises strain ⇒ Va

εq

Qualitative Analysis

)(t r e

27

Discussion and Perspectives

Seismic & numerical model shows increasing stresses around the main production area

Spatialization problem • Uncertainties in events location & model uncertainties • Reactivation of preexisting structure not taken into account in the model

Cumulating seismic and numerical parameters in space correlations are improved

Improvement of events detection and location • Detection: bigger number of seismometers and continuous recordings • Location: new location methods based on double difference and multiples analysis

Include fractures in the model • Bigger fractures of CC may be modeled based on focal mechanism of MSE • Smaller fractures of RC cannot be included in the model with the actual mesh ⇒ reduce mechanical properties of some meshes?

Taking creep into account • Seismic swarms are long lasting in time (more than 1 month) • Stress measurements have sown a differential response of deformations during time

Questions ? References

DeSantis, F., Contrucci, I., Lizeur, A., Tonnellier, A., Matrullo, E., Bernard, P., Nyström, A., 2017. Numerical approach for evaluating microseismic array performances: case study of a deep metal mine monioring network, in: Proceedings of the 9th International Symposium on Rockbursts and Seismicity in Mines.

Tonnellier, A., Bouffier, C., Renaud, V., Bigarré, P., Mozaffari, S., Nyström, A., Fjellström, P., 2016. Integrating microseismic and 3D stress monitoring with numerical modeling to improve ground hazard assessment, in: Proceedings of the Eighth International Symposium on Ground Support in Mining and Underground Constructions. eds. E. Nordlund et al., University of Technology, Luleå, Sweden.

Souley, M., Renaud, V., Al Heib, M., Bouffier, C., Lahaie, F., Nyström, A., 2018. Numerical investigation of the development of the excavation damaged zone around a deep polymetallic ore mine. International Journal of Rock Mechanics and Mining Sciences, 106, 165-175.

Wyss, M., Brune, J.N., 1968. Seismic moment, stress, and source dimensions for earthquakes in the California-Nevada region. J. Geophys. Res. 73, 4681–4694.

28

1000-3000m3 3000-4000 m3 4000-5000 m3 5000-6000m3 ≥ 6000 m3

Extracted rock mass volume for Production Blasts:

29

n° events per hour

cum. events number

Time [year-month]

Cu

mu

lati

ve M

SE n

um

ber

MSE

nu

mb

er

Mining Production & Microseismic Activity Space-time distribution

X ≈ 300 m3

Extracted rock mass volume for Development Blasts:

Pro

du

ctio

n B

last

s

Dev

elo

pm

en

t B

last

s