134
POSIVA OY Olkiluoto FI-27160 EURAJOKI, FINLAND Tel +358-2-8372 31 Fax +358-2-8372 3709 Hanna Remes Harri Kuula Petteri Somervuori Matti Hakala August 2009 Working Report 2009-55 ONKALO Rock Mechanics Model (RMM) Version 1.0

ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

P O S I V A O Y

O l k i l u o t o

F I -27160 EURAJOKI , F INLAND

Te l +358-2-8372 31

Fax +358-2-8372 3709

Hanna Remes

Har r i Kuu la

Pet te r i Somervuor i

Mat t i Haka la

August 2009

Work ing Repor t 2009 -55

ONKALO Rock Mechanics Model (RMM)Version 1.0

Page 2: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

August 2009

Base maps: ©National Land Survey, permission 41/MML/09

Working Reports contain information on work in progress

or pending completion.

The conclusions and viewpoints presented in the report

are those of author(s) and do not necessarily

coincide with those of Posiva.

Hanna Remes

Harr i Kuu la

Pet te r i Somervuor i

WSP F in l and L td

Matt i Haka la

KMS Haka la Oy

Work ing Report 2009 -55

ONKALO Rock Mechanics Model (RMM)Version 1.0

Page 3: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

ONKALO ROCK MECHANICS MODEL (RMM) - VERSION 1.0 ABSTRACT

The Rock Mechanics Model of the ONKALO rock volume is a description of the

significant features and parameters related to rock mechanics. The main objective is to

develop a tool to predict the rock quality and the potential for stress failure which can

then be used for continuing design of the ONKALO and the repository.

This is the first implementation of the Rock Mechanics Model and it includes sub-

models of the intact rock strength, rock mass spalling strength, in situ stress, potential

for stress failure, seismic velocities, thermal properties, major fracture sets, rock mass

quality and properties of the brittle deformation zones. Because of the varying quantities

of available data for the different parameters, the types of presentations also vary: some

data sets can be presented in the style of a 3D block model but, in other cases, a single

distribution represents the whole rock volume hosting the ONKALO.

Keywords: Rock mechanics model, Block model, Geostatistics, Rock mass quality, Rock strength, In situ stress, Spalling

Page 4: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

ONKALON KALLIOMEKAANINEN MALLI (RMM) - VERSIO 1.0 TIIVISTELMÄ

ONKALOn kalliomekaaninen malli on kuvaus kalliomekaanisesti merkittävistä omi-

naisuuksista ja parametreista. Työn tavoitteena on ollut kehittää työkalu, jolla voidaan

arvioida kalliolaatua ja mahdollisia jännitystilasta aiheutuvia vaurioita edelleen hyödyn-

nettäviksi ONKALOn tilojen suunnittelussa.

Tämä on ensimmäinen versio ONKALOn kalliomekaanisesta mallista, jonka osamalleja

ovat: lohkomalli kalliolaadusta, hauraat deformaatiovyöhykkeet ja niiden ominaisuudet,

kiven lujuus, kalliomassan vaurioituminen (spalling), kallion in situ jännitys, kallion

seisminen nopeus ja kallion termiset ominaisuudet. Koska eri parametrien lähtö-

tietomäärät vaihtelevat, eräät parametrit on esitetty 3D-lohkomallimuodossa, kun taas

eräät on esitetty yksinä jakautumina käsittäen koko ONKALOn kalliotilavuuden.

Avainsanat: kalliomekaaninen malli, lohkomalli, geostatistiikka, kalliolaatu, kallion lujuus, kallion in situ jännitys, kallion hilseily.

Page 5: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

1

TABLE OF CONTENTS

ABSTRACT

TIIVISTELMÄ

1 INTRODUCTION ................................................................................................. 3

2 INPUT DATA ....................................................................................................... 5

3 ROCK MASS QUALITY ....................................................................................... 7 3.1 Comparing pilot hole data with tunnel mapping data ................................... 9 3.2 Geostatistical analysis of the GSI and RQD rock mass classification indices ..................................................................................................... 13

3.2.1 Basic statistics ............................................................................... 14 3.2.2 Variogram modelling ..................................................................... 15

4 ROCK MECHANICS PROPERTIES OF MAJOR BRITTLE DEFORMATION ZONES (BDZS) .................................................................................................. 21

5 ROCK STRENGTH ............................................................................................ 25 5.1 Intact rock strength ................................................................................... 25 5.2 In Situ Spalling Strength ........................................................................... 26

6 IN SITU ROCK STRESS ................................................................................... 27

7 THERMAL PROPERTIES .................................................................................. 29

8 BLOCK MODEL ................................................................................................. 31 8.1 Block model dimensions and estimation method ...................................... 31 8.2 Block model results................................................................................... 33 8.3 Uncertainties in the RMM.......................................................................... 38

9 GEOPHYSICAL DATA....................................................................................... 41

10 SPALLING PREDICTION .................................................................................. 43 10.1 Spalling prediction for the ONKALO: Chainages 3117 - 4340 m ............... 43

11 FUTHER DEVELOPMENT AND PRELIMINARY IDEAS FOR VERSION 2 OF THE ROCK MECHANICS MODEL ..................................................................... 45

REFERENCES ........................................................................................................... 47

LIST OF APPENDICES .............................................................................................. 49

APPENDIX 1 .............................................................................................................. 51

APPENDIX 2 .............................................................................................................. 53

APPENDIX 3 .............................................................................................................. 57

APPENDIX 4 ............................................................................................................ 103

Page 6: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

2

Page 7: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

3

1 INTRODUCTION

The Rock Mechanics Model (RMM) is a description of significant features and

parameters related to rock mechanics. The main objective is to develop a tool to predict

rock quality and the potential for stress failure which can be used for the design of

ONKALO and repository facilities. This report provides an overview of the progress of

the RMM project.

The RMM is based on the current Geological Model (Mattila & al. 2007), i.e. no new

interpretation of the main geological features has been attempted. The adopted features

are the rock type domains, brittle deformation zones and the general orientation of

foliation. Other input data consist of the rock mass classification data from drillholes,

pilot holes and tunnels, rock mechanics laboratory and field test data, and geophysical

velocity and density data. The geometry is modelled using the Gemcom Surpac®

software and all data are stored in an Access database. Surpac supplies an interface for

the database and it is used as an interpretation and visualization tool.

Based on the input data, sub-models of intact rock strength, rock mass spalling strength,

in situ stress, stress failure, seismic velocities, thermal properties, major fracture sets,

rock mass quality, and the quality of the brittle deformation zones are made. Because of

the amount of available data and the nature of the parameter in question, the different

sub-models vary: some models can be presented in the style of a 3D block model; in

other cases one distribution represents the whole ONKALO volume.

The function of the model has also changed during the working period. In the beginning

of the project, the rock mass quality and the rock mass strength were considered as key

parameters but, during the development of the project, it transpired that the in situ

stress, spalling strength of the rock and the foliation are the most critical parameters. In

addition, the long fractures and brittle deformation zones are important and these can be

adopted from the geological model.

The phases in the RMM development project were

- Selection and evaluation of input data,

- Preparation of input data

- Geostatistical analysis of rock mass quality

- Block modelling

- Evaluation of block model results

- Spalling strength of rock, in situ state of stress

This is the first phase of the RMM development. In the next version, focus will be

placed on the prediction of stress damage/failure.

Page 8: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

4

Page 9: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

5

2 INPUT DATA

The geometry of the RMM is based on the current geological model. The adopted

geological features in the RMM 1.0 are the major brittle deformation zones. The

lithological units are included in the RMM but they are not used directly as rock

mechanics domains because the correlation between the rock mechanics data and the

rock types is not clear.

The input data for rock mass classification and the general orientation of foliation are

from surface drillholes (OL-KR1 – OL-KR40), pilot holes (OL-PH1 and ONK-PH2 –

ONK-PH8) and ONKALO tunnel mapping. These data include rock mass quality

indices (GSI, Q ) and the constituent Q parameters: RQD, Jn, Jr, and Ja. The tunnel

mapping was undertaken in 5 m or longer increments and the resultant data are

considered as the most reliable data for the rock mass quality. These tunnel mapping

data were available for this RMM 1.0 from the beginning of the ONKALO to chainage

3116 m.

Data from the surface drillholes were available from OL-KR1 – OL-KR40, representing

a total of 19840 m of logged core. The lengths of the logged sections in the surface

drillholes and ONKALO pilot holes varies from 0.05 m to 47.62 m, the average

sectional length being 5.9 m. Data from the pilot holes (OL-PH1 and ONK-PH2 –

ONK-PH7) are used only in the statistical studies to establish if the correlation factor

between drillhole data and tunnel mapping data could be determined. In the block model

estimation, data from OL-PH1 and ONK-PH2 – ONK-PH7 are not used because the

tunnel mapping data are available from the same locations.

These input data are stored in the RMM work database. RMM work database include

same data as in Posiva investigation database and some modified data tables. In

modified data tables brittle deformation zone intersections have been were filtered from

the data.

The intact rock strength data consist of point load tests from drillholes OL-KR1 – OL-

KR39 and laboratory test data from drillholes OL-KR1, OL-KR2, OL-KR4, OL-KR5,

OL-KR10, OL-KR12, OL-KR14 and OL-KR24.

The in situ stress field is taken directly from the interpretation as presented in Posiva’s

Site Report 2008 (Posiva 2009).

The thermal property data for the rocks at Olkiluoto consist of laboratory measurements

of drill cores (thermal conductivity, specific heat and density).

data from drillholes OL-KR1, OL-KR2, OL-KR4, OL-KR9, OL-KR11,

calculated thermal conductivity and specific heat from modal analysis of mineral

composition (OL-KR1, OL-KR2),

anisotropy of thermal conductivity from drill core samples (OL-KR1, OL-KR2),

thermal expansion properties of the Olkiluoto migmatitic gneiss estimated from mineral

composition (average) and in situ TERO measurements of thermal conductivity and

diffusivity (OL-KR2, and OL-KR14).

Page 10: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

6

The thermal property model is built from interpretation presented in Posiva‟s Site

Report 2008.

Geophysical drillhole logging data include density as well as P- and S-wave velocity

data from drillholes OL-KR1 – OL-KR4, OL-KR6 – OL-KR20, OL-KR22 – OL-KR48,

OL-PH1, ONK-PH2 – ONK -PH4, and ONK- PH6 – ONK-PH9 (Öhman et al. 2008).

Petrophysical sample data of the P-velocity and density are available from OL-KR1 to

OL-KR39, OL-PH1 and ONK-PH2 – ONK-PH07 (Öhman et al. 2009). The VSP

reflection seismic data were compiled to tomographic P- and S-wave velocity models

from drillholes OL-KR1, OL-KR2, OL-KR4, OL-KR8, OL-KR10, OL-KR14, OL-

KR27 and OL-KR38 (Appendix 3). The correlation of this large scale seismic velocity

model was reviewed (Appendix 4) against more detailed drillhole and sample data, and

against lithological, alteration and deformation intersections in these drillholes

(Paulamäki et al. 2006).

Page 11: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

7

3 ROCK MASS QUALITY

Rock mass quality is determined by using Q-classification (Barton et al. 1974; Grimstad

& Barton 1993) in the ONKALO access tunnel. Q- and Q´-value is calculated with the

following equations ((Barton et al. 1974; Grimstad & Barton 1993):

andSRF

J

J

J

J

RQDQ w

a

r

n

(3-1)

a

r

n J

J

J

RQDQ´ (3-2)

Where RQD = rock quality designation,

Jn = joint set number

Jr = joint roughness number

Ja = joint alteration number

Jw = joint water reduction number

SRF = stress reduction factor

The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and

Bawden (1995) provides a number which, when combined with the intact rock

properties, can be used for estimating the reduction in rock mass strength for different

geological conditions. This system for blocky rock masses is presented in Figure 3-1.

GSI can be estimated from Q‟-value with following equation (Hoek et al. 1995)

44)'ln(9 QGSI (3-3)

The GSI value varies between 0-100, 0 indicating bad and 100 indicating excellent rock

quality. When using equation 3-3 GSI may get values over 100 in exceptional good

rock.

Page 12: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

8

Figure 3-1. General chart for GSI estimates from the geological observations (Hoek & Marinos, 2000)

Page 13: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

9

3.1 Comparing pilot hole data with tunnel mapping data

The differences between drill core data and tunnel mapping data were studied to

determine whether the drill core data could be adjusted to be compatible with the tunnel

mapping data. This was done by comparing data from pilot holes (OL-PH1, ONK-PH2

to ONK-PH7) with the tunnel mapping data. For tunnel mapping data, the median

values of the mapped section for each parameter are used in comparison. The data from

ONK-PH8 were not used in the comparison because the tunnel mapping data in the

ONK-PH8 vicinity were not yet available.

The tunnel mapping data and pilot hole data were first compared using the original

logging sections (Figure 3-2). In this comparison, it is clearly seen that the observed

GSI-values vary more in the pilot holes. This is because the drill core logging sections

are defined based on changes of rock type, fracture frequency, deformation zones, etc,

and tunnel mapping is undertaken in 5 m or longer sections.

40

50

60

70

80

90

100

110

980 1030 1080 1130 1180

chainage

GS

I O

rig

ina

l

ONK-VT1

ONK-PH5

Figure 3-2. Example of the comparison of GSI-values from the pilot holes with the tunnel mapping values based on the original logging sections. The orange area represents the deformation zone intersection (ONK19C) observed in the tunnel; the grey area represents deformation zone intersection (ONK103) observed in ONK-PH5.

A second comparison was achieved by allocating the data from the pilot holes to

sections which were commensurate with the tunnel mapping, so that there would be a

more 1:1 sectional comparison. Parameters for rock mass quality (RQD, Jr and Ja) were

recalculated from the fracture logging data. Before recalculating the parameters,

fractures with high Jr and low Ja number (i.e. Jr = 3 or higher and Ja = 1 or lower) were

removed from the drillhole data, because these fractures are likely to be short ones. In

the newly constituted sections, the originally mapped values for the parameter Jn were

used. If two or more Jn values were originally logged in the newly constituted sections,

the Jn value from the longest section was used.

The comparisons between the ONK-PH5 and tunnel mapping results are presented in

Figures 3-3 to 3-7 as examples.

Page 14: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

10

40

50

60

70

80

90

100

110

980 1030 1080 1130 1180chainage

GS

I

ONK-VT1

ONK-PH5

Figure 3-3. GSI values from ONK-PH5 and tunnel mapping. The data in ONK-PH5 have been reconstituted so that the ONK-PH5 sections match the tunnel mapping sections. The orange area represents the deformation zone intersection (ONK19C) observed in the tunnel; the grey area represents the deformation zone intersection (ONK103) observed in ONK-PH5.

70

75

80

85

90

95

100

980 1030 1080 1130 1180chainage

RQ

D

ONK-VT1

ONK-PH5

Figure 3-4. RQD values from ONK-PH5 and tunnel mapping. The data in ONK-PH5 have been reconstituted so that the ONK-PH5 sections match the tunnel mapping sections. The orange area represents the deformation zone intersection (ONK19C) observed in the tunnel; the grey area represent the deformation zone intersection (ONK103) observed in ONK-PH5.

Page 15: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

11

0

1

2

3

4

5

6

7

8

9

980 1030 1080 1130 1180

chainage

Jn

ONK-VT1

ONK-PH5

Figure 3-5. Joint set number values from ONK-PH5 and tunnel mapping. The data in ONK-PH5 have been reconstituted so that the ONK-PH5 sections match the tunnel mapping sections. The orange area represents the deformation zone intersection (ONK19C) observed in the tunnel; the grey area represent the deformation zone intersection (ONK103) observed in ONK-PH5.

0

1

2

3

4

5

6

7

980 1030 1080 1130 1180Chainage

Ja

ONK-VT1

ONK-PH5

Figure 3-6. Joint alternation number values from ONK-PH5 and tunnel mapping. The data in ONK-PH5 have been reconstituted so that the ONK-PH5 sections match the tunnel mapping sections. The orange area represents the deformation zone intersection (ONK19C) observed in the tunnel; the grey area represent the deformation zone intersection (ONK103) observed in ONK-PH5.

Page 16: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

12

0

1

2

3

4

5

980 1030 1080 1130 1180Chainage

Jr

ONK-VT1

ONK-PH5

Figure 3-7. Joint roughness number values from ONK-PH5 and tunnel mapping. The data in ONK-PH5 have been reconstituted so that the ONK-PH5 sections match the tunnel mapping sections. The orange area represents the deformation zone intersection (ONK19C) observed in the tunnel; the grey area represent the deformation zone intersection (ONK103) observed in ONK-PH5.

In Figure 3-8, the differences in the GSI values between the tunnel mapping and PH

core logging data are presented. 50 % of the data values exhibit a difference of around

seven GSI units and 25 % of the data a difference of 12 units.

0

5

10

15

20

25

30

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190

Number of samples

GS

I T

un

ne

l -

GS

I P

H

25%50%

75%

Figure 3-8. Differences in GSI values between tunnel mapping data and pilot holes (One Q class corresponds to 8-12 units in the GSI value).

Page 17: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

13

It is not possible to say which measurement type, tunnel mapping or drill core logging,

providers a better estimate of rock mass quality, although the variation in rock mass

quality measured from the drill cores is higher. Of the compared parameters, the RQD

correlates fairly well with the tunnel mapping data; however, the joint alteration number

(Ja) and joint roughness number (Jr) do not correlate as well.

Because no single trend as found, it is not possible to establish any „correction‟ factors

to adjust the drillhole data to match the tunnel mapping data. Nevertheless, it can be

stated as an overall conclusion that the rock mass quality from drill cores varies as ± one

rock mass quality class (Q class) compared to the tunnel mapping data.

3.2 Geostatistical analysis of the GSI and RQD rock mass classification indices

Geostatistical analyses are used to determine interpolation/extrapolation distances for

the RMM input data. Variograms were also used to determine search ellipsoid

parameters used in the block model calculation.

Statistical and geostatistical analyses for the rock mass quality, GSI and RQD indices

were made using 1 m composites. In compositing original samples, which may vary in

size and length, are divided/combined so that each composite sample represents equal

length of drillhole. The available data from tunnel mapping, surface drillholes and

ONK-PH8 were used. Before calculating the composites, brittle deformation zone

intersections were filtered from the data (Figure 3-9). This filtering prevents low GSI-

values observed in the brittle deformation zones from spreading out the results. Filtering

was done using geological interpretations of the deformation zone intersections.

Figure 3-9. The ONKALO volume and two of the major brittle deformation zones (OL-BFZ 018 and OL-BFZ080. Brittle deformation zone intersections were filtered from the logging data before the composites were calculated.

Page 18: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

14

3.2.1 Basic statistics

To provide an overview of the available logging data, a basic statistical study was made

of the GSI and RQD values from tunnel mapping, surface drillholes and ONK-PH8

(Figure 3-10). Unlike GSI data from the tunnel, the drillhole data exhibit a bimodal

distribution. In terms of the RQD values, no major differences exist (Figure 3-11). The

data statistics are presented in Table 3-1.

Histogram GSI values

0

1000

2000

3000

4000

5000

6000

40 45 50 55 60 65 70 75 80 85 90 95 100 More

GSI

Nu

mb

er

Of

Sa

mp

les

0 %

20 %

40 %

60 %

80 %

100 %

120 %

Cu

mu

lati

ve

%

All data

Drill hole data

Tunnel data

Cumulative, all data

Cumulative, drill hole data

Cumulative, tunnel data

Figure 3-10. Histogram of GSI values from surface drillholes, ONK-PH8 and tunnel mapping data between chainage 0-3100.

Histogram RQD values

1

10

100

1000

10000

100000

10 20 30 40 50 60 70 80 90 100

RQD

Nu

mb

er

Of

Sa

mp

les

0 %

20 %

40 %

60 %

80 %

100 %

120 %

Cu

mu

lati

ve

%

All data

Drillhole data

Tunnel data

Cumulative, all data

Cumulative, drill hole data

Cumulative, tunnel data

Figure 3-11. Histogram of RQD values from surface drillholes, ONK-PH8 and tunnel mapping data between chainage 0-3116.

Page 19: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

15

Table 3-1. Statistics of GSI and RQD input data.

GSI RQD

All Tunnel Drillholes

(surface + PH8)

All Tunnel Drillholes

(surface + PH8)

No of samples 22,783 3,189 19,594 22,773 3,179 19,594

Min 34 47 34 10 50 10

Max 109* 106* 109* 100 100 100

Mean 83.82 85.62 83.53 96.10 97.85 95.83

Median 79.21 85.45 79.00 99.34 100 99.14

Std deviation 13.20 13.15 13.19 9.88 5.58 10.39

* GSI value calculated from equation 3-3.

3.2.2 Variogram modelling

The variogram used in geostatistics is a graphical tool which can be used to describe

spatial continuity of data (Isaaks & Srivastava 1989). It is a measure of variability and

its value increases when samples become more dissimilar.

In mathematical terms, the variogram for lag distance, h, is defined as the mean squared

difference of values separated by distance h:

)(2

))()(()(

2

hN

hxyxyh , (3-4)

where h represents a distance vector between two spatial locations and N(h) is the

number of pairs for the lag distance h.

The variogram range is the maximum distance at which there is some correlation

between the values of a parameter at two points, e.g. the distance from a drillhole at

which the drillhole information no longer provides any information. In other words,

there is no statistical relation for values located farther apart than the range. The

variogram plateau at the range is termed the sill (Figure 3-12).

Page 20: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

16

Figure 3-12.Characteristics of the Variogram.

Variograms were made using data from the surface drillholes and tunnel mapping both

together and separately. An omni-directional variogram, where data pairs are selected

based only on their separation distance and not their direction, is presented in Figure 3-

13. Although there is no clear plateau in this variogram, the effective range is about 30

m. Thus, on the average and for the measurements being considered here, knowledge of

the rock mechanics parameters at a given point cannot be reliably extrapolated for a

distance of more than 30 m.

Figure 3-13. Omni-directional variogram of GSI values for tunnel mapping and surface drillhole data, lag 3 m, the green line represents the variance of all the data.

Page 21: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

17

In directional sample variograms, it is often observed that the range and sill change as

the direction of sampling changes. Typical observed variogram behaviours can be

trends, fluctuations, and anisotropic dependence. The type of anisotropy where the

range changes while the sill remains constant is known as geometric anisotropy. A case

where the sill changes with direction, while the range remains constant is called zonal

anisotropy. (Isaaks & Srivastava 1989). Examples of oriented variogram behaviour for

three different two-dimensional geological structures are presented in Figure 3-14.

Figure 3-14. Three different geological images with the corresponding directional variograms. Note the fluctuations, trends, geometric anisotropy and zonal anisotropy. The red lines represents horizontal sample variograms and the blue lines represents vertical sample variograms; the black horizontal line represents the variance of the whole model data. (Gringarten & Deutsch 1999).

Page 22: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

18

The factor of anisotropy is significant for the rocks at Olkiluoto because they are

foliated, causing different mechanical properties to be exhibited parallel and

perpendicular to the foliation, and at the various orientations between these two end

cases. To study this anisotropy variograms were produced from data in the horizontal

plane and in the plane of foliation. For the overall foliation orientation, a dip of 32º and

dip direction of 138º were used (Mattila & al WR-92-2007), based on stereograms of

the foliation directions which indicate the average foliation orientations within the block

model area. In the plane of foliation, correlation is better compared to the direction

perpendicular to the foliation (Figure 3-15). In plane of foliation no clear anisotropy

direction, where range would be greater, were obtained.

Figure 3-15. Variograms of the GSI values in the direction of foliation (top) and perpendicular to the foliation (bottom). The red lines represent the best fit variogram models in the direction 32/138. Tunnel mapping and surface drillhole data; lag 7 m; the green lines represent the variance of all data.

Page 23: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

19

The top variogram in Figure 3-15, representing values in the direction of foliation,

provides better correlation as the distance between samples increases; whereas, the

variogram representing distances perpendicular to the foliation, the lower variogram in

Figure 3-15 bottom graph increases more quickly to the average variance, providing less

correlation as the distances between samples increases. The predictability is good for an

interpolation/extrapolation distance of 10 m or less in all directions. With an

interpolation/extrapolation distance of 30 m, predictability is good in the direction of

foliation but not in the direction perpendicular to the foliation.

Page 24: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

20

Page 25: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

21

4 ROCK MECHANICS PROPERTIES OF MAJOR BRITTLE DEFORMATION ZONES (BDZs)

In the RMM, brittle deformation zones (BDZs) are visualized and presented as 3D

planes adopted directly from the Geological model (Figure 4-1).

Figure 4-1. 3D view of the ONKALO with the major brittle deformation zones coloured based on their GSI-values.

Estimation of the mechanical properties of the BDZs is based on data from the Olkiluoto area drillholes and the ONKALO tunnel mapping. For this report, ten fracture zones have been analysed. Identification of the location and size of the zones is described in Kemppainen et al. (2007). As described in that report, each zone has been checked and described from those drillholes penetrating the zone being considered. All the intersection points are connected to each other using geophysical and hydrogeological information and, from those points, a 3D plane (to the upper and lower boundaries of the BDZ) is created using the Gemcom Surpac® program. Units have been modelled to date from the information in Mattila et al. 2007, Kemppainen et al. 2007, and Posiva 2007.

Page 26: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

22

Figure 4-2. Method of sub-dividing the brittle deformation zone as used by SKB and for the numerical modelling in Glamheden et al. 2007.

BDZs can be divided into the core part and transition zones, here the transitional and core components (Figure 4-2). Seven BDZs intersect the ONKALO tunnel in the 0 -2400 m tunnel chainage range. From three of these zones (BFZ18, BFZ43 and BFZ100) core zone and transitional zones have been mapped. From zone, BFZ15, only the core has been mapped. From zones BFZ11, BFZ101 and BFZ118, only the logged Q´ median value for a 5 m long chainage increment is available (Table 4-1).

Page 27: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

23

Table 4-1. Summary of the brittle deformation zones intersections.

Name of Brittle

Deformation ZoneOther name

Intersects ONKALO

tunnel at 0-2400

chainages

Mapped pre, core

and post zones

Intersections in drill

holes (number)

OL-BFZ11, OL-BFZ18,

OL-BFZ51ONK19A, HZ19C x - 22

OL-BFZ18, OL-BFZ60 ONK19C, HZ19C x x 20

OL-BFZ19, OL-BFZ21,

OL-BFZ98ONK20A, HZ20A - - 19

OL-BFZ98, OL-BFZ80,

OL-BFZ22ONK20B, HZ20B - - 17

OL-BFZ43 ONK43 x x 1

OL-BFZ10, OL-BFZ20,

OL-BFZ77, OL-BFZ84ONK56 - - 9

OL-BFZ100 ONK100 x x 6

OL-BFZ101 ONK101 x - 1 (pilot hole)

OL-BFZ15 ONK 103 x only core mapped 9

OL-BFZ118 ONK110 x - 1 (pilot hole)

The GSI values of the BDZs listed in Table 4-2 are based on either drill core logging or

tunnel mapping. In cases where the core has been established, the GSI value for the

BDZ is the value for the zone core. In cases where the GSI value (calculated from the

Q´ value) from tunnel mapping data is an average value for the mapped chainage, the

drillhole data are used. BDZs which do not intersect the tunnel are classified using the

drill hole logging information.

In the drillhole intersections, the GSI value is a lower quartile value of the mapped data.

Both approaches are conservative because the widths of the modelled zone are much

wider than the intersections. However, at this stage for rock mechanics modelling

purposes, it was decided to characterize the BDZs by the value of the weakest region

existing in the zone.

Page 28: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

24

Table 4-2. Strength and deformability properties of brittle deformation zones.

OL-BFZ11 OL-BFZ18 OL-BFZ19 OL-BFZ22 OL-BFZ43 OL-BFZ10 OL-BFZ100 OL-BFZ101 OL-BFZ15 OL-BFZ118

OL-BFZ18 OL-BFZ60 OL-BFZ21 OL-BFZ80 OL-BFZ20

OL-BFZ51 OL-BFZ98 OL-BFZ98 OL-BFZ77

OL-BFZ84

(ONK19A) (ONK19C) (ONK20A) (ONK20B) (ONK43) (ONK56) (ONK100) (ONK101) (ONK103) (ONK110)

Width of fault zone core (m) - 0.2 - - 0.15 - 0.8 - 5.0 -

Hoek Brown Classification

GSI 53 64 53 54 61 51 41 44 71 64

1st quartile

of drill core

intersections

mapped core

value from

tunnel

intersection

1st quartile

of drill core

intersections

1st quartile

of drill core

intersections

mapped core

value from

tunnel

intersection

1st quartile

of drill core

intersections

mapped core

value from

tunnel

intersection

mapped from

one drill core

intersection

1st quartile

of drill core

intersections

mapped from

one drill core

intersection

sigci (MPa) 22 22 22 22 22 22 22 22 22 22

mi 9.99 9.99 9.99 9.99 9.99 9.99 9.99 9.99 9.99 9.99

D 0 0 0 0 0 0 0 0 0 0

Ei (GPa) 63 63 63 63 63 63 63 63 63 63

Hoek Brown Criterion

mb 1.86 2.76 1.86 1.93 2.48 1.74 1.21 1.35 3.55 2.76

s 0.0054 0.0183 0.0054 0.0060 0.0131 0.0043 0.0014 0.0020 0.0399 0.0183

a 0.50 0.50 0.50 0.50 0.50 0.51 0.51 0.51 0.50 0.50

Failure Envelope Range

Application Custom Custom Custom Custom Custom Custom Custom Custom Custom Custom

sig3max (MPa) 28.5 28.5 28.5 28.5 28.5 28.5 28.5 28.5 28.5 28.5

Mohr-Coulomb Fit

cohesion (MPa) 3.2 3.8 3.2 3.3 3.7 3.1 2.7 2.8 4.2 3.8

friction angle (°) 19.0 21.7 19.0 19.2 20.9 18.6 16.4 17.0 23.4 21.7

Rock Mass Parameters

sigt (MPa) -0.06 -0.15 -0.06 -0.07 -0.12 -0.05 -0.03 -0.03 -0.25 -0.15

sigc (MPa) 1.6 3.0 1.6 1.7 2.5 1.4 0.8 0.9 4.4 3.0

sigcm (MPa) 4.1 5.3 4.1 4.2 4.9 3.9 3.1 3.4 6.4 5.3

Erm (GPa) 23.1 38.4 23.1 24.4 34.2 20.5 10.8 13.2 47.3 38.4

Kn = E / width (GPa/m) * - 192.1 - - 227.9 - 13.5 - 9.5 -

G = E / 2 (1+n), n = 0.25 (GPa) - 15.4 - - 13.7 - 4.3 - 18.9 -

Ks = G / width (GPa/m) - 76.9 - - 91.2 - 5.4 - 3.8 -

*) Width of the zone core varies in the drill core intersections. That is why the stiffness parameters has not been determined.

Page 29: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

25

5 ROCK STRENGTH

5.1 Intact rock strength

The Site Report 2008 (Posiva 2009) presents the distribution for the critical stress states

of the metamorphic rock types encountered at Olkiluoto (Figure 5-1). For the igneous

rock types, mainly pegmatite, the same critical stress states can be considered to apply

in compression, but the tensile strength is less.

Figure 5-1. Cumulative probability distributions for the normalised critical stress states for the ONKALO area metamorphic rock types.

0.0 %

12.5 %

25.0 %

37.5 %

50.0 %

62.5 %

75.0 %

87.5 %

100.0 %

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00

Normalized strength

Cu

mu

lati

ve

pro

ba

bil

ity

Peak strengthmean=115 MPa, n=77

Crack damage stressmean=99 MPa, n=82

Crack initiation stressmean=52 MPa, n=83

Point load Indexmean=118 MPa, n=972

Indirect tensile strengthmean=12 MPa, n=53

Page 30: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

26

For the RMM, the information in Figure 5-1 is represented by a triangular distribution

between the measured cumulative probabilities of 5%, 50% and 95% (Table 5-1). In this

first version of the RMM, this strength distribution is applied for all rock types, and no

spatial variation is assumed.

Table 5-1. Triangular distribution for the critical compressive stress states for the ONKALO volume rock.

lower limit mean upper limit

Uniaxial compressive strength (MPa) 81 115 157

Crack damage stress (MPa) 64 98 149

Crack initiation stress (MPa) 36 50 69

Indirect tensile strength (MPa) 6.7 11.9 16.8

5.2 In Situ Spalling Strength

For the in situ spalling strength, a reduction factor of 0.57 is applied to the uniaxial

compressive strength of the intact rock (Table 5-2). The value of 0.57 comes from in

situ testexperiments executed in Äspö and URL (Martin & Christiansson 2008). No

spatial variation is assumed.

Table 5-2. Triangular distribution for the in situ spalling strength of the ONKALO volume rock types (values are 0.57 of the values in Table 5-1).

lower limit mean upper limit

In situ spalling strength (MPa) 46 65 90

Page 31: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

27

6 IN SITU ROCK STRESS

The Site Report 2008 (Posiva 2009) presents values for the ONKALO volume in situ

state of stress in terms of the values of the three principal stresses as a function of depth

(Table 6-1). Similar to the characterization of the variability in rock strength presented

in Section 5, in the RMM and spalling analyses, a triangular representation is used for

the rock stress between the effective lower limit, mean and upper limit, i.e. the

cumulative probability for the lower limit is 0 %, 50 % for the mean value and 100 %

for the upper limit.

Table 6-1. Estimated mean, and lower and upper limits for the horizontal and vertical stress components at the Olkiluoto site for two stress domains corresponding to 0 to 300 m and 300 to 900 m vertical depth (Posiva 2009).

Range Stress component Vertical depth range [m]

H [MPa]

Mean zH 042.010

0 < z < 300 m

Lower limit zH 042.06

Upper limit zH 042.014

Orientation [°]

(mean; lower–upper) 10 (0 –20)

Mean zH 030.06.13

300 < z < 900 m

Lower limit zH 030.06.7

Upper limit zH 030.06.19

Orientation [°]

(mean; lower–upper) 90 (70 –110)

h [MPa]

Mean zh 0265.06

0 < z < 300 m Lower limit zh 0265.03

Upper limit zh 0265.09

Mean zh 015.045.9

300 < z < 900 m Lower limit zh 015.045.5

Upper limit zh 015.045.13

v [MPa]

Mean zv 0265.0

0 < z < 900 m Lower limit zv 0239.0

Upper limit zv 0292.0

z = depth below ground surface [m]

Page 32: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

28

Page 33: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

29

7 THERMAL PROPERTIES

The thermal properties investigated at Olkiluoto include thermal conductivity, specific

heat capacity, thermal diffusivity, density and thermal expansion coefficient. Site

Report 2008 presents the information for the Olkiluoto rocks. The samples used in the

laboratory measurements were mainly of veined gneiss, but also included granodiorite,

granite and pegmatite. In the RMM, the thermal properties of the veined gneiss are

assigned to the whole model.

As stated in the Site Report 2008 (Posiva 2009), the thermal expansion property of the

Olkiluoto migmatitic gneiss was studied with theoretical estimators, using the weighted

arithmetic mean of the mineral properties from literature data, as well as from numerical

modelling (Huotari & Kukkonen 2004). In the migmatitic gneiss, quartz and biotite

have the highest thermal expansion values. Theoretical estimations suggest linear

thermal expansion coefficient values in the range of 7 – 10 ·10-6

C-1

at 20-60º C, in

agreement with published data for similar rock types. For the RMM, a conservative

assumption of 9.5 10-6

C-1

is used.

The temperature-corrected mean thermal conductivity and diffusivity data for Olkiluoto

veined gneiss are shown in Table 7-1.

Table 7-1. Temperature-corrected drillhole means and standard deviations for the thermal properties of the Olkiluoto veined gneiss. The number of samples is shown in parentheses after the conductivity value at 22 ºC. Posiva 2009.

OL-KR1 OL-KR2 OL-KR4 OL-KR9 OL-KR11 All samples

Conductivity (W m-1

K-1

)

22ºC 2.84±0.84 (11) 3.09±0.52(135)1

2.83± 0.40 (8) 2.78±0.47 (8) 2.32±0.35 (8) 3.01±0.53 (170)

60ºC 2.74 2.86 2.64 2.68 2.24 2.78

98.7ºC 2.62 2.57 2.61 2.57 2.14 2.50

Specific heat

capacity (J kg-1

K-1

)

22ºC 738 732 730 746 738 737

60ºC 786 779 777 794 785 784

98.7ºC 834 ±26 827±15 825±12 842±15 833±16 832±19

Diffusivity (10-6

m2 s

-1)

Uncorrected 1.24±0.22 1.37±0.241

1.25±0.18 1.20±0.22 1.02±0.16 1.37±0.24

22ºC 1.40 1.56 1.42 1.35 1.14 1.50

60ºC 1.27 1.36 1.24 1.22 1.04 1.31

98.7ºC 1.14 1.15 1.16 1.11 0.93 1.11

Density (kg m-3) 2748±46 2707±51 2735± 26 2756±26 2753±21 2715±51

1) Includes samples in all measured directions; averages in direction of hole axis are 2.85 ± 0.81 W m-1

K-1

and 1.26 ± 0.19 10-6

m2 s

-1

Page 34: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

30

Page 35: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

31

8 BLOCK MODEL

8.1 Block model dimensions and estimation method

A block model covering the ONKALO volume has been developed (Figure 8-1). The

dimensions of the model are:

Y (m) X (m) Z (m)

Min co-ordinates 6791750 1525150 -600

Max co-ordinates 6792500 1526500 20

The basic block size is 10 x 10 x 10 m and the sub-block size is 2.5 x 2.5 x 2.5 m.

Figure 8-1. Location of the block model dimensions showing the ONKALO ramp and the drillholes from which the GSI and RQD input data estimations were made.

The GSI and RQD values in the block model were estimated using an inverse distance

with a power of 2 as the calculation method. Estimated values are based on the weighted

values of data points closest to the each block centroid. The weighting is the inverse of

distance of the data point from the block centroid raised to a power of two. No

lithological boundaries were used when estimating the block values.

The block model was calculated using 1 m composites determined from the surface

drillholes and tunnel mapping data. To avoid the lower RQD and GSI values from the

BDZ intersection zones for „spreading‟ to the surrounding rock mass, the BDZ

intersections were filtered from the data before calculating the values for the

composites. As a consequence, the estimated GSI and RQD values in the block model

represent the rock mass between the BDZs.

Calculations were made in three steps with different search parameters to classify the

level of confidence. In the first two steps, a spherical search was used when calculating

Page 36: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

32

the block values: the search radius in step one (class 1) was 10 m; and, in step two (class

2), it was 30 m. The minimum number of samples selected in steps one and two was 3,

and the maximum was 15. In the third step (class 3), all blocks which were not allocated

an estimated value in the first two steps were assigned a value. A search ellipsoid

oriented in the plane of foliation (32º/138º) was used. The search radius in the foliation

plane was set to infinity to make sure all the blocks would be estimated. In the direction

perpendicular to the plane of foliation, the maximum search distance was limited to 30

m. The maximum number of samples was increased to 50.

Attributes for in situ stress, intact rock strength and spalling strength were added to the

block model according to the data described in previous sections of this report. The

depth below ground surface, which is used when calculating the in situ stress

component, is calculated from the block centroid. Thermal properties (conductivity,

specific heat capacity and diffusivity) are added to the block model. Lithological units

were assigned to the blocks using 3D solid models from the lithology model. A

summary of the main attributes in the block model is presented in Table 8-1.

Table 8-1. Main attributes in the RM block model.

Attribute name Description

Block_depth Depth from ground level to block centroid

Classification GSI and RQD estimation class

Crack_damage, mean Mean crack damage stress for ONKALO area rock types

Crack_initiation, mean Mean crack initiation stress for ONKALO area rock types

Indirect_tensile, mean Mean indirect tensile strength for ONKALO area rock

types

Rock type Rock type

RQD Estimated RQD value

Sigma_hmax,_mean σH, mean horizontal stress component at Olkiluoto site

Sigma_hmax,_bearing Bearing for σH, mean horizontal stress component at

Olkiluoto site

Sigma_hmin,_mean σh, mean horizontal stress component at Olkiluoto site

Sigma_v,_mean σv, mean vertical stress component at Olkiluoto site

Uniaxial_compressive

strength, mean

Mean uniaxial compressive strength for ONKALO area

rock types

Spalling_strength Mean spalling strength for ONKALO area rock types

Conductivity_22 Conductivity at 22 ºC

Heat_capacity_22 Specific heat capacity at 22 ºC

Diffusivity_22 Diffusivity at 22 ºC

Th_exp Thermal expansion coefficient

Page 37: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

33

8.2 Block model results

The estimated GSI values are presented in Figures 8-2 to 8-4. Because the widths of the

major BDZs are small compared to the block size, the BDZs are presented as planes in

the model (see Section 4). The estimated GSI values vary from 50 to 110 (see Section 3)

which correspond to rock mass qualities of Poor to Exceptionally Good in the Q´

classification. Inside the tunnel blocks after about depth Z = -100 m (chainage 1100 m),

the rock mass quality (GSI value) increases. The mean GSI value inside the tunnel and

above Z = -100 is 73 (Good). Below Z = -100, the mean GSI value is 86 (extremely

Good) (Figure 8-2 and Figure 8-3). The mean RQD value inside the tunnel and above Z

= -100 is 93.8%. Below Z = -100, the mean RQD value is 97.6% (Figure 8-5 and Figure

8-6). The rock types and mean in situ stress component magnitudes are visualized in

Figures 8-7 to 8-11. In Appendix 1 and 2 more examples of the RMM visualization are

presented.

Figure 8-2. Block model with estimated GSI values inside the ONKALO tunnel.

50

70

90

110

-450-400-350-300-250-200-150-100-500

Z

GS

I

Figure 8-3. Estimated GSI values inside the ONKALO tunnel.

Page 38: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

34

The data presentation in Figures 8-2 and 8-3 show a clear cut trend in the rock quality –

as indicated by the GSI rock mass classification index: for depths of 0-125 m, the GSI

values are generally between 60 and 90; below 125 m, the GSI values are above 75.

This indicates that the rock mass near the surface, i.e. for depths 0-125 m, has been

reduced in mechanical quality, probably by the action of repeated glacial influence,

whereas the rock mass below 125 m depth has been unaffected and has a relatively

constant mean quality and variation with depth.

Similarly, and with reference to Figure 8-5, the mean RQD value inside tunnel above Z

= -100 m is 93.8% whereas below Z = -100 m, the mean RQD- value is 97.6%.

Figure 8-4. Estimated GSI values and major brittle deformation zones (the colours of the BDZs indicates their GSI-values).

Page 39: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

35

Figure 8-5. Estimated RQD value inside the ONKALO tunnel.

Figure 8-6. Estimated RQD values in the block model. The values represent the general rock mass (here the BDZs are not included in the blocks).

Page 40: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

36

Figure 8-7. Rock types inside the ONKALO tunnel ramp.

Figure 8-8. Assigned rock types in the block model.

Page 41: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

37

Figure 8-9. Mean horizontal in situ stress component σH around the ONKALO tunnel.

Figure 8-10. Mean horizontal in situ stress component σh around the ONKALO tunnel.

Figure 8-11. Mean vertical in situ stress component σv around the ONKALO tunnel.

Page 42: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

38

8.3 Uncertainties in the RMM

One of the main uncertainties in the model is incomplete coverage using the available

data. Although the average rock mass quality is well known, the estimation of local

anomalies is difficult. Based on the variogram studies, the maximum interpolation/

extrapolation distance for the GSI and RQD values is of the order of 30 m. Beyond this

distance, the level of confidence decreases. Block values estimated in the lowest class of

level of confidence are significantly averaged. Because the distances between drillholes

are relatively large compared to the block model dimensions, 92.8% of the block model

volume falls in the lowest class of level of confidence (Figures 8-12 and 8-13) –

although this refers to the blocks some distance away from the ONKALO tunnel.

Figure 8-12. Search radius in different calculation phases. Grey line presents drillholes i.e. input data.

Figure 8-13. Search radius in different calculation phases in tunnel blocks.

Page 43: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

39

In Figures 8-14 to 8-16 are presented GSI values inside the tunnel blocks compared to

the original tunnel mapping data. The estimation of block values was made using data

only from the surface drillholes and pilot holes. The maximum search distance was 30

m, which corresponds to the first two classes in level of confidence. The accuracy of the

drillhole estimates of rock mass quality is ± one Q class.

Figure 8-14. Estimated GSI values from drillholes (left) compared to the originally mapped rock mass quality in the tunnel section (right) between chainage 0-760 m. The grey area represents areas where no sufficiently reliable geostatistical prediction can be made.

Figure 8-15. Estimated GSI values from drillholes (left) compared to the originally mapped rock mass quality in the tunnel section (right) between chainages 760-1820 m. The grey area represents areas where no sufficiently reliable geostatistical prediction can be made.

Page 44: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

40

Figure 8-16. Estimated GSI values from drill hole data (left) compared to the originally mapped rock mass quality in the tunnel section (right) between chainages 1740-3000 m. The grey area represents areas where no sufficiently reliable geostatistical prediction can be made.

Page 45: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

41

9 GEOPHYSICAL DATA

It was suggested that the seismic travel times from existing 3D VSP investigations

(surface drillhole reflection) conducted during years 1989-2005 could be used in

compiling a travel time tomographic representation of the near-ONKALO volume. The

target was set to estimate whether 5…20 m scale rock properties can be characterized

from the seismic travel time tomographic results.

The VSP data reprocessing includes gathering of first arrival times, inversion

computation with developed 2D bend tray tracing method and the tomographic

computations. In order to review the results against the drillhole geophysical data and

geological logging data the tomograms were transferred into Gemcom Surpac®

software. (Appendix 3.)

Conclusions based on the tentative processing phase of the VSP traveltime tomography

is that the technique is feasible for displaying medium scale (20-50 m thick layer)

variation in the rock mass, mostly due to lithological variation. Interpretation of velocity

anomalies in the model and an associated report are presented in Appendix 4.

At this stage, the geophysical data are not used in the RMM because more studies of the

correlation between rock mass quality and P-wave velocity need to be undertaken.

Page 46: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

42

Page 47: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

43

10 SPALLING PREDICTION

The spalling prediction is based on methodology presented by Martin & Christianson

(2008) and Hakala et al. (2008). The distribution for the maximum tangential stress

around the excavation is calculated based on the distribution of in situ stress, excavation

geometry, orientation and depth location. The probability of spalling is presented in the

form of the probability of the Factor of Safety (FOS) values. The FOS is calculated via

equation 10-1:

FOS = Factor of Safety = SPALLING / (10-1)

where SPALLING = spalling strength of the rock

= maximum tangential stress

The probability of spalling is the probability of FOS<1, and it is defined by three

categories based on the mean value of the FOS:

FOS Probability of spalling

>1.25 No

1.25-1.0 Minor or No

< 1.0 Extensive

The severity of spalling is defined by the spalling depth, which can be estimated by the

empirical equation 10-2 or from the results of numerical simulation by equation 10-3.

For simplicity, equation 10-2 is used in the RMM v 1.0.

Sd = a (0.5 ( / SPALLING) – 0.52) (10-2)

where Sd = spalling depth

a = average radius of excavation profile

Sd = depth where ( 1 - 3) > SPALLING (10-3)

where 1 = major principal stress

= minor principal stress

10.1 Spalling prediction for the ONKALO: Chainages 3117 - 4340 m

For the RMM version 1.0, the spalling prediction was made only for chainages 3117-

4340 m. The chainages considered were divided into five sections (Table 10-1). The

elastic tangential stresses around the excavations were calculated with Examine3D, a

three dimensional boundary element code. The spalling probability and depth

calculation for each tunnel section were obtained by Monte Carlo simulation with

20,000 realisations.

Page 48: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

44

The results indicate minor spalling for the turn and the narrow profiles after the turn

(Table 10-1 and Figure 10-1).

Table 10-1. Tunnel sections for spalling estimation and the resulting values.

Profile from Depth

(m)

to Depth

(m) mean FOS PoS* Sd (m)

normal -295 -365 1.37 6 % 0.10

niche -295 -365 1.48 2 % 0.07

turn -365 -373 1.14 22 % 0.13

normal -373 -416 1.23 14 % 0.12

niche -373 -416 1.29 8 % 0.10

* PoS = probability of spalling

Figure 10-1. Spalling prediction for ONKALO chainages 3117-4340 m, i.e. contract phase TU4.

Page 49: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

45

11 FUTHER DEVELOPMENT AND PRELIMINARY IDEAS FOR VERSION 2 OF THE ROCK MECHANICS MODEL

A fully featured and functioning RMM Version 2.0 should include more extensive

estimates of the secondary stresses and spalling around the excavations. Also, the

foliation information should be used in more detail. Moreover, efforts should be made

to use the rock type, foliation and geophysical information to see if the spatial variation

of rock strength can be estimated through these parameters.

The geometry and rock mechanics properties of the major BDZs should be updated. The

BDZ intersections in drillholes and the ONKALO tunnel should be further interpreted

also from the rock mechanics point of view.

In future, the RMM will be extended to the final disposal repository volume. In these

areas, the rock mechanics properties and the effect of smaller scale BDZs and long

fractures (TCF=Tunnel Crosscutting Fractures) should also be considered.

The geophysical tomography data will be studied more carefully and, if possible and

depending on the current studies, it will be used in the RMM Version 2.

Further detailed geostatistical analyses of rock mass quality will be considered. For

example, statistical analysis in the different rock mechanics domains could be

considered.

Thermal property model will be updated when more data is available. Possible

anisotropy should also be considered in RMM.

Page 50: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

46

Page 51: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

47

REFERENCES

Barton, N. Lien & R. Lunde, J. 1974. Engineering classification of rock masses for the

design of tunnel support. Rock Mechanics, Vol 6, No 4, p. 189-236.

Glamheden, R., Hansen, L. M., Fredriksson, A., Bergkvist, L., Markström, I., Elfström,

M. 2007. Mechanical modelling of the Singö deformation zone. SKB report R-07-06,

Stockholm. 85 p.

Grimstad, E. & barton, N., 1993. Updating of the Q-system for NMT. Proceedings of

the International Symposium of Sprayed Concrete, 18-21 December 1993. Fagernäs,

Norway. Kompen, E. Opsahl & Berg. Norwegian Concrete Association, pp. 46-66.

Gringarten, E. & Deutsch, C. V., 1999. Methodology for Variogram Interpretation and

Modelling for Improved Reservoir Characterization, Paper SPE 56654 presented at the

SPE Annual Technical Conference and Exhibition held in Houston, Texas, Oct 3-6

1999, 13 p.

Hakala, M., Hudson, J.A., Harrison, J.P. & Johansson, E. 2008. Assessment of the

Potential for Rock Spalling at the Olkiluoto Site. Posiva Working Report 2008-83.

Hoek, E., 1994. Strenght of rock and rock masses, ISRM News Journal, 2(2), 4-16

Hoek, E., Kaiser, P. K. & Bawden, W.F. 1995. Support of Underground Excavations in

Hard Rock. Balkema, Rotterdam, 215 p.

Hoek, E. & Marinos, P., 2000. GSI: a geologically friendly tool for rock mass strength

estimation. In: Proceedings of the GeoEng2000, international conference on

geotechnical and geological engineering, Melbourne, Technomic publishers, Lancaster,

pp 1422–1446

Huotari, T. & Kukkonen, I. 2004. Thermal expansion properties of rocks: Literature survey and estimation of thermal expansion coefficient for Olkiluoto mica gneiss. Posiva Oy, Working Report 2004-04.

Isaaks, E. H. & Srivastava, R. M. 1989. An Introduction to Applied Geostatistics.

Oxford University press, New York, 561 p.

Kemppainen, K., Ahokas,T., Ahokas, H., Paulamäki, S., Paananen, M., Gehör, S. &

Front, K. 2007. The Onkalo Area Model, version 1. Working Report 2007-71. Posiva

Oy, Eurajoki, 141 p.

Martin, C. D. & Christiansson, R. 2008. Estimating the potential for spalling around a deep nuclear waste repository in crystalline rock. Int. J. Rock Mech. Min. Sci., 46, pp. 219-228

Page 52: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

48

Mattila, J., Aaltonen, I., Kemppainen, K., Wikström, L., Paananen, M., Paulamäki, S.,

Front, K., Gehör, S., Kärki, A. & Ahokas, T. 2007. Geological Model of the Olkiluoto

Site, Version 1.0. Working Report 2007-92. Posiva Oy, Eurajoki. 514 p.

Paulamäki, S., Paananen, M., Gehör, S., Kärki, A., Front, K., Aaltonen, I., Ahokas, T.,

Kemppainen, K., Mattila, J., Wikström, L., 2006. Geological Model of the Olkiluoto

site, Version 0, Working Report 2006-37. Posiva Oy, Eurajoki, 355 p.

Posiva 2007. Olkiluoto Site Description 2006, Posiva Report 2007-03

Posiva 2009. Olkiluoto Site Description 2008, Posiva Report 2009-01

Öhman, I., Heikkinen, E., Säävuori, H., Vuorinen, S., Paulamäki, S., Aaltonen, I. 2009.

Summary of petrophysical analysis of Olkiluoto core samples 1990 - 2008. Working

Report 2009-11. Posiva Oy, Eurajoki. 212 p.

Öhman, I., Palmén, J. & Heikkinen, E. 2008. Unification of acoustic drillhole logging

data. Working Report 2008-18. Posiva Oy, Eurajoki. 57 p.

Page 53: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

49

LIST OF APPENDICES

APPENDIX 1

GSI and RQD values in a defined sub-horizontal section

APPENDIX 2

Search radii and GSI and RQD values in a defined vertical section

APPENDIX 3

Cosma, C., Cozma, M. & Enescu. N. 2008. P- and S-wave 3D velocity inversion of

VSP data collected at Olkiluoto. A memorandum.

APPENDIX 4

Heikkinen, E. 2009. Review of Seismic 3D Velocity Model Based On 3D-VSP Data. A

memorandum.

Page 54: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

50

Page 55: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

51

APPENDIX 1

Figure A1-0-1. Defined sub-horizontal section

Figure A1-0-2. Search radius in different calculation phases (left) and estimated GSI-values and major brittle deformation zones (right) at defined section. Small dots in the left figure represent drillhole intersections.

Page 56: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

52

Figure A1-0-3. Estimated RQD values (left) and assigned rock types (right) at defined section. Major brittle deformation zones are not shown in the Figures.

Page 57: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

53

APPENDIX 2

Figure A2-0-1. Defined vertical section.

Figure A2-0-2. Search radii in the different model calculation phases.

Page 58: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

54

Figure A1-0-3. Estimated GSI values and major brittle deformation zones at defined vertical section.

Figure A1-0-4. Estimated RQD values at defined vertical section. Major brittle deformation zones are not shown in the Figure.

Page 59: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

55

Figure A2-0-5. Assigned rock types at defined vertical section.

Page 60: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

56

Page 61: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

57

VIBROMETRIC

APPENDIX 3

P- and S-wave 3D velocity inversion

of VSP data collected at Olkiluoto

Vibrometric Oy

2008

Page 62: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

58

VIBROMETRIC

Page 63: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

59

VIBROMETRIC

TABLE OF CONTENTS

1 THE PURPOSE OF THE STUDY ...................................................................... 61

2 THE EQUIPMENT ............................................................................................. 63 2.1 The seismic source ................................................................................... 63 2.2 The seismic receivers ............................................................................... 64

2.2.1 The Recording station ................................................................... 65

3 THE TOMOGRAPHIC INVERSION METHOD ................................................... 67

4 THE ELASTIC PARAMETERS EQUATIONS .................................................... 73

5 CONCLUSIONS ................................................................................................ 75

APPENDIX 1 – SUMMARY OF THE VSP BOREHOLES USED FOR THE CURRENT STUDY ....................................................................................................................... 77

APPENDIX 2 – 3D VSP TOMOGRAMS...................................................................... 79 P wave tomograms ............................................................................................ 79 S wave tomograms ............................................................................................ 87 Tomograms obtained by simultaneous inversion of VSP data from several boreholes ........................................................................................................... 91

APPENDIX 3 ............................................................................................................ 103

Page 64: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

60

VIBROMETRIC

Page 65: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

61

VIBROMETRIC

1 THE PURPOSE OF THE STUDY

The objective of this study has been to determine the distribution of the P and S seismic

velocities within a rock volume covered by VSP measurements and to use them to infer

elastic parameters. Eight boreholes investigated by VSP have been selected based on

their location and data quality. The measurements were carried out between years 1990

and 2005, by Vibrometric Oy.

Figure A3- 1. Location of the drillholes (labeled blue triangles) and shot points (red circles). The projection of the Onkalo tunnel is represented by the black polygon.

Page 66: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

62

VIBROMETRIC

Page 67: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

63

VIBROMETRIC

2 THE EQUIPMENT

2.1 The seismic source

Surveys conducted prior to the year 2000 used dynamite charges of typically 125 or 250

grams. After that, the VIBSIST-1000, a time-distributed multi-impact source, replaced

the dynamite, see Table A3- 2. The VIBSIST-1000 seismic source was mounted on a

wheeled or tracked excavator equipped with hydraulic breaker (Figure A3-2).

The dynamite charges were fired in boreholes, directly coupled to the rock, whilst the

VIBSIST-1000 source was coupled through the overburden. The different coupling lead

to differences in the signal characteristics, e.g. slightly delayed arrivals with the

VIBSIST-1000. These were compensated in order to make the measurements

comparable.

Figure A3- 2. The Vibsist-1000 seismic source used at Olkiluoto.

Page 68: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

64

VIBROMETRIC

2.2 The seismic receivers

Several versions of the R8-XYZ-C geophone chain were used through time with the

VSP surveys at Olkiluoto (Figure A3-3). However, the essential specifications remained

largely the same, all versions consisting of eight 3-component modules with the Z-

component directed along the hole and the X- and Y-components perpendicular to the

Z-component and to each other. Each module is equipped with a side arm for clamping,

contains three 28-Hz geophones and down-hole preamplifiers. The frequency response

extends upwards to 1000 Hz, which is largely sufficient for both dynamite- and

VIBSIST- generated signals.

Figure A3- 3. The R8-XYZ-C geophone receiver chain.

Page 69: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

65

VIBROMETRIC

2.2.1 The Recording station

A PC-based acquisition system (Figure A3-4) has been used, equipped initially with 16-

bit A/D converters and later upgraded to 24-bit converters in the year 2000.

Figure A3- 4. The acquisition station used for the surveys at Olkiluoto.

Page 70: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

66

VIBROMETRIC

Page 71: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

67

VIBROMETRIC

3 THE TOMOGRAPHIC INVERSION METHOD

A well-conditioned tomographic velocity-inversion exercise would require the rock to

be probed evenly and from view angles as diverse as possible. 3D inversion can in

principle be achieved by conducting multiple cross-hole surveys among densely drilled

set of boreholes. However, only two cross-hole tomographic surveys have been

performed at Olkiluoto: in 2002, between boreholes KR14 and KR15 and in 2003,

between boreholes KR04 and KR10. High resolution tomographic images of the P-wave

fields were produced, in local scale.

The current attempt to re-use the VSP data collected previously, as a part of the site

characterization programme, proves the feasibility of comprehensive mapping of the

velocity fields, hence of the elastic parameters, at the entire site scale, even if the

resolution is poorer than the one obtained by cross-hole tomographic inversion, as it can

be seen in Figure A3- 5.

Figure A3- 5. 3D P-wave velocity field computed now from the KR04 VSP data, together with the cross-hole tomography panel between boreholes KR04 and KR10, as calculated in 2003. View from East. Velocity range: 5650 m/s (blue)-5850 m/s (red).

Page 72: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

68

VIBROMETRIC

The analysis done within this study consisted essentially of the following:

1) Independent tomographic inversions for each VSP shot gather, in the plane

defined by the respective borehole and the respective shot. This approach

produced fairly realistic velocity distributions with a good match of the synthetic

travel times computed on these distributions and the actual travel times (Figure -

A3- 7). However, the velocity distributions obtained independently for each shot

do not produce the same solution along the borehole. Therefore, such solution is

deemed unsatisfactory.

Figure A3- 6. Typical setup for tomographic inversion from VSP data. Example from KR08, six shot points.

Figure A3- 7. Reduced velocity shot gathers from borehole KR08 with P wave first arrivals (red line) modeled on the inverted velocity models computed independently on each shot gather.

Page 73: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

69

VIBROMETRIC

The VSP geometry produces ill conditioned tomographic images because each inversion

panel is a single shot-gather, the result being essentially a velocity vs. depth chart. This

representation is nonetheless a valid expression of the velocity distribution within the

plane determined by the source and the receiver array. The validity of this

representation is demonstrated by plotting synthetic arrival times computed on the

velocity model against the data profiles as presented in Figure A3- 7. It can be noticed

that the synthetic first arrivals line up very closely with the real data, which gives

confidence to both computed ray pattern and velocity solution. The cell size (horizontal

x vertical) for this step was equal to 10 m x 5 m, for both P and S waves.

Page 74: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

70

VIBROMETRIC

Figure A3- 8. P-wave field reconstructed from the KR01 VSP data – before (up) and after (down) adding the sonic logs processed data. View from South. Velocity ranges from 5000 m/s (blue) to 6000 m/s (red).

Page 75: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

71

VIBROMETRIC

2) Simultaneous tomographic inversions for all VSP shot gathers measured in the

same borehole. This leads to a 3D tomographic solution consisting of several

planes, one for each shot point. At this stage the solutions obtained for different

shot points become consistent with each other, as they are obtained by

simultaneous inversion. However, the first arrivals computed on this velocity

model became less consistent with the actual first arrivals, when compared with

Step 1. The cause is related to shot-specific time delays produced either by local

overburden conditions and/or by the different definition of the zero-time for the

measurements done with explosive sources and the VIBSIST-1000. The

subtraction of the shot-specific time delays deemed as „static‟ corrections‟

improves the quality of the solution (Figure A3- 8). Sonic log data has been used

to aid with the computation of these delays, the rationale being that along the

borehole both VSP and sonic data should produce similar velocity distributions,

while the causes of the delays are confined near the surface. The cell size

(horizontal x vertical) for this step was equal to 20 m x 10 m, for both P and S

waves.

3) Simultaneous tomographic inversion of VSP data from all the eight holes, with

delay-reductions similar to Step 2. The cell size (horizontal x vertical) was also

equal to the one used in Step 2, i.e. 20 m x 10 m for both P and S waves. This

approach produces the most stable solution, although, somehow unexpectedly,

also the smoothest. A tentative conclusion is that more local velocity variations

are in fact possible measuring artefacts and that a consistent and reliable image

obtainable from the VSP measurements is in fact quite smooth (Figure A3- 9).

Figure A3- 9. 3D P-wave velocity field derived by simultaneous tomographic inversion of VSP data from all the eight holes.

Page 76: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

72

VIBROMETRIC

The data processing steps are described in Table A3- 1.

Table A3- 1. The data processing steps.

Item Processing step

1 Data transfer to Vibrometric format. Data conditioning & coordinates

2 Band-pass filtering 20 – 200 Hz

3 Rotate the X, Y, Z components to the Axial, Radial and Transversal

components

4 P wave arrival times picking on the Axial component

5 S wave arrival times picking on the Transversal component

6 Integrate the sonic logs P arrival times with the seismic ones

7 3D ray tracing based on the arrivals calculated at step 6

8 P wave tomographic inversion for each borehole, with ray paths computed

at Step 7

9 P wave tomographic inversion for all the boreholes, with modelled arrival

times computed at Step 8

10-12 Repeat steps 7 - 9 for the S waves arrival times

13 Calculate the elastic parameters based on the P and S waves tomograms

14 Model the elastic parameters as arrival times

15 Run the inversion for elastic parameters

Page 77: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

73

VIBROMETRIC

4 THE ELASTIC PARAMETERS EQUATIONS

Poisson‟s ratio: )(2

2

22

22

SP

SP

vv

vv

Shear modulus: 2

Sv

Young‟s modulus: )1(2E

For the Shear and Young‟s modulus, a constant density ( ) value of 2780 kg/m3 has

been used.

Page 78: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

74

VIBROMETRIC

Page 79: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

75

VIBROMETRIC

5 CONCLUSIONS

It is possible to use the VSP data for obtaining 3D tomographic maps of the P- and S-

wave velocities and of the corresponding elastic parameters. The volume considered in

this study is approximately 1000 m x 1000 m x 800 m, see Figure A3 -1. The site-

integrated results are presented in Appendix 2, in Figure A3- 22, Figure A3- 25, Figure

A3- 28, Figure A3- 30 and Figure A3- 32.

Main facts:

1) Difficulties: the main problem of this approach was represented by the relative

low 3D cube coverage by seismic signals ray paths, see paragraph “The

tomographic inversion method”.

2) Resolution: cell size (horizontal x vertical) was equal to 10 m x 5 m for the

processing step no. 7 and 20 m x 10 m for the processing steps no. 8 and 9.

3) Reliability: the synthetic arrival times computed on the velocity model line up

very closely with the real data, increasing the approach level of confidence, see

Figure A3- 7.

4) Sonic log information: proved to be useful but not mandatory, see Figure A3- 8.

Applied tomography algorithm: modified SIRT.

Recommended further works: a total of 19 boreholes have been used for VSP

measurements in the Olkiluoto area. The current study has focused on 8 boreholes in the

region of the ONKALO tunnel, see Figure A3- 1. The future work should integrate

some more boreholes in the already obtained velocity and elastic parameters model,

increasing both the scale and the resolution of the model.

Page 80: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

76

76

Page 81: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

77

VIBROMETRIC

APPENDIX 1 – SUMMARY OF THE VSP BOREHOLES USED FOR THE CURRENT STUDY

Table A3- 2. Summary of the VSP boreholes used for the current study.

No. BoreholeYear(s)

measuredSource type

P wave

tomo

S wave

tomo

Poisson's

ratio tomo

no. of

sources

no. of

detectors

min.

depth (m)

max.

depth (m)

Top borehole

elevation (m)

1 KR01 1990 Dynamite Yes No No 5 176 40 915 9.91

2 KR02 1990-1995 Dynamite Yes No No 7 200 30 1025 8.62

3 KR04 1990-1995 Dynamite Yes Yes Yes 7 172 40 895 9.15

4 KR08 1995-2005 Dynamite & Vibsist Yes Yes Yes 7 64 300 615 10.56

5 KR10 1996 Dynamite Yes Yes Yes 7 114 40 605 9.32

6 KR14 2002 Vibsist Yes Yes Yes 10 88 40 475 8.27

7 KR27 2005 Vibsist Yes No No 7 96 25 500 7.25

8 KR38 2005 Vibsist Yes Yes Yes 7 96 30 505 9.58

Total 8 8 5 5 57 1006

All the results are reported in a reduced coordinate system, as shown below:

X Y Z Northing Easting Elevation

1525540.5 6792362.4 9.91 = y - 6700000= x - 1500000 = z

92362.42 25540.54 9.91

Original coordinates Reduced coordinates

77

Page 82: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

78

VIB

RO

ME

TR

IC

78

Page 83: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

79

VIBROMETRIC

APPENDIX 2 – 3D VSP TOMOGRAMS

P wave tomograms

Figure A3- 10. KR01 – P wave tomograms.

79

Page 84: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

80

VIBROMETRIC

Figure A3- 11. KR02 - P wave tomograms.

80

Page 85: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

81

VIBROMETRIC

Figure A3- 12. KR04 - P wave tomograms.

81

Page 86: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

82

VIBROMETRIC

Figure A3- 13. KR08 - P wave tomograms.

82

Page 87: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

83

VIBROMETRIC

Figure A3- 14. KR10 - P wave tomograms.

83

Page 88: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

84

VIBROMETRIC

Figure A3- 15. KR14 - P wave tomograms.

84

Page 89: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

85

VIBROMETRIC

Figure A3- 16. KR27 - P wave tomograms.

85

Page 90: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

86

VIBROMETRIC

Figure A3- 17. KR38 - P wave tomograms.

86

Page 91: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

87

VIBROMETRIC

S wave tomograms

Figure A3- 18. KR04 – S wave tomograms.

87

Page 92: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

88

VIBROMETRIC

Figure A3- 19. KR10 – S wave tomograms.

88

Page 93: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

89

VIBROMETRIC

Figure A3- 20. KR14 – S wave tomograms.

89

Page 94: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

90

VIBROMETRIC

Figure A3- 21. KR38 – S wave tomograms.

90

Page 95: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

91

VIBROMETRIC

Tomograms obtained by simultaneous inversion of VSP data from several boreholes

Figure A3- 22. All site – P wave tomograms after processing step 9.

91

Page 96: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

92

VIBROMETRIC

Figure A3- 23. All site – P wave tomograms after processing step 9 – rotated 180 degrees.

92

Page 97: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

93

VIBROMETRIC

Figure A3- 24. All site – P wave tomograms after processing step 9 – top view.

93

Page 98: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

94

VIBROMETRIC

Figure A3- 25. All site – S wave tomograms after processing step 12.

94

Page 99: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

95

VIBROMETRIC

Figure A3- 26. All site – S wave tomograms after processing step 12 – rotated 90 degrees clockwise.

95

Page 100: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

96

VIBROMETRIC

Figure A3- 27. All site – S wave tomograms after processing step 12 – top view.

96

Page 101: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

97

VIBROMETRIC

Figure A3- 28. All site – Poisson’s ratio tomograms after processing step 13.

97

Page 102: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

98

VIBROMETRIC

Figure A3- 29. All site – Poisson’s ratio tomograms after processing step 13 – rotated 90 degrees clockwise.

98

Page 103: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

99

VIBROMETRIC

Figure A3- 30. All site – Young’s modulus tomograms after processing step 13.

99

Page 104: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

100

VIBROMETRIC

Figure A3- 31. All site – Young’s modulus tomograms after processing step 13 – rotated 90 degrees clockwise.

100

Page 105: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

101

VIBROMETRIC

Figure A3- 32. All site – Shear modulus tomograms after processing step 13.

101

Page 106: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

102

VIBROMETRIC

Figure A3- 33. All site – Shear modulus tomograms after processing step 13 – rotated 90 degrees clockwise.

102

Page 107: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

103

APPENDIX 4

Memo

REVIEW OF SEISMIC 3D VELOCITY MODEL BASED ON 3D-VSP DATA

ABSTRACT .............................................................................................................. 105

1 GENERAL ....................................................................................................... 107

2 ASSESSMENT OF INFORMATION INCLUDED IN SEISMIC DATA ............... 109 2.1 Sample comparisons .............................................................................. 110 2.2 Comparisons of samples to drillhole logging ........................................... 110 2.3 Comparison of velocity model to drillhole data ........................................ 111

3 REVIEW OF RESULTS ................................................................................... 115

4 DISCUSSION ON METHODOLOGY ............................................................... 125

5 CONCLUSIONS .............................................................................................. 127

REFERENCES ......................................................................................................... 129

Pöyry Environment Oy

P.O.Box 50 (Jaakonkatu 2) FI-01621 Vantaa Finland Domicile Helsinki, Finland Business ID. 0196118-8 Tel. +358 10 3311 Fax +358 10 33 26761 www.environment.poyry.fi Date Jun. 23, 2009 Ref. No Project 67070638.BGF

Eero Heikkinen Tel. +358 10 33 26751 [email protected] Page 103 (21)

Page 108: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

104

Page 109: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

105

ABSTRACT

A model of P and S wave velocities in Olkiluoto has been compiled on basis of 3D VSP

surveys during 1990 – 2005. Model uses arrival time and drillhole deviation data from 8

drillholes, OL-KR01, OL-KR02, OL-KR04, OL-KR08, OL-KR10, OL-KR14, OL-

KR27 and OL-KR38 located near ONKALO.

The model indicates that the velocities increase with depth. Model is highly averaging.

The velocities in the model are ranging at 5280 - 6120 m/s for P-wave (average 5500

m/s) and at 2620 - 3420 m/s (average 3100 m/s) for S-wave. Lowest velocities are

encountered at shallowest 0 - 100...200 m depth interval.

Apart of the depth trend there are observed also zones where velocity is varying from

the average, high velocities indicating rock type variation e.g. locations occupied by

granite pegmatite, TGG gneiss or competent mafic rocks, and lower (<5300 m/s or

<3000 m/s) velocities indicating locations of alteration or deformation in rock mass.

Most indications of these zones are associated to depth intervals 300-380 m, 500 m and

600-800 m.

The explanation for velocity variation in the drillholes is found from more detailed

petrophysical sample and drillhole logging data. Variations arise from lithology,

alteration and deformation. Fresh (non altered, non deformed) rock types have

characteristic velocities, granite pegmatite 5900 - 6100 m/s, some of the mafic gneisses

6000 - 6500 m/s, and diatexitic, veined, mica and tonalitic gneiss, partly also mafic

gneiss rather varying 5300 - 5900 m/s. The velocity of gneisses will depend on

leucosome content, and is decreasing according to increasing density, i.e. according to

increase of mafic silicates (biotite and amphibole), which is associated also with texture

of the rock mass.

High banding intensity is associated to lowered velocity. Velocities decreasing below

5100 - 5300 m/s start to indicate increase in degree of alteration or deformation due to

increased fracturing or porosity. Decrease of velocity below 4500 - 5100 m/s is

indicating presence of significant deformation or alteration. Tomography averages out

this low values, as minimum directly observable layer thickness is order of 20 – 50 m.

Drillhole logging data of P and S velocity can be used to delineate location and

thickness of anomalous velocity zones. Velocity model can be used to estimate

continuity, orientation and extent of these zones between and outside of drillholes. The

seismic reflection interpretations from VSP surveys and from 3D surface reflection

surveys provide tools to delineate the boundaries of these anomalous velocity zones.

Page 110: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

106

Page 111: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

107

1 GENERAL

Seismic P and S wave velocity data and reflection features can be used for rock mass

characterization.

The reflecting features are indicating fairly accurately location of elastic contrasts

(impedance, consisting of velocity and density). Reflection measurements do not

indicate directly the rock mass properties.

Transmission measurements and attribute interpretations can provide averaged rock

mass properties of P and S wave velocity and Poisson‟s ratio. These properties are

indirectly associated to lithology, fracture frequency, deformation and alteration of the

rock mass. The properties are linked in varying manner over different scales of

observation.

Vibrometric Oy has reprocessed the 3D VSP survey data from eight drillholes in

Olkiluoto in order to create tomographic scans. Purpose was to produce P and S wave

velocity data to be used for rock mechanical modelling of Olkiluoto.

This memorandum represents the comparison of results to available drillhole data.

Page 112: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

108

Page 113: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

109

2 ASSESSMENT OF INFORMATION INCLUDED IN SEISMIC DATA

The P and S wave velocities are affected by mineral content, texture, porosity and

fracturing, deformation and alteration of the rock mass as well as by stress field.

Seismic measurement data on velocity has been obtained in following scales and

geometries:

- Petrophysical samples on 50-100 mm sample scale, which can be

compared to other sample data, and to some extent to core logging and

drillhole measurement data (Öhman et al. 2009a)

- Sonic logging in all drillholes in Olkiluoto, a continuous measurement of

P and S wave velocities immediately in the drillhole wall over 1 m

interval (Öhman et al. 2009b).

- Seismic refraction survey in most of the Olkiluoto area at 50 m line

interval (bidirectional). The velocity is describing the surface part of

bedrock down to 20-30 m depth level (Lehtimäki, 2002).

- Seismic refraction survey in tunnel (Cosma et al. 2008)

- Crosshole tomographic surveys in KR14-KR15 and KR4-KR10 (Enescu

et al. 2003, 2004).

Velocity model or constant velocity has been used in processing of 3D VSP surveys

during 1990 – 2005 and in surface based 3D reflection surveys 2006 and 2007 (Juhlin &

Cosma 2007, Cosma et al. 2008). The processing velocity does not adequately

characterize the rock properties.

There would be limited possibilities to process more detailed velocity variation from the

3D seismic survey data, either with pre-stack amplitude vs. offset analysis and dip

moveout modeling (DMO) or using inversion modeling in the stacked reflection cube.

Purpose of the currently presented work is to create a velocity model from 3D VSP

travel times.

The available sample and drillhole logging data, as well as geological data, are from

different scale and from different acquisition geometry as the tomographic results. The

source of velocity variation cannot be explained by direct comparison, which would

easily lead to erroneous deductions. Results from OL-KR01 were used to explain the

velocity variation. Validation chain used below is the following:

1) Velocity sample measurement data (“petrophysics”, Öhman et al. 2009a) is

compared to lithology and alteration of the samples, and to other physical

properties (density, porosity, resistivity) from same location or same

samples.

2) Velocity sample measurement data is compared to drillhole sonic logging

data (Öhman et al. 2009b) from same locations

3) Drillhole sonic logging data is compared to lithology, fracture frequency,

deformation and alteration of the core logging (Paulamäki et al. 2006) from

same locations.

Page 114: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

110

4) Drillhole sonic logging data is compared to the tomographic velocity at the

drillhole location.

Comparisons have been carried out also between the scales to demonstrate possible

correlation.

2.1 Sample comparisons

From 1350 samples (Öhman et al. 2009a) where porosity and P-wave velocity has been

measured, a general trend of decreasing velocity with increasing porosity is seen.

Correlation coefficient is -0.42. Correlation is more distinct in gneisses (-0.43 - -0.51)

than in pegmatite granite (-0.18). Correlation does not exist or is reverse in TGG, MDB,

QGN and KPF.

P wave velocity in the samples correlates weakly with lithology (coefficient 0.24).

Review was performed by giving the rock types the median of velocity in the rock type

as numerical value.

Correlation of P wave velocity in the samples is very weak on deformation and

alteration (-0.09, velocity is slightly decreasing in altered samples). Review was made

by assigning the samples “1” when belonging to alteration or deformation interval, “2”

when belonging to both, and “0” when not belonging to either of these.

Correlation of P wave velocity and density is weakly negative (-0.11). Velocity is

decreasing with increasing density in DGN, KFP, PGR and SGN (correlation -0.11 -

.0.24, probably due to melanosome content), in MFGN and QGN velocity increases

with increasing density (correlation 0.33 – 0.69). In MGN, TGG and VGN there is only

slight or no correlation between density and velocity.

Correlation of P wave velocity and resistivity is weak, 0.15. Resistivity and P wave

velocity is increasing while porosity is decreasing, especially in DGN, MGN, MFGN

and VGN (0.21 – 0.27). Correlation is very weak in TGG (0.12), does not exist in PGR

and SGN and is reverse in MDB, KFP and QGN (resistivity is decreasing as velocity

increases).

2.2 Comparisons of samples to drillhole logging

The 95 petrophysical samples from OL-KR01 were compared to P wave velocity

logging data from the same drillhole location. There is a visible trend of increasing

velocity. Correlation is weak 0.15. There are clear differences in drillhole logging and

petrophysical samples, which may be explained with:

- averaging of logging data over interval

- positioning errors

- measurement errors on both data sets

- stress field effect on rock mass in in situ measurement.

Rejecting most distinct outliers from comparison will increase correlation to 0.46.

Page 115: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

111

Comparison of drillhole logging to geological core data

Geophysical P wave velocity logging has a rather distinct correlation to fracture

frequency (-0.51). Correlation to core loss and Ri zone indications is slightly weaker (-

0.37). Correlation to alteration zone occurrences is also -0.38. Highest correlation is

observed to kaolinite and illite alteration (-0.32 – -0.37), and lower to sulphide alteration

(-0.16). Most effect has the pervasive kaolinite alteration (-0.31). Correlation to

deformation zones is slightly lower, -0.26. Combined correlation of alteration and

deformation is -0.39.

The velocity decreases as fracture frequency increases, and in presence of fracture

zones, alteration or deformation.

Correlation to lithology is higher than for the samples, 0.34. The median of the

velocities for rock types is different for petrophysical samples and for logging, which is

caused by stress field and orientation effects on rock mass different in laboratory and in

situ, and due to fact that laboratory samples do not contain fractures whereas the

logging will encounter both rock mass and fracturing.

The presented correlations were reviewed in OL-KR01 only and can be different in

other drillholes and other parts of the area.

2.3 Comparison of velocity model to drillhole data

The velocity model data from different steps of tomographic inversion were compared

with lithology, alteration, deformation, fracture frequency and sonic logging data.

Values of tomogram were picked at the drillhole path.

The most obvious correlation should occur between the P wave velocity from sonic

logging and tomography. The single slice tomography (an example from source S1

above the drillhole) in OL-KR01 has correlation 0.42. The complete hole tomogram of

OL-KR01 compared to logging in OL-KR01 has correlation of 0.394. The complete

area tomogram in OL-KR01 receives correlation 0.388. Different tomographic step

results (slice, whole drillhole and whole area) correlate 0.68 – 0.73.

Comparison to petrophysical data from the same depth level leads to somewhat

confusing result. Velocities have a negative correlation of -0.27 – -0.31. This is

probably due to different scale and the directional and in situ effects of tomographic

survey. Petrophysical data is highly scattered when compared to imaging, and cannot be

directly compared to the image results.

Correlation of seismic measurements between different scales is not perfect, as the

scales are different. Also the tomographic image describes the rock mass from larger

interval and also from the volume outside of the drillhole. Features may also image to

slightly different position due to dip or other geometric reasons.

Page 116: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

112

On the tomographic imaging scale lithology on drillhole logging does not correlate to

seismic P wave velocity (-0.05 – 0.09). Shifting the tomographic result some 25 m

downwards the correlation gets sligthly better, 0.15 - 0.22.

Fracture frequency has correlation of – 0.12 - -0.17 (shifted downwards -0.16 - -0.21).

This indicates the P wave velocity is expected to decrease when fracture frequency

increases.

Fracture zone or core loss indications (which are rather narrow) have a correlation of -

0.1.

Deformation zone indications alone explains fairly little (-0.1), including the influence

zones may have some effect.

The alteration indications (typically from rather wide section) indicate correlation of c. -

0.22 to seismic P wave velocity from tomogram. According to this, the alteration has

some reducing effect on seismic velocities in this scale. Strongest correlation (-0.31) has

kaolinite alteration. Also sulphide alteration has a slight effect (-0.23).

Comparison of seismic P velocities from petrophysical samples, logging and different

tomographic processing steps in OL-KR01 is shown in Figure A4- 1.

Page 117: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

113

Figure A4- 1. Geological information from OL-KR01 (Paulamäki et al. 2006) and seismic P wave velocity data in different scales.

Page 118: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

114

Page 119: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

115

3 REVIEW OF RESULTS

Results were compared to drillhole lithological and acoustic logging results. The near

offset shot point has produced fairly similar results as the sonic logging does, but the

variation is displaying the largest zones of both high and low velocities (Figure A4- 2).

Figure A4- 2. Tomogram of near offset shot in OL-KR2. Velocity range 5100 – 5900 m/s (blue-red). Velocity high and low positions are fairly similar. TGG (yellow) and PGR (red) indicate higher velocities than DGN and VGN gneiss (light and dark blue).

Page 120: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

116

The single slice from one shot point to one drillhole is often describing continuity or

termination and apparent angles of low and high velocity zones met in the drillhole. Part

of this information may suffer from geometric or processing artifacts.

Velocity distribution consists typically of low velocity zone at the surface (Figure A4-

3).

Figure A4- 3. Low velocity at top part of rock, OL-KR10 shot point towards the east. View from S.

Page 121: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

117

Another low velocity zone is seen in all drillholes at 200-400 m depth level (HZ20).

Yet another low velocity zone is seen at c. 500 m depth level, see Figure A4- 4 below.

Figure A4- 4.Low velocity zones at 300 – 400 m and 500 m, shot point towards the east, view from S.

Page 122: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

118

One more low velocity zone is met at 600 – 800 m depth level (HZ21), Figure A4- 5.

Figure A4- 5. Low velocity zone at 500 – 600 m, OL-KR01 shot point towards the north, view from E.

Page 123: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

119

High velocity zones are met on several profiles in OL-KR1 and OL-KR2 at upper part

of rock mass in the north and west. The zone seems to dip towards south and east and to

terminate before reaching the drillhole OL-KR1. (Figure A4- 6).

Figure A4- 6. High velocity zone at 200 – 300 m, OL-KR01 shot point towards northeast, view from E.

Page 124: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

120

Possibly the higher velocity is explained by TGG gneiss in OL-KR2 and OL-KR14, and

also in OL-KR20 (not included into survey). Rock type TGG is typically sparsely

fractured. Another zone of higher velocity is observed in mid part of OL-KR10. Results

rendered from the whole drillhole results show the directions and continuity of the low

velocity zones, as compared with the drillhole observations (Figure A4- 7).

Figure A4- 7. E – W view of OL-KR01 tomogram with seismic P-wave velocity (blue) and density (green) along the drillhole. High and low velocities in logging and in tomograms are matching. Velocity model shows discontinuous features

Page 125: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

121

The imaging results applying all data simultaneously smooth out the minor details.

Results indicate that the low velocity zone in the surface extents deeper (-300 m level)

in the north and also in the east (OL-KR27). The HZ20 and HZ21 zones are seen also in

this result as extensive low velocity areas (Figure A4- 8).

Figure A4- 8. Section along Easting = 1525.880 (KKJ1). Velocity low is seen to reach deeper in the north (up to -300 m level). Low velocity zones are observed at 300 – 500 m and 600 – 700 m levels.

Page 126: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

122

Figure A4- 9. Section along Northing = 6792.200 (KKJ1). Velocity low are seen at the top down to 200 – 300 m, ta c 400 m and at 600 – 700 m.

Page 127: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

123

It is possible to compare the tomographic results also with other modeling or seismic

reflection data. Below is presented a reflection image from 3D survey data, overlain

with VSP tomographic velocity slice (Figure A4- 10).

Figure A4- 10.The seismic 3D reflection result and intersection of VSP tomographic result from OL-KR1. Red is high velocity and blue velocity. Green and black indicate high reflectivity.

Page 128: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

124

Page 129: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

125

4 DISCUSSION ON METHODOLOGY

Seismic P and S wave velocity model was constructed on base of 3D VSP surface to

drillhole measurements. The model is focused near to the ONKALO volume.

Model is sparse, consisting of sub-planar slices from single sources to drillhole. Slices

(5-7 for each drillhole) are arranged in a star-shaped pattern. Coverage deeper down in

rock mass is concentrated in a narrow area near drillholes. Results leave gaps between

different profiles and drillholes.

Modelling was performed in three steps:

1) tomographic inversion of one source and all receivers (slice)

2) inversion of all slices for one drillhole, harmonising the velocities for

drillhole.

3) inversion of all drillholes, regularizing the result over all directions.

The initial modeling step produces the highest resolution and best correlation to

drillhole mapping and logging data but is prone to processing artifacts. Rock properties

are projected to the plane of image and may represent apparent dips rather than real

orientations of anomalies. Result would be still useful in delineation of velocity

contrasts. Features which can be tentatively detected in this material are possible

velocity high and low areas, and potential discontinuities in these areas or even vertical

displacements.

Single drillhole model is maintaining most of the details of the individual slices, and

joining the velocity levels comparable. The result starts to smooth out minor details.

The whole site model is averaging the velocity strongly. Velocity levels are consistent

over the site volume. Result is indicating the most significant and continuous velocity

low and high areas.

There are several opportunities to continue and improve in this work, if found

necessary.

Coverage:

- now eight drillholes and 60 profiles near ONKALO were selected for

processing of P-waves and half of this (four drillholes/30 profiles) for S-

waves.

- adding up with remaining measured drillholes would enhance coverage;

these are OL-KR03, OL-KR05, OL-KR06, OL-KR07, OL-KR09, OL-

KR11, OL-KR12, OL-KR13, OL-KR19 and OL-KR29.

- adding up with moving source VSP (WVSP) in OL-KR04, OL-KR08, OL-

KR10 and OL-KR14.

- other drillholes in the near volume which have not been measured with VSP

can be used to supplement the image in ONKALO area (OL-KR15, OL-

KR20, OL-KR22, OL-KR23, OL-KR25, OL-KR28, OL-KR37, OL-KR39)

and the drillholes in the east-northeast part area can be used to extent the

image (OL-KR40 – OL-KR42, OL-KR44 – OL-KR50).

Page 130: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

126

Enhancing the measurement technique:

- surface to drillhole reflection technique (3D VSP) has not been designed for

tomographic imaging but was used with success.

- new measurements should take into account a precise timing and signal

consistency to avoid problems in processing

- several source stations would be useful to be placed on each azimuth, at

several offset distances. Overlapping traces would enhance imaging.

- the same source stations are useful to be used for several drillholes, to make

the traces spatially overlapping and to allow interferometry.

- several drillholes can be measured in a surface to drillhole profile, which

would bring in continuity in a 2D plane.

New geometries to be considered:

- apart of surface to drillhole and crosshole geometries, other possibilities for

tomography are available

- tunnel excavation blast travel times in microseismic network have a good

coverage and will construct over time

- profiles between tunnel sections and ground surface (to image vertical

structures)

- profiles between tunnel sections

- profiles between tunnel sections and drillholes.

Processing tecniques:

- using the traveltime difference from a same source to receivers in two

drillholes (at a time) together with path between receivers will produce

interferometric imaging, and will assist in filling gaps between drillholes

and in horizontal direction.

- imaging can be used as a starting model for numerical inversion, which

would produce sharper model boundaries. This kind of modeling would

require denser coverage and constrains taken from drillhole logging.

- application of attribute analysis to existing surface 3D reflection results, to

obtain velocity field (the accuracy of stacking velocities is inadequate for

this purpose).

Page 131: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

127

5 CONCLUSIONS

A model of P and S wave velocities was compiled using 3D VSP surface to drillhole

reflection data. Measurements have been designed for reflection, and techniques applied

vary from dynamite source to engineered impact series source. This has caused

problems for travel time picking and processing which need to be overcome.

The model is representing well the velocity distribution in Olkiluoto area. Velocity

ranges are conforming with the previously known levels from petrophysics, sonic

logging and crosshole tomographic results. Velocity field is smooth. Imaging using all

drillhole data simultaneously is averaging the model, but showing correct velocity levels

and indicating the most significant low velocity volumes in the area. These are

potentially linked to alteration and deformation features.

The single hole images and images created between single source station and one

drillhole are representing in more detailed manner the local velocity variation, and may

display the low and high velocity zones in more detail. The results are not on exact

velocity level regionally, and may contain processing and geometric artefacts.

The results compare well with drillhole sonic logging and large scale geological features

(alteration and to some degree lithology). Results do not compare well with sample data

or detailed geological information from core. However as the sonic logging does

correlate with detailed information, this provides a link with which the observations can

be explained.

Some part of low and high velocity zones are explained with lithological or textural

variation. Significant lowering of velocity is explained with deformation and alteration

zones, mainly with increase of fracturing and porosity. Features which are seen as

velocity anomalies in tomography need to be large and continuous.

The velocities range at 5100 – 6100 m/s (P) and 2900 – 3300 m/s (S). Lowering of

velocities below 5300 m/s (P) and 3000 m/s (S) is indicating some degree of

deformation or alteration.

The near surface fracturing intensity and effect of increasing stress are seen as

increasing velocity according to depth. The S wave velocity is more sensitive to this

effect. Locally the low velocity zones can be distinguished from this trend. Most

significant low velocity areas extent from surface to 100…300 m depth (deepest in the

north part of area), and between 300-400 m (HZ20), at 500 m and between 600-800 m

(HZ21). One velocity low area is met at bottom of drillhole OL-KR08.

The technique has proven the potential to characterize the rock mass with seismic P and

S wave velocities. The velocities are depending on lithology, texture, porosity,

fracturing, deformation, alteration and stress field. When using the results it is necessary

to explain the source of observed anomalies with drillhole data. Results are showing

lateral variation of velocity outside of drillhole covered areas. There is a possibility to

use the results in delimiting anomalous volumes (best together with reflection data).

Page 132: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

128

Tomography will allow assigning the material properties to rock mass. Priority for the

material property presentation is the P and S wave velocities and Poisson ratio which

are directly obtained from data. There are some possibilities to use the results to

characterize alteration, fracturing and lithology, but correlations from core mapping

scale are rather weak.

It would be useful to apply the tomographic images together with reflection data where

the boundaries are seen more accurately.

The tomographic model may be supplemented with existing data if found necessary. It

is also possible to design possible future 3D VSP measurements in a way the processing

would be more readily obtainable.

Page 133: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

129

REFERENCES

Cosma, C., Cozma, M. & Enescu. N. 2008. P- and S-wave 3D velocity inversion of

VSP data collected at Olkiluoto. A memorandum.

Cosma, C., Heikkinen, P., Honkanen, S. & Keskinen, J. 1990. VSP-survey in Olkiluoto,

boreholes KR1 - KR3. TVO Site Investigations, Work Report 90-39.

Cosma, C., Heikkinen, P., Honkanen, S. & Keskinen, J. 1991. VSP-survey in Olkiluoto,

boreholes KR4 - KR5. TVO Site Investigations, Work Report 90-36.

Cosma, C., Heikkinen, P., Honkanen, S., Keskinen, J. 1996. VSP-survey at Olkiluoto in

Eurajoki, borehole OL-KR8 and extended parts of boreholes OL-KR2 and OL-KR4.

Posiva Work Report PATU-96-11e.

Cosma, C. Enescu, N., Heikkinen, P. & Keskinen, J. 1996. VSP-survey at Olkiluoto in

Eurajoki, boreholes OL-KR9 and OL-KR10. Posiva Work Report PATU-96-60e.

Cosma, C., Enescu, N., Adam, E & Balu, L. 2003. Vertical and horizontal seismic

profiling investigations at Olkiluoto, 2001. Report POSIVA-2001-01. ISBN-951-652-

065-0.

Cosma, C., Cozma, M., Juhlin, C. & Enescu, N. 2008. 3D Seismic Investigations at

Olkiluoto, 2007. Factual Report. Posiva Working Report 2008-43, 49 p.

Enescu, N., Cosma, C., Balu, L. 2003. Seismic VSP and crosshole investigations in

Olkiluoto, 2002. Posiva, Working Report 2003-13.

Enescu, N., Cosma, C., Balu, L. 2005. Seismic VSP Investigations at Olkiluoto, 2005.

Posiva Working Report 2007-72, 147 p.

Cosma, C., Cozma, M., Balu, L. & Enescu, N. 2008. Rock mass seismic imaging

around the ONKALO tunnel, Olkiluoto 2007. Posiva Working Report 2008-64, 29 p.

Enescu, N., Cosma, C., Balu, L. 2004. Reflection seismics using boreholes at Olkiluoto

in 2003 - from investigation design to result validation. Volume 1. Posiva Working

Report 2004-62.

Heikkinen, E., Vaittinen, T., Saksa, P. , Palmén, J., Nummela, J. 2004. Reflection

seismics using boreholes at Olkiluoto in 2003 - from investigation design to result

validation. Volume 2. Posiva Working Report 2004-62.

Heikkinen, E. (ed), Paananen, M., Kurimo, M., Öhberg, A., Ahokas, H., Okko, O.,

Front, K., Hassinen, P., Pitkänen, P., Cosma, C., Heikkinen, P., Keskinen, J., Honkanen,

S., & Korhonen, R. 1992. Geophysical investigations in the Olkiluoto area, Finland.

Summary report. Nuclear Waste Commission of Finnish Power Companies, Report

YJT-92-34.

Page 134: ONKALO Rock Mechanics Model (RMM) · 2009-12-14 · The Geological Strenght Index (GSI), introduced by Hoek (1994) and Hoek, Kaiser and Bawden (1995) provides a number which, when

130

Juhlin, C. & Cosma, C. 2007. A 3D Surface Seismic Pilot Study at Olkiluoto, Finland:

Acquisition and Processing Report. Posiva Working Report 2007-65, 47 p.

Keskinen, J., Cosma, C. & Heikkinen, P. 1992. Seismic VSP and HSP surveys on

preliminary investigation areas in Finland for final disposal of spent nuclear fuel.

Nuclear Waste Commission of Finnish Power Companies, Report YJT-92-19.

Lehtimäki, T. 2003. Combined interpretation and processing of seismic refraction data

at Olkiluoto. Posiva Working Report 2003-62. 41 p.

Paulamäki, S., Paananen, M., Gehör, S., Kärki, A., Front, K., Aaltonen, I., Ahokas, T.,

Kemppainen, K., Mattila, J., Wikström, L. 2006. Geological Model of the Olkiluoto

Site, Version 0. Posiva Working Report 2006-37, 355 p.

Öhman, I., Heikkinen, E., Säävuori, H., Vuoriainen, S., Paulamäki, S., Aaltonen, I.

2008. Summary of petrophysical analysis of Olkiluoto core samples 1990 - 2008.

Working report 2009-11. Posiva Oy, Eurajoki. 212 p.

Öhman, I., Palmén, J. and Heikkinen, E. 2008. Unification of acoustic drillhole logging

data. Working report 2008-18. Posiva Oy, Eurajoki. 57 p.