UR14_2009 Final version_nso - School of the Built Environment

Transcription

INFLUENCE
OF THE DESICCATION FINE
FISSURING ON THE STABILITY OF FLOOD
EMBANKMENTS
FINAL REPORT
INCLUDING LABORATORY STUDIES
Philippe Sentenac
Marcin Zielinski
Mark Dyer
Strathclyde University
Strathclyde University
Trinity College Dublin
March 2009
FRMRC Research Report
Project Web: www.floodrisk.org.uk
Desiccation fine fissuring
FMRC Research Report II UR14
FRMRC Partners
The FRMRC Partners are:
• University of Bristol
• Heriot Watt University
• HR Wallingford
• Imperial College, London
• University of Lancaster
• University of Manchester
• University of Nottingham
• University of Sheffield
Project Secretariat
ARP
Directorate of Planning and Academic Services
University of Manchester
Sackville Street,
Manchester
PO Box 88
M60 1QD
Tel: +44 (0)161 306 3626
Fax: +44 (0)161 306 3627
Web: www.floodrisk.org.uk
UR14_desiccation_fine_fissuring_WP4_1_v4_0 Final
01/04/2009
2
SUMMARY
The aim of this project was to investigate the effect of desiccation and fissuring of clay fill
on the geotechnical stability of flood defence embankments by: a) monitoring and measuring the extent of fine fissuring in the field using visual inspection, analysis of soil properties such as plasticity/shrinkage and soil suction, b) measuring the reduction in mass permeability caused by fissuring and further investigate the flow of water through fissured
soil in the laboratory that could lead to slope instability and breach initiation, c) using current models and theories to predict the development of fine fissuring in full-scale embankments, d) assimilating the results into a useable limit equilibrium model to calculate
geotechnical stability of flood embankments under different flood loading conditions for
later use in EA/DEFRA PAMS programme.
Although the type of desiccation cracking may be different from that observed in the field
(width, spacing, type and strength of the soil) the results from this study clearly demonstrate that breaching of a flood embankment constructed from clay soil is controlled by
weakness in the embankment construction such as highly permeable underlying strata
which was not the case in the model constructed and which partly explain that the scaled
embankment didn’t collapse. Furthermore breaching also is controlled by deterioration
(such as desiccation) and not the classical Bishop style slope failure. In fact the classical
slope instability is rarely recorded for clay flood embankments except during the construction phase when the underlying soil may be too weak to initially support the embankment
(such as occurred for the new embankment at Thorngumbald). This is a very important result that shows that fault trees and corresponding fragility curves need to focus on potential weakness that could lead to breach initiation and avoid the misconception that classical
slope failure is a primary failure mode during a flood event (even though a large amount of
research effort and resources have been directed towards the construction of fragility
curves to characterise classical slope failure mechanism for breach initiation).
The miniature geophysical technique proved to be useful tool for detecting vertical cracks
down to a depth of 50 cm.
01/04/2009
3
DOCUMENT DETAILS
Document History
Version
1_0
Date
Lead Authors
Institution
Joint
authors
Comments
Dec 2006
University of
Strathclyde
First interim
report version 1
2_0
April 2007
University of
Strathclyde
3_0
Dec 2008
University of
Strathclyde
First interim
report final
version
Final report
version 1
4_0
April 2009
Mark Dyer
Stefano Utili
Marcin Zielinski
Mark Dyer
Stefano Utili
Marcin Zielinski
Marcin Zielinski
Philippe Sentenac
Mark Dyer
Marcin Zielinski
Philippe Sentenac
Mark Dyer
University of
Strathclyde
Final report
final version
Changes in Staffing
•
•
Mark Dyer was PI from September 2004 to January 2008
Philippe Sentenac was PI from January 2008 to June 2008
Industrial Partners
•
•
Mark Morris, HR Wallingford
Phillip Smith, Royal Haskoning
Acknowledgement
This research was performed as part of a multi-disciplinary programme undertaken by the Flood Risk
Management Research Consortium. The Consortium is funded by the UK Engineering and Physical
Sciences Research Council under grant GR/S76304/01, with co-funding from:
• Defra and the Environment Agency through their Joint R&D programme on Flood and Coastal
Erosion Risk Management,
• UKWIR
• NERC
• The Scottish Executive
• Rivers Agency Northern Ireland
Disclaimer
This document reflects only the authors’ views and not those of the FRMRC Funders. This work may
rely on data from sources external to the FRMRC Partners. The FRMRC Partners do not accept
liability for loss or damage suffered by any third party as a result of errors or inaccuracies in such data.
The information in this document is provided “as is” and no guarantee or warranty is given that the
information is fit for any particular purpose. The user thereof uses the information at its sole risk and
neither the FRMRC Funders nor any FRMRC Partners is liable for any use that may be made of the
information.
© Copyright 2009
The content of this report remains the copyright of the FRMRC Partners, unless specifically
acknowledged in the text below or as ceded to the funders under the FRMRC contract by the Partners.
01/04/2009
4
TABLE OF CONTENTS
SUMMARY.................................................................................................................................................... 3
DOCUMENT DETAILS ................................................................................................................................... 4
TABLE OF CONTENTS.................................................................................................................................... 5
TABLE OF FIGURES ....................................................................................................................................... 6
TABLE OF TABLES ........................................................................................................................................ 8
1
INTRODUCTION.............................................................................................................................. 9
2
OBJECTIVES .................................................................................................................................. 10
3
EXPERIMENTAL SET-UP ............................................................................................................. 11
3.1
Construction of macro-scale embankment model ................................................................... 11
3.1.1 Introduction........................................................................................................................ 11
3.1.2 in-field Soil Rating ............................................................................................................. 12
3.1.3 Laboratory investigation of soil properties......................................................................... 13
3.1.4 Construction of the embankment model............................................................................. 14
3.1.5 Soil compaction properties during construction. ................................................................ 16
3.2
Environmental Chamber ......................................................................................................... 18
4
INVESTIGATING USE OF GEOPHYSICS IN CRACKING DETECTION .......................................................... 19
4.1
Pilot test .................................................................................................................................. 19
4.1.1 Introduction........................................................................................................................ 19
4.1.2 Experimental set up............................................................................................................ 20
4.1.3 Experimental procedure ..................................................................................................... 22
4.1.4 results and discussion ......................................................................................................... 23
5
MEASURING EQUIPMENT ................................................................................................................... 26
5.1
Introduction............................................................................................................................. 26
5.2
Moisture content probes – tdr’s .............................................................................................. 26
5.3
Soil water tension sensor – SWTS 1 ....................................................................................... 27
5.4
Soil water tension sensor – T4e .............................................................................................. 29
5.5
Relative humidity & air temperature sensors – Rht2nl........................................................... 31
5.6
Miniature resistivity array ....................................................................................................... 32
6
DESICCATION EXPERIMENT .......................................................................................................... 32
6.1
Phase one – desiccation with one weekly precipitation........................................................... 32
6.2
Phase two – desiccation with two weekly precipitation .......................................................... 38
6.3
Phase three – desiccation without precipitation ...................................................................... 40
7
FLOODING EXPERIMENT ............................................................................................................... 44
7.1
Introduction............................................................................................................................. 44
7.2
Experimental set up................................................................................................................. 44
7.3
Experiment.............................................................................................................................. 45
8
DISCUSSION ABOUT USER FOCUS OUTCOMES ..................................................................................... 57
8.1
Stated objectives and outcomes .............................................................................................. 57
8.1.1 Failure mechanism for breach initiation ............................................................................. 57
8.1.2 SOIL PROPERTIES .......................................................................................................... 60
8.1.3 GeoFlood model................................................................................................................. 61
8.1.4 Monitoring and measuring the extend of fine fissuring in the field.................................... 61
8.1.5 Monitoring and measuring the depth of cracking using suction and moisture content probes
61
8.1.6 database of fissuring of embankments................................................................................ 64
8.1.7 Characteristics of desiccation fissuring .............................................................................. 66
8.1.8 Use of geophysics in cracking detection ............................................................................ 67
9
POSSIBLE NOVEL SOLUTIONS ............................................................................................................. 67
10 FURTHER STUDIES ............................................................................................................................. 69
10.1
LABORATORY STUDIES.................................................................................................... 69
10.2
Field STUDIES....................................................................................................................... 70
01/04/2009
5
11
CONCLUSION ..................................................................................................................................... 71
11.1
Measuring THE EXTENT of desiccation cracking ................................................................ 71
11.2
Geophysical measurements of the extent of fine fissuring ...................................................... 71
11.3
Failure mechanism of the embankment model ........................................................................ 71
12 REFERENCES ..................................................................................................................................... 73
13 ANNEX .............................................................................................................................................. 75
13.1
First day of desiccation phase ................................................................................................. 75
13.2
second day of desiccation phase ............................................................................................. 76
13.3
Third day of desiccation phase................................................................................................ 77
13.4
Fourth day of desiccation phase.............................................................................................. 78
13.5
Fifth day of desiccation phase................................................................................................. 79
13.6
sixth day of desiccation phase................................................................................................. 81
13.7
seventh day of desiccation phase ............................................................................................ 82
13.8
Resistivity scans...................................................................................................................... 83
TABLE OF FIGURES
Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
Figure 3.7
Figure 3.8
Figure 3.9
Figure 3.10
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
Figure 4.9
Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.5
Figure 5.6
Figure 5.7
Figure 5.8
Figure 6.1
Figure 6.2
Figure 6.3
Figure 6.4
Figure 6.5
Figure 6.6
Figure 6.7
Desiccation cracking of the clay used in the experiment as seen under natural state. ....... 12
Proctor compaction curve. ................................................................................................ 13
Concrete flume with diffuser and windows where the model was constructed.................. 14
Soil prepared for compaction. ........................................................................................... 14
85 kg vibrating plate used for compaction. ....................................................................... 15
Scraped surface prepared for the next layer placing.......................................................... 16
U100 sample taken after compaction................................................................................... 16
Dry density measured for compacted layers...................................................................... 17
Undrained shear strength for compacted layers................................................................. 18
Environmental chamber with embankment and installed sensors................................. 19
Non-corrosive electrodes mounted in block terminals. ..................................................... 21
Experimental setup with infra-red heater and electrodes installed. ................................... 21
Four levels array of plotting points chosen for data analysis............................................. 22
Temperature variations for 12 days drying process........................................................... 22
Inversed Resistivity map taken before drying process and after 4 hours of heating.......... 23
Comparison between two measurements taken after 20 hours and 27 hours of drying. .... 24
Comparison between two measurements taken after 43 hours and 50 hours of drying. .... 24
Moisture content profile measured during excavation. ..................................................... 25
Inverted resistivity map with corresponding picture of the cracked surface...................... 25
Example of TDR probe used in the experiment. ............................................................... 26
Specific calibration curve for used soil. ............................................................................ 27
Single calibration curve for tensiometer - SWTS1............................................................ 28
Downwards installation with the angle.............................................................................. 28
Example of SWTS 1 used in experiment........................................................................... 29
Example of T4e used in the experiment. ........................................................................... 30
Single calibration curve for tensiometer – T4e. ................................................................ 31
Example of RHT2nl sensor in the solar radiation shield. .................................................. 31
First cracks observed on the embankment......................................................................... 33
Cracks observed on the upstream side of the embankment. .............................................. 33
Cracks observed on the crest after 7 days of drying. ......................................................... 34
Moisture content profiles representative for first 6 weeks of experiment.......................... 34
Suction profiles representative for first 6 weeks of experiment. ....................................... 35
Inversed resistivity scan taken before the experiment. ...................................................... 36
Resistivity scan with corresponding visual observation after 7 days of drying. ................ 37
01/04/2009
6
Figure 6.8
Inversed resistivity scan taken before the experiment. ...................................................... 38
Figure 6.9
Moisture content profiles representative for 4 weeks of experiment................................. 38
Figure 6.10
Suction profiles representative for 4 weeks of experiment. .......................................... 39
Figure 6.11
2 inversed resistivity maps for scans taken before and after the rainfall....................... 40
Figure 6.12
Moisture content profiles representative for 5 weeks of drying at 20oC....................... 40
Figure 6.13
Suction profiles representative for 5 weeks of drying phase. ....................................... 41
Figure 6.14
Inversed resistivity map taken before flooding when the soil was dry.......................... 42
Figure 6.15
Observed desiccation cracking inside embankment body............................................. 42
Figure 6.16
Observed desiccation cracking inside embankment body............................................. 43
Figure 6.17
Crack parallel to the slope. ........................................................................................... 43
Figure 7.1
Embankment model with the experimental set up. ............................................................ 44
Figure 7.2
Inversed resistivity map taken before flooding.................................................................. 45
Figure 7.3
Channel filled with water to the level of first tensiometer................................................. 45
Figure 7.4
Downstream side of the embankment after one hour of flooding...................................... 46
Figure 7.5
Upstream side of the embankment with the water level 20 cm below the crest................. 46
Figure 7.6
First observed leaks on the downstream side. ................................................................... 47
Figure 7.7
Local seepage.................................................................................................................... 47
Figure 7.8
Local seepage from the bottom of the slope...................................................................... 48
Figure 7.9
Two inversed resistivity maps taken during first stage of flooding. .................................. 48
Figure 7.10
Water penetrating cracks on the crest........................................................................... 49
Figure 7.11
Overflowed crest. ......................................................................................................... 49
Figure 7.12
Eroded crack inside embankment close to the edge of the slope. ................................. 50
Figure 7.13
Observed phreatic zone. ............................................................................................... 50
Figure 7.14
Eroded slope and the half crest..................................................................................... 51
Figure 7.15
5 cm deep cracks on the crest after first day of flooding. ............................................. 51
Figure 7.16
Eroded slope with the two flow paths visible in the middle of the slope...................... 52
Figure 7.17
Increased water head with the wooden barrier. ............................................................ 52
Figure 7.18
Undercutting of clay block. .......................................................................................... 53
Figure 7.19
Visible crack in collapsing block.................................................................................. 53
Figure 7.20
Collapsing clay block. .................................................................................................. 54
Figure 7.21
Clay block pushed away. .............................................................................................. 54
Figure 7.22
Visible small collapse in the clay block........................................................................ 54
Figure 7.23
Soil moisture content profiles during flooding experiment. ......................................... 55
Figure 7.24
Suction/positive pore water pressure profiles during flooding experiment................... 56
Figure 8.1 Failure caused by seepage pressures in underlying coarse layers (after Cooling et al5). .... 57
Figure 8.2 Schematic process of uplift induced slope failure (after Van et al26). ................................ 58
Figure 8.3
Measurements of pore water pressure during flooding at 20 cm depth. ............................ 59
Figure 8.4
Measurements of pore water pressure during flooding at 40 cm depth. ............................ 59
Figure 8.5
Exposed undamaged clay core (Easington, E. Yorkshire), (after Marsland11) .................. 60
Figure 8.6
Soil suction variations in the upper layer of the embankment model. ............................... 61
Figure 8.7
Soil suction variations at 40 cm depth of the embankment model. ................................... 62
Figure 8.8
Soil suction profiles before and after the rainfall indicating depth of cracking................. 62
Figure 8.9
Soil suction in desiccated clay under seasonal variations (after Jennings12 1961). ........... 63
Figure 8.10
Moisture content variation in the upper layer of the embankment model..................... 63
Figure 8.11
Moisture content variation at 40 cm depth of the embankment model. ........................ 64
Figure 8.12
Investigated places of fissured embankments between 2006 and 2007. ....................... 65
Figure 8.13
Superficial cracking found on embankments in 2007................................................... 65
Figure 8.14
Pebbles mixed with top soil with some visible loose material...................................... 66
Figure 13.1
Downstream side of the embankment........................................................................... 75
Figure 13.2
Crest of the embankment. ............................................................................................. 75
Figure 13.3
Upstream side of the embankment................................................................................ 76
Figure 13.4
Figure 13.5
01/04/2009
7
Figure 13.6
Figure 13.7
Figure 13.8
Figure 13.9
Figure 13.10
Figure 13.11
Figure 13.12
Figure 13.13
Figure 13.14
Figure 13.15
Figure 13.16
Figure 13.17
Figure 13.18
Figure 13.19
Figure 13.20
Figure 13.21
Figure 13.22
Figure 13.23
Figure 13.24
Figure 13.25
Figure 13.26
Figure 13.27
Figure 13.28
Figure 13.29
Figure 13.30
Inversed resistivity map taken before experiment . ...................................................... 83
Inversed resistivity map from the second day............................................................... 83
Inversed resistivity map from the third day. ................................................................. 84
Inversed resistivity map from the fourth day. ............................................................... 84
Inversed resistivity map from the fifth day. .................................................................. 84
Inversed resistivity map from the sixth day. ................................................................. 84
Inversed resistivity map from the seventh day.............................................................. 84
TABLE OF TABLES
Table 1 Measured undrained shear strength...................................................................................12
Table 2 Soil properties...................................................................................................................13
01/04/2009
8
1 INTRODUCTION
The results from field observations (UFMO UR119) showed that desiccation fissuring occurs to a depth of typically 60 cm within the outer surface of a flood embankment constructed from clay fill and can occur within 2 years of construction. In some other cases it
has been seem that cracking can penetrate embankments to a depth of 1m. The critical
condition occurs when desiccation creates an interconnected network of sub-vertical and
sub-horizontal fissures that increases the mass permeability of the fill material to that of
coarse sand or gravel and hence allows rapid seepage of flood water through the surface
layer of the embankment (crest and side slopes).
Instead of leading to an increase of pore water with a corresponding reduction of increase
of effective stress as proposed by Marsland and Cooling15, it is understood that the high
rates of seepage can potentially lead to localised erosion of the upper rubbilised zone and
internal erosion of the embankment.
The laboratory work was based on the investigation of the soil properties used for the construction of the embankment model, measurements of soil moisture and suction profiles,
geophysical survey as well as visual observation of the onset of desiccation cracking.
This research report also examined the resilience of the embankment made of stiff clay
(“boulder clay”), subjected to the seepage and overflow conditions.
The present study is a separate experimental research to investigate the behaviour and performance of laboratory embankment. Hence, it may or may not represent field behaviour.
01/04/2009
9
2 OBJECTIVES
The first part of this project along with the detailed theory survey has been described in the
first Interim Report in April 2007.
The second part of the project, which is mainly based on the laboratory studies, will examine the effects of fine fissuring of clay fills on the geotechnical stability and breach initiation of flood embankments on the macro-scaled embankment model constructed in the
laboratory and subjected to different weather conditions.
In particular, this report will show how the desiccation and fissuring of clay embankments
affects the long-term performance by:
a) monitoring of the extent of fine fissuring on the examined embankment model using
visual observation,
b) measuring changes in moisture content profile during: dry, wet and extremely dry conditions,
c) measuring changes in suction profile during: dry, wet and extremely dry conditions,
d) comparing results obtained from the suction measurements on the crest with suction on
the slope,
e) investigating of non-invasive geophysical method for cracking initiation and detection,
f) monitoring the behaviour of fissured embankment model under flooding conditions using visual observation and geophysics,
g) monitoring failure of embankment subjected to overtopping conditions using visual observation,
h) investigating the use of soil instrumentation to detect extent of desiccation fissuring.
01/04/2009
10
3 EXPERIMENTAL SET-UP
3.1
3.1.1
CONSTRUCTION OF MACRO-SCALE EMBANKMENT MODEL
INTRODUCTION
The past research on the desiccation cracking of flood embankments which has been presented in the first version of the report in 2007 has shown that under natural conditions
embankments constructed from local alluvium or clayey soils are always tend to crack
when dry period occurs for a long time. It has been found that during drying process interconnected network of cracking can occur on the upper layer of flood defence and moreover it can progressively affect defence construction to a depth of 1m.
A lot of laboratory research has been done about the cracking of the soils, as well as the
behaviour of the embankments has been investigated by many researchers. Most of the
field based research focussed on measuring the moisture content and pore water pressures
during different seasons. Ridley et al.19, have investigated pore water pressures changes in
the highway embankments using different measuring devices and trying to find the best
available method for direct measurements. Smethurst et al22. were looking at seasonal
changes in pore water pressures and changes in the moisture content at different depths in
a grass-covered cut slope in London Clay. Experiments, performed by Take and Bolton23
have investigated the effect of seasonal moisture changes on clay slopes using microscaled centrifuge model.
One of the major questions asked before this experiment, was: How to replicate the cracking in the laboratory and how to obtain the same extent of cracking as observed by Cooling
and Marsland5?
Cracking in the soils occurs when is subjected to drying, but how the construction behaves
when is cracked and when extreme conditions occur or when passing from the dry period
to wet period and than from the wet period to very dry period? This report will try to answer these unknowns as well as how the cracked structure behaves when subjected to
flooding and overtopping conditions.
Weather conditions are very complex and are compound of many different phenomenons.
When trying to replicate them, use of as close as possible facilities must be used to achieve
results similar to the natural ones. If possible, extreme and minimum conditions should be
applied to the experimental set-up to obtain the result from the optimum range of measurements. It can be also useful to visually observe the construction when the experimental
extremum is considered. This extremum can be called “experimental stroke”, and it can
give more relevant results.
01/04/2009
11
3.1.2
IN-FIELD SOIL RATING
To prove that, assumed failure mode as an effect of uplift mechanism of cracked upper
layer, a macro-scale embankment model has been constructed in a concrete flume, covered
by environmental chamber.
The biggest priority in this experiment was to find natural source of clay which could be
use for the construction and for further experimental work. It has been also considered to
use the same material which is used for the real embankments and earth structures in order
to further possible field investigation after successful completion of the laboratory work.
Thus, construction of the new embankment in Galston, on the south-west from Glasgow,
which is a part of flood prevention scheme in Ayrshire, has been used as an indicator. First
inspection of the quarry which was the clay source has indicated very high tendency of
desiccation cracking, as it is shown in Figure 3.1.
Figure 3.1
Desiccation cracking of the clay used in the experiment as seen under natural
state.
In-field measurements using hand vane tester have been taken in four different places of
the quarry in order to classify consistency of the soil and to measure undrained shear
strength. The results which are presented below in the Table 1 were analysed in accordance to BS 8004:19863 and the clay has been identified as stiff clay.
Ref no
1
2
3
4
Table 1 Measured undrained shear strength.
Undrained Shear
Consistency in accordance
Strength
Field indication
with BS59304
[kN/m2]
121
Stiff
Cannot be moulded in the fingers
133
Stiff
“
152
Stiff
“
146
Stiff
“
01/04/2009
12
3.1.3
LABORATORY INVESTIGATION OF SOIL PROPERTIES
After the field reconnaissance and visual inspection, disturbed soil samples were taken to
carry out series of laboratory experiments investigate soil properties. The results from the
experiments are presented in the Table 2 below.
Table 2 Soil properties.
Property
Natural Moisture Content MC: %
Optimum Moisture Content OMC: %
Linear shrinkage LS: %
Liquid Limit LL: %
Plastic Limit PL: %
Plasticity Index ID: %
Maximum Dry Density ρd: Mg/m3
Particle Dry Density ρs: g/ml3
* Average from 3 independent tests
Value
17.91*
11.5
10.13*
35.75*
16.40*
36.2*
1.948
2.58*
According to the Specification for Highway Works – Earthworks, Series 600011, it was
identified that the end of the compacted product might be at 95% of maximum dry density,
which means that the moisture content of the compacted soil might be within the range of
6.2-17%, as it is shown in the Figure 3.2.
Standard Proctor Compaction Curve
2.000
Max DD
1.950
Dry Density (Mg/m3)
Max DD - Maximum Dry Density
OMC - Optimum Moisture Content
1.900
95% of Max DD
1.850
1.800
1.750
OMC - 11.6%
Min - 6.2%
Max - 17.0%
1.700
5.0
10.0
15.0
20.0
25.0
30.0
Actual Moisture Content (%)
Figure 3.2
Proctor compaction curve.
01/04/2009
13
3.1.4
CONSTRUCTION OF THE EMBANKMENT MODEL
Embankment model has been constructed in the concrete flume where the big amount of
water can be supplied in order to simulate the flooding conditions. The concrete 1m high
and 20 m long channel consists of two side windows which were used to observe the nature of cracking inside construction and the diffuser on the one side to supply the water
from the tank, as it is shown in Figure 3.3.
Figure 3.3
Concrete flume with diffuser and windows where the model was constructed.
Soil was placed into 12 cm high shattering and than broken down to the smaller lumps
(see Figure 3.4), to achieve very solid and homogenous layers and especially to avoid the
air packets formation during the compaction of cohesive soils. Under normal conditions,
when the embankments are constructed, the grid rollers are used to break down the big
lumps of the fill and to achieve the homogenous product.
Figure 3.4
Soil prepared for compaction.
01/04/2009
14
As it was mentioned before, the moisture content plays a major role in the compaction
process. Thus the soil samples where taken from the prepared soil to measure the moisture
content before the compaction of every layer. When the moisture content was bigger than
the one from the range calculated from the Standard Proctor Method, than the soil was left
over the night to let the water evaporate. After this, the moisture content has been measured again to check if it is in the expected range.
Because of the space and the scale of the embankment 85 kg heavy vibrating plate has
been chosen as the compaction effort in accordance to the Specification for Highway
Works – Earthworks, Series 600011, which is shown in Figure 3.5.
Figure 3.5
85 kg vibrating plate used for compaction.
Embankment was constructed in 10 layers. The compaction has been split to two parts.
First five layers were compacted in 10cm high layers and last 5 layers were divided by
two. In order to investigate the effect of the thickness of the layers in terms of cracking,
one half of the embankment was compacted in 10 cm as previous ones, and the other half
in 5 cm thick layers, to finally get 10cm height in total on both sides. To make the best
contact between two layers and to avoid the weak surface between them which could act
as the initiation for the horizontal cracking, each layer was scraped using rake, as it is
shown in Figure 3.6. After every day of compaction, embankment was covered by the
plastic sheet in order to protect it against the moisture lose.
01/04/2009
15
Figure 3.6
3.1.5
Scraped surface prepared for the next layer placing.
SOIL COMPACTION PROPERTIES DURING CONSTRUCTION.
To indicate the state of the compaction, dry density tests have been done using U100 steel
tubes. One undisturbed soil sample from each side of the compacted layer has been taken.
The sample was extruded from the tube and the weight along with the sample volume has
been measured. Example of the extruded sample is shown in Figure 3.7.
Figure 3.7
U sample taken after compaction.
100
01/04/2009
16
When the final compacted product after minimum number of compactor passes didn’t
meet required compaction state, additional passes were applied to the compacted layer.
Dry density tests were repeated and the final results were compared with the results obtained from the Standard Proctor Tests. The results presented in the Figure 3.8 below show
the measured dry density required for the construction.
Proctor compaction curve and dry density for compacted layers
2.000
Dry density from the
Proctor Test
1.950
Dry Density (Mg/m3)
Range of the dry density
for compacted layers
1.900
95% of Maximum Dry Density
1.850
1.800
1.750
1.700
5.0
10.0
15.0
20.0
25.0
30.0
Moisture Content (%)
Figure 3.8
Dry density measured for compacted layers.
The shear strength in the randomly chosen 6-9 points on each compacted layer has been
also measured. The measurements have been taken during the construction phase using
hand vane tester and the undrained shear strength was indicated to vary in the range of 6496 kPa, with the Moisture Content varied from 13% to 14.87%. All the measurements are
presented in the Figure 3.9 which shows that according to BS 5930:19994 the final
achieved consistency has been classified to be Firm and mostly Firm to Stiff. This again
meets the required in Specification for Highway Works – Earthworks, Series 600011 compaction state, and has confirmed that used compaction effort was appropriate. This is always very important that construction is designed as well as constructed with the standards
required.
01/04/2009
17
Undrained Shear Strength For Compacted Layers
Layer 1'
C o ns is t e nc y in
a c c o rda nc e wit h
B S 5930
150
Layer 1''
140
Layer 2'
Layer 2''
130
Layer 3'
Layer 4''
Layer 5'
Layer 5''
Layer 6'
Layer 6''
Layer 7'
Layer 7''
Layer 8'
Layer 8''
Layer 9'
Undrained Shear Strenth [kPa]
Layer 4'
Stiff
120
Layer 3''
Layer 9''
110
100
90
Firm to stiff
80
70
Firm
60
50
Soft to firm
40
Soft
30
Layer 10'
Layer 10''
20
10
Very soft
0
10
11
12
13
14
15
16
Moisture Content [%]
Figure 3.9
3.2
Undrained shear strength for compacted layers.
ENVIRONMENTAL CHAMBER
Environmental chamber was constructed on the top of the concrete flume in order to control the environment above the embankment. 1m high, 8 m long and 1.8 m wide wooden
frame was covered by the Perspex sheets and sealed by the silicone sealant. The chamber
shown in Figure 3.10, consists of 6 1.2kW infra-red heaters, 2 high speed fans, two pipes
with sprinklers, rain gauge and the thermostat which was coupled with the heaters and fans
to control the temperature inside. Whole heating system was automatically controlled and
was based on relays which were switching between heating and ventilating systems once
the adjusted temperature has been reached by thermostat. Rain gauge installed inside the
chamber, was set to 55 ml/m2 and when the expected precipitation has been achieved than
it alarmed automatically and the sprinkling system was closed manually.
01/04/2009
18
Figure 3.10
Environmental chamber with embankment and installed sensors.
4 INVESTIGATING USE OF GEOPHYSICS IN CRACKING
DETECTION
4.1
4.1.1
PILOT TEST
INTRODUCTION
In additional to the original aim of study it was decided to investigate whether geophysical
resistivity tests were a useful tool for detecting the onset and development of desiccation
cracking using geophysics.
Geophysical methods based on miniature resistivity arrays could be the solution as they
have proved to be reliable to monitor contaminant transport in soils scaled models in centrifuge experiments (Depountis8,1999). This technique is non-invasive and hence is reducing considerably the disturbance of the soil, improving the accuracy of the measurements.
For the assessment of embankments this technique could be very useful for long term
monitoring of clay sealings which are not accessible.
The subsurface soil properties are determined by measuring the distribution of resistivity.
The basis of the technique is to pass a direct current through the soil between a pair of
electrodes. This process is observed by monitoring the distortion of the equipotentials (assuming the soil to be a homogeneous half-space) using another pair of potential electrodes
located at the ground surface (Barker1, 1997). This provides a simple, repeatable technique
01/04/2009
19
that can be applied where any contrast in electrical conductivity exists in space (or time).
Lataste (2003) used the resistivity technique with a device made of four electrodes spaced
out 5 or 10 cm, arranged in a square to study the cracks on a damage concrete slab. He
used a numerical modelling approach rather than an inversion model showing a qualitative
similar disturbance of apparent resistivity right to cracks, for depth or opening variations.
On a wider scope at geology level, Nguyen17 (2005) proposed a methodology to locate
automatically limits or boundaries between synthetic faults and layer boundaries in two
dimension electrical tomography using a crest line extraction process in gradient images.
He found that the method showed poor results when vertical gradients are greater than
horizontal ones but otherwise should be systematically used to improve tomography interpretation.
In most of the Geo-electrical surveys the resistivity technique usually involves a computer
controlled multi-electrode arrays to give a tomography contour model of the subsurface in
two and three dimension (Griffiths and Barker1, 1993). One of the first team to use miniature resistivity imaging to detect cracks of cm size was Samouelian20 (2003). They used
porous special electrodes filled with CuSO4 similar to the one currently used for self potential measurements to improve the electrical contact in creating a wet contact with the
surrounding dried soil. They created artificially a crack of 2mm width with a saw at varying depths (1, 2, 3, and 4 cm deep) in order to obtain four cracking stages. The highest interpreted electrical resistivity was detected in the top 1.5-cm depth of their soil sample,
whereas the crack developed down to 4 cm. The electrical images obtained from these
electrodes enabled the detection of structures at the millimeter scale. More recently Tabbagh24 (2007) developed a new inversion model for assessing and simulating the electrical
response and the main physical parameters of cracks in soils. Their model allowed a faster
inversion of the experimental results. Rather than recreating artificially cracks and improve the inversion model, the present research has focused on the natural apparition of
cracks and their direct detection using resistivity arrays and then using viscous tracers.
4.1.2
EXPERIMENTAL SET UP
The same soil as for the embankment construction has been chosen for this pilot test in order to investigate the use of geophysics for cracking detection. 190 kg of clay was oven
dried to remove all the moisture and sieved using 20 mm mesh sieves. The sample was
mixed with water to obtain 15 % of soil moisture content. Then it was left for 24 hours for
a curing period. The chosen value of moisture content was within the range of moisture for
95% of maximum dry density, and was close to the natural moisture content measured in
the field.
After the soil preparation, a 2.5 kg compaction load was applied to several layers 5 cm
deep, over the 1.5m length, 0.25m wide and 0.4m high Perspex tank, secured by 3 steel
clamps. The transparency of the Perspex tank allowed checking the uniformity of the clay
compaction.
To identify the soil layers location already known, a miniature resistivity array was
adapted to be used with the ARES earth meter equipment purchased from the company Gf
instruments. 48 non-corrosive 1.5 mm diameter and 6 cm long electrodes (see Figure 4.1)
were wired up and connected with the automatic resistivity system using double 24 ways
connectors.
01/04/2009
20
Figure 4.1
Non-corrosive electrodes mounted in block terminals.
The electrodes were pushed 3cm into the compacted clay keeping a 3 cm spacing between
them. To initiate and perform the desiccation and drying process a 1.2 kW infra-red heater
was placed 0.9m above the clay surface, as shown in Figure 4.2.
Figure 4.2
Experimental setup with infra-red heater and electrodes installed.
The measurements selected option was a 2 Dimensional multi-electrodes resistivity profile. A Wenner-Schlumberger array profiling method as shown on Figure 4.3, was chosen
in this study because it is the most sensitive configuration to vertical resistivity changes
(horizontal structures) in the soil strata and the groundwater table, and it is also more sensitive than other arrays to the horizontal resistivity changes (vertical structures). Further-
01/04/2009
21
more, the extensive horizontal coverage and greater number of data points than other arrays justified its choice.
Four levels array of plotting points chosen for data analysis.
Figure 4.3
In the case presented here, only the first four levels and 168 plotting points were chosen
for data analysis in relation to the measurements taken outside the physical boundaries
(depth) of the flume model, which were identified as high resistivity measurements due to
the plastic interferences at the bottom of the Perspex tank.
4.1.3
EXPERIMENTAL PROCEDURE
The experiments were carried out for twelve days. Every morning the infra-red heater was
switched on and left for 6 hours to initiate and perform desiccation cracking. During this
drying stage the variations in soil temperature were recorded every hour. Then the Geoelectrical scan was carried out, and the equipment was switched off overnight. The initial
temperature of the soil in the morning was recorded at 20 cm depth every day before starting a new drying stage. Each last measurement was taken after drying has been finished
(see in Figure 4.4).
Temperature variations
40
Temperature [oC]
35
30
25
20
15
10
0
20
40
60
80
100
120
140
160
180
200
220
240
260
280
Elapsed time [h]
Figure 4.4
Temperature variations for 12 days drying process.
01/04/2009
22
4.1.4
RESULTS AND DISCUSSION
As it was mentioned before, the Wenner-Schlumberger method was used to measure the
resistivity changes in the soil during the desiccation process.
The experiments were carried out in the same spirit as during a field survey on a real embankment as it is the final goal. The numbering and detection of vertical crack was the
main target taking into account the limitation of the scanning method which was the electrodes geometry as the spacing kept between them was the same as the contact depth into
the soil.
The first scan was taken after the clay compaction and before drying in order to confirm
that the model was homogenous and was compacted to the expected state. The baseline
map presented in Figure 4.5, shows that the compaction was fully achieved and the measured resistivity was in the range of 20-30 ohm*m. This can be seen as a deep-blue color,
which is the resistivity contour obtained after inversion of the experimental measurements.
The temperature was increased from 15.0 ºC up to 37.0 ºC. The heat generated from the
infra-red lamp, generated a temperature of 35.3 ºC above the model, similar to a very hot
summer day.
Figure 4.5
Inversed Resistivity map taken before drying process and after 4 hours of
heating.
The resistivity measurements were taken every morning, before the IR heater was switched
on and every afternoon after the drying was finished. Due to the small changes in readings
caused by evaporation process only morning’s measurements were analyzed.
Figure 4.5 and Figure 4.7Figure 4.6 show how the resistivity changed with time and with
cracks formation.
01/04/2009
23
Figure 4.6
Comparison between two measurements taken after 20 hours and 27 hours
of drying.
Figure 4.7
Comparison between two measurements taken after 43 hours and 50 hours
of drying.
It can be clearly seen on the resistivity profile presented on Figure 4.6 c) and Figure
4.6Figure 4.7 d) that several vertical openings occurred and became wider with time (vertical purple channels) corresponding visually to the same vertical cracks location on the
laboratory clay sample.
However, it can be explained that during the fast heating process, the decrease in soil resistivity may be due to two phenomenons. The first can be described as micro swelling of the
clay and closing up of the cracks caused by evaporating water represented by the disappearance of the light green vertical afternoon contours shown on Figure 4.7. The second
01/04/2009
24
phenomenon could be due to the moisture content redistribution in the model due to
evaporation and the water rising to the surface from deeper regions. A post mortem excavation of the model, and measurements of the moisture content (Figure 4.8) at different
depths, have shown a good agreement with the second assumption. The moisture content
in the sample was between 8 and 9% below 5cm depth and decreased sharply in the top
5cm where the cracks developed.
Moisture Content Profile
0
1
2
3
4
5
6
7
8
9
10
0
2
4
D ep th [c m ]
6
8
10
12
14
16
Figure 4.8
Moisture content profile measured during excavation.
The inversed resistivity map, shown of Figure 4.9, has confirmed that the air space created
inside the crack should give the high resistivity response in analyzed measurements shown
as vertical purple (dark) wide channels. It has to be noted that the photographic picture of
the cracks was made of 3 different pictures and was not at the same scale as the resistivity
profile. The corresponding location of the cracks was given as an indication.
Figure 4.9
Inverted resistivity map with corresponding picture of the cracked surface.
The resistivity results seem to reflect well the real desiccation cracking by comparing them
with visual observations of the clay model.
01/04/2009
25
5 MEASURING EQUIPMENT
5.1
INTRODUCTION
An excellent verification framework has been provided by Peck18 in “Observational Approach” (1969) for geotechnical engineering and the same approach needs to be applied to
unsaturated soil problems. Only in this way is it possible to develop confidence in the application of unsaturated soil theories. Matric suction and moisture content are the measurements that can provide verification information on unsaturated soil behaviour. That’s
why the engineers must have access and consider use of devices that will help them evaluate the adequacy of an engineering design (Fredlund10 2006).
5.2
MOISTURE CONTENT PROBES – TDR’S
Measurements of gravimetric moisture content provide a reference for the amount of water
contained in a soil. They are very important and can play a big role in verification. However, this method is an intrusive, destructive and not satisfactory method even it provides
indirect and quick measurements as an advantage. Time domain reflectometry sensor consists of 4 metal rods which should be inserted into soil during measurements. An electrical
pulse is sent to the end of the rods (and returns) and the results provide a measure of dielectric constant of the soil. The dielectric property is dependent on the amount of water in
the soil and the measurement can be converted to the volumetric water content (Fredlund10
2006). Figure 5.1 shows an example of used TDR probe.
Figure 5.1
Example of TDR probe used in the experiment.
A specific calibration curve for the soil needs to be used to achieve the proper accuracy of
the measurements. The example of the calibration curve is shown in Figure 2. In some
situations the salt content may affect the measurements; however it is not subjected to hysteretic effects.
01/04/2009
26
Calibration curve for TDR sensor - 4
1000
900
800
Output Voltage [mV]
y = 0.0637x3 - 4.373x2 + 106.43x - 47.067
700
600
500
400
300
200
Moisture content
100
Poly. (Moisture
content)
0
0.0
5.0
10.0
15.0
20.0
25.0
30.0
Figure 2 Specific calibration curve for used soil.
5.3
SOIL WATER TENSION SENSOR – SWTS 1
Generally, the most preferred measurements in unsaturated soil mechanics are the measurements of matric soil suction. Over the years, there has been proliferation of sensors and
devices, particularly related to the measurement of soil suction in the agronomy or agriculture-related disciplines (Dane and Topp7 2002). Geotechnical engineers need to ensure that
the devices have sufficient accuracy for engineering purposes. SWTS 1 sensor compromises three direct different measurements. It can measure suction of the unsaturated soils
as well as positive pore water pressure in saturated soils, and also it can measure the temperature of the water contained in the soil which can be considered as the temperature of
the soil. The range of the sensor is from -100kPa to +100kPa, and in some cases this can
be not enough range when measurements are taken in very dry soils. One of the advantages
of this sensor is that, when it cavitates it refills itself once the water gets inside the sensor
after the rainfall. There is no maintenance needed over the measuring period especially after dry and hot summers. Before the installation in the soil, tensiometer doesn’t need to be
calibrated for the specific soil. It contains single calibration curve which is supplied by
manufacturer, as it is shown in Figure 3.
01/04/2009
27
Single calibration curve for Tensiometer - SWTS1
150
100
Pressure [kPa]
50
0
0
500
1000
1500
2000
2500
-50
-100
-150
Output vlotage [mV]
Figure 3
Single calibration curve for tensiometer - SWTS1.
The sensor can be installed either downwards or upwards but it has to be inserted in the
angle and in the position when the typical flow is not disturbed by the tensiometer, as it is
shown in Figure 4. For the air removal from the ceramic cup the ideal angle for the installation from the surface is 25o to 65o from the vertical line. Thus, the correction offset needs
to be applied to the measurements.
Figure 4
Downwards installation with the angle.
01/04/2009
28
Angle installation allows sensor in correct self-refilling and helps completely removing air
bubbles from the cup. It has been also discovered that during the rainfall water according
to the gravity movements of the water reaches the ceramic cup faster than in case of vertical installation. Figure5 shows the example of used sensor.
Figure5
Example of SWTS 1 used in experiment.
The most important thing during installation is to inject the slurry paste of the soil which is
going to be investigated. This needs to be injected inside the drilled hole to make a good
contact in between the ceramic cup and soil.
Described tensiometers have been installed on four different depths the crest (0.2m, 0.4m,
0.6m, 0.8m) with the angle of 25 o. 25% moisture content slurry paste has been injected inside the augered hole using silicone tube supported by the silicone gun. Hole was filled in
on its whole depth in order to seal the gap between the sensor and the soil. All the ceramic
cups were directed on the upstream side of the embankment.
5.4
SOIL WATER TENSION SENSOR – T4E
All water movements in soil are depending on the soil water tension as water – in soils as
well as on the surface – always will move from a point of higher potential to a point of
lower potential (UMS25 2007). The most of the soil water movements are taking place
when the small water tensions are occurring. Tensiometers are the best sensors which allow direct and very precise measurements of these small forces which are called tensions.
In the heterogeneous soils not only precipitation and evaporation effect the process, but
also texture, particle size distribution, cracks, compaction, roots and air voids. Due to
these heterogeneities the soil water tension may vary and change dramatically especially
when the soil is cracked. In general, the purpose of use the tensiometers, is to measure the
soil water tension respectively of matrix potential. Presented tensiometer, can measure the
01/04/2009
29
pore water pressure from +100kPa (water pressure) to -85kPa (suction of the soil). If the
soil is drier than -85kPa, tensiometer gets dry and cavitates and needs to be reffiled before
the soil is moist again. The tensiometer will dry very quick even it is refilled as long as
the wet period does not come in. Water contained in the soil has a contact with the tensiometer through the ceramic cup which is porous and lets the water get inside the cup.
This creates an ideal interface between both soil and water. The pressure transducer inside
the ceramic cup gives direct and continuous measurements of water tension. The reference
pressure is measured through membrane installed on the cable. Chosen T4e is equipped
with two capillary tubes (refilling tubes) and can be refilled in the field without removing
it from the soil. Figure 6 shows an example of used T4e Tensiometers.
Figure 6
Example of T4e used in the experiment.
However, T4e must be installed with the care and in the way as described in the previous
paragraph. The single calibration curve, shown in Figure 7 for the T4e tensiometer is also
supplied by the manufacturer, and the sensor doesn’t need to be calibrated for the specific
soils.
01/04/2009
30
Calibration curve for Tensiometer - T4e
0
0
10
20
30
40
50
60
70
80
90
-10
-20
Suction [kPa]
-30
-40
-50
-60
-70
-80
-90
Voltage [mV]
Figure 7
Single calibration curve for tensiometer – T4e.
Four of these tensiometers were installed on the upstream slope, where two were installed
at a depth of 20 cm from the surface and the other two at 40 cm from the slope surface and
the same installation procedure has been used.
5.5
RELATIVE HUMIDITY & AIR TEMPERATURE SENSORS – RHT2NL
The RHT2nl sensor compromises two independent sensors Relative Humidity and air temperature transducer which are housed in the radiation shield. The transducers require
power and provide two output signals for the RH and air temperature. The RHT2nl gives
very high precision measurements with the output range 0 to 100% for Relative Humidity
and -50oC to 150oC for air temperature. These sensors are design to work as a part of the
weather station, but when needed, they may work independently when plugged into the
Data Logger. Figure 8 shows an example of the used sensor.
Figure 8
Example of RHT2nl sensor in the solar radiation shield.
01/04/2009
31
Both sensors give very accurate and direct measurements and as previously described sensors they don’t need to be calibrated before use. Calibration curve is supplied by manufacturer.
RHT2nl sensors were mounted in the solar radiation shield 1 m above embankment.
All the sensors described in the previous paragraphs, have been particularly tested and than
connected to the Delta-T DL2e Data Logger, which was programmed to take measurements every 30 s during the experiment.
5.6
MINIATURE RESISTIVITY ARRAY
Non-invasive geophysical technique has been used on the downstream side of the embankment in order to avoid any disturbance of the slope which in the further study was
subjected to flooding. Two miniature resistivity arrays as used in the pilot test have been
installed in two directions; 48 electrodes of the first array have been pushed with 3.5cm
spacing in the central part across the slope, and another 48 electrodes have been installed
along the walls.
6 DESICCATION EXPERIMENT
6.1
PHASE ONE – DESICCATION WITH ONE WEEKLY PRECIPITATION
After successful installation of all of the sensors, one week of the testing procedure has
been performed in order to check if the sensors are working in the proper way and responding to the changes in behaviour. During the testing time all tesiometers reached the
equilibrium with the surrounding soil. This was because the moisture content of the slurry
(25.44%) injected inside the installation hole was higher than the exact moisture content of
the embankment body.
Having available facilities it was impossible to replicate very low temperatures especially
to reduce the temperature below the ambient temperature in the laboratory. Thus, why for
all of the desiccation phases, temperature on the thermostat was set to 20oC and only the
weekly precipitation has changed over the testing period. A visual observation from the
first few hours of the experiment has shown that the cracks started appearing on the surface. It can be clearly seen in Figure 6.1 and Figure 6.2 a few tiny cracks which have appeared on the crest as well as on the upstream side of the embankment.
01/04/2009
32
Figure 6.1
Figure 6.2
First cracks observed on the embankment.
Cracks observed on the upstream side of the embankment.
During the first six weeks, 55mm/m2 of rainfall has been introduced every week inside the
environment. After the expected precipitation has been reached and measured by installed
rain gauge, system has been closed, and the drying system has been switched on again.
Data logging system was continuously measuring and recording any change in the environment. During this experimental phase averaged recorded humidity was around 50% and
was only increasing after precipitation due to the evaporation effect. It has been observed
that during the rainfall water on the crest of the embankment was disappearing inside the
cracks, and was flowing out from the cracks on the slope. It has been also seen that water
flowing on the slope surface, once reached the cracks edge was flowing inside them and
distributing by cracking network.
Under normal circumstances, when the soil is in the equilibrium with water, clay swell
when gets wet, but observed wide cracks which are shown in Figure 6.3 that have created
after the first week of drying, have stayed open even the moisture content of the soil increased. More pictures from the first week of drying are attached in the Annex of this report.
01/04/2009
33
Figure 6.3
Cracks observed on the crest after 7 days of drying.
This can be described as an “additional evaporation” effect, when the soil is fractured and
it helps the water to evaporate quicker, not only through the surface but also from inside of
the structure through the cracks. This assumption has been confirmed in further data analysis from the measurements taken by TDR sensors installed inside the embankment. It can
be clearly seen from the moisture content profile in Figure 6.4 that in the upper part of the
embankment which was cracked and open to the atmosphere moisture content varied in
the big range.
Moisture Content Profile - 1 cycle of rainfall
12
13
14
15
16
17
18
19
20
21
22
23
24
25
0
0.1
0.2
Depth [m]
0.3
0.4
0.5
0.6
1st week
0.7
2nd week
3rd week
0.8
4th week
5th week
6th week
0.9
Figure 6.4
Moisture content profiles representative for first 6 weeks of experiment.
01/04/2009
34
In the same time, decrease of the moisture content has been recorded on the bottom of the
embankment along with the almost constant moisture in the middle part. This may be due
to the moisture content transport from the central part of the embankment to the drier (upper) part, from the upper part (after the rainfall) to the middle part and from the moistest
bottom part to the drier - central part.
The shape of the moisture content curve matched the one found by Black and Croney2
(1958) in their studies.
A similar effect has been observed in the suction profile. Presented results are the measurements recorded by the SWTS1 tensiometers installed on the crest of the embankment.
Over the 6 weeks of the experimental period even the suction on the top of the embankment decreased after the rainfall and primary increased because of drying the suction on
the bottom was still increasing. The direction of the increase in suction is shown by the red
arrow in the Figure 6.5 below. Again, as it was seen in the moisture content profile, the
suction in the middle of the structure stayed almost constant with only small increase.
From the observations made during this phase of experiment it can be seen that the suction
is mainly dependent on the moisture content, its movements and changes. Observed
change in the suction profile is an effect of the moisture content transport between three
water regimes which are described above.
Suction Profile - 1 Cycle of Rainfall
Suction [kPa]
-110
-90
-70
-50
-30
-10
0
0.1
0.2
Depth [m]
0.3
0.4
0.5
0.6
0.7
1st week
2nd week
0.8
3r d week
4th week
5th week
6th week
Figure 6.5
0.9
Suction profiles representative for first 6 weeks of experiment.
01/04/2009
35
It has been also observed that after and during every rainfall fines from the surface have
been washed away from the surface and visible erosion has been noticed. However, the
structure was still stable and no slide or disturbance has been recorded after the moisture
content increase and the clay softening.
Measurements using resistivity array have been taken every morning and every afternoon
as well as additional scans after the rainfall in order to map any changes in the embankment body. Scans taken after the rainfall have helped to understand the physical processes
especially to see if the water could remain inside the cracking network or if there is any
self-healing of the cracks.
The initial scan taken before drying experiment has confirmed that the body of the embankment is homogenous, and no cracking occurred since the construction was finished.
Inversed resistivity model shown in Figure 6.6 is the initial representative scan taken by
the across array shows that the uniformity of the moisture content has been achieved.
Figure 6.6
Inversed resistivity scan taken before the experiment.
A deep blue colour on the picture represents very low resistivity (~50Ωm) and only small
changes can be seen on the bottom of the model which might be due to the disturbance
caused by the boundaries effect. The same was observed in the pilot test presented earlier
in this report, where the miniature resistivity array has been tested and some interferences
on the bottom of the model have been caused by the Perspex.
01/04/2009
36
Figure 6.7
Resistivity scan with corresponding visual observation after 7 days of drying.
Figure 6.7 shown above is representative for the visual observation of the embankment after 7 days of drying. It can be clearly seen that changes in the structure, especially cracking
was mapped by the miniature resistivity array. Deep purple contour represents places with
the high resistivity values such as air voids. These discontinuities, which have created
within the period of seven days, were seen appearing and becoming larger every day. Inversed resistivity baseline maps from the first week of experiment are attached in the Annex of this report.
It has been seen that after the rainfall, when the soil became moist, measured resistivity
values decreased as some of the high resistivity places have change their contour colour.
As it is shown in Figure 6.8, the upper inversed resistivity scan has been taken in the
morning, before the rain was introduced on the embankment. It can be clearly recognized
where the places with high resistivity have widely grown after one week of drying. On the
second scan which has been taken in the afternoon after the rainfall a deep purple colour
which is representative for the high resistivity values in some places became smaller as the
conductivity has changed after moisture content increased. This can be also due to the water which remains in the horizontal cracks until it soaks in the soil or evaporates.
01/04/2009
37
Figure 6.8
6.2
Inversed resistivity scan taken before the experiment.
PHASE TWO – DESICCATION WITH TWO WEEKLY PRECIPITATION
In the second experimental phase which is presented in this paragraph one additional precipitation has been introduced every week and the experiment was run over the period of
four weeks. In every mid-week, the same amount of rainfall as on the beginning of the
week (55mm/m2) was simulated with the same procedure as described previously in order
to investigate behaviour of cracked embankment when subjected to the wetter conditions.
Presented below, moisture content and suction profiles have been measured by the TDR’s
and SWTS1’s installed from the crest of the embankment and than plotted in the averaged
daily profiles where each colour corresponds to the one week of measurements.
Moisture Content Profile - 2 Cycles of Rainfall
12
13
14
15
16
17
18
19
20
21
22
23
24
25
0
0.1
0.2
Depth [m]
0.3
0.4
0.5
0.6
0.7
1st week
2nd week
3rd week
4th week
0.8
0.9
Figure 6.9
Moisture content profiles representative for 4 weeks of experiment.
01/04/2009
38
It can be seen from the Figure 6.9 that the moisture content increased in the upper layer,
when the weekly precipitation increased. The soil water regime has became constant in the
middle part of the embankment, with the only small changes as well as the bottom part,
which has stabilized close to the moisture content readings from the first phase. A small
increase has been observed in the last few days. This means that sufficient amount of water has been supplied to stop the water transport between three water regimes (uppermiddle-bottom) and that drying time between rainfalls was to short to induce this transport.
The increase of suction seen in Figure 6.10 which has finally changed to the increase in
positive pore water pressure was partially caused by the rainfall and it is again part of the
physical processes and movements of the water inside the embankment, which was accumulated in the centre of the structure.
Suction profile - 2 Cycles of Rainfall
Suction/Positive pore water pressure [kPa]
-110
-90
-70
-50
-30
-10
0
0.1
0.2
Depth [m]
0.3
0.4
0.5
0.6
0.7
1st week
2nd week
3rd week
4th week
0.8
0.9
Figure 6.10
Suction profiles representative for 4 weeks of experiment.
Figure 6.11 shows the difference and changes in resistivity measurements which were
taken the day before and after the second rainfall introduced in the middle of the week.
Measured resistivity values have confirmed that observed increase in moisture content significantly changed the water regime and soil became more conductive. This scans has confirmed the assumption that this water transport was taking place, as the soil was much
more resistive on the bottom and top of the embankment. Fact that some of the high resistivity vertical places cannot be seen might be explained as the cracks sealing by the fines
eroded from the surface. However, this cannot be taken as an advantage of this process because this fine material can be washed away from the cracks again.
01/04/2009
39
2 inversed resistivity maps for scans taken before and after the rainfall.
Figure 6.11
6.3
PHASE THREE – DESICCATION WITHOUT PRECIPITATION
During last desiccation phase extremely dry conditions have been applied to the experiment. Embankment has been subjected to drying conditions in order to dry out the structure that can be subjected to the flooding conditions. Thus, the experimental temperature
of 20oC has been kept constant inside the environmental chamber along with data collection and geophysical scanning over the period of 5 weeks. It can be seen in Figure 6.12
that the structure, on its full depth was drying out. The middle part of the embankment had
higher moisture content than the upper and bottom part, which again is due to the moisture
transport/exchange effect between three parts as it was described in the previous paragraph. The moisture content of 14.64% has been measured in the upper layer with 16.58%
indicated in the bottom part and 19.46% in the middle part. Collected data show how big
differences have formed inside and how the structure behaved.
Moisture Content Profile - No Rainfall
12
13
14
15
16
17
18
19
20
21
22
23
0
0.1
0.2
Depth [m]
0.3
0.4
0.5
0.6
1st week
0.7
2nd week
3r d week
0.8
4th week
5th week
0.9
Figure 6.12
Moisture content profiles representative for 5 weeks of drying at 20 C.
o
01/04/2009
40
Looking at the suction profiles presented in Figure 6.13 it can bee seen that initiated drying
phase has resulted in very fast suction increase. Big “jumps” in the negative pore water
pressure had an effect in drying out of the ceramic cups and letting the air get inside the
sensors, which are only working when they are filled with water. Thus, the upper sensor,
which was installed only 20 cm below the surface has cavitated as first and was only recording the maximum of its range. This means that 100kPa of suction has been achieved
after the first week of drying and possibly exceeded as the embankment was still drying
out. Another 3 sensors which were installed deeper, have been working until they exceeded their range and when the drying front has reached their depth. However, this happened after four weeks of drying and until this the whole embankment sustained wet. It has
been observed that the structure was drying faster, as the cracks have opened more and
progressed inside the embankment. This allowed water to evaporate from the body and to
disturb the water regime inside. The upper-middle-bottom moisture exchange was still observed and can be clearly seen on the graph as a bulge in the middle part on daily suction
profiles. This lower suction value in the centre of the embankment has followed presented
trend for about four weeks, and disappeared one week before the drying phase was finished. This might be due to the vertical cracking reaching the central part of the embankment, and letting the water evaporate quicker from the area where the moisture was accumulating.
Suction profile - No Rainfall
Suction [kPa]
-110
-90
-70
-50
-30
-10
0
0.1
0.2
Depth [m]
0.3
0.4
0.5
0.6
0.7
1st week
2nd week
3rd week
4th week
5th week
0.8
0.9
Figure 6.13
Suction profiles representative for 5 weeks of drying phase.
Figure 6.14 shows the results from the resistivity measurements taken before flooding experiment, when the soil was expected to be extremely dry as it was seen from the moisture
content and suction measurements when the soil moisture decreased and the suction draUR14_desiccation_fine_fissuring_WP4_1_v4_0 Final
01/04/2009
41
matically increased. It can be seen on the map presented below, how the places with high
resistivity have enlarged and the whole structure became more resistive.
Figure 6.14
Inversed resistivity map taken before flooding when the soil was dry.
Desiccation cracking shown in Figure 6.15 and Figure 6.16 which has created during desiccation process inside the embankment body has confirmed that vertical cracking occurs
from the surface to the certain depth. It is not so obvious at what depth horizontal crack
will create. However, in this experiment horizontal crack has created at a depth of about
25 cm, and was connected with the vertical cracks from the upper layer as well as with the
vertical cracks underneath horizontal crack. Vertical cracks which were seen in the upper
surface layer were wider than deeper ones as this layer was more subjected to the weather
variations. It is difficult to say that for every type of the material and every type of construction, cracks will create in the same place, at the same depth and with the same extent.
This strictly depends on many factors, such as: drying temperature, type of soil, density of
the soil, variations in wetting-drying processes and many others.
Figure 6.15
Observed desiccation cracking inside embankment body.
01/04/2009
42
Figure 6.16
Observed desiccation cracking inside embankment body.
It has been also proved that horizontal cracks follow the shape of the embankment as proposed by Dyer et al. It can be seen in Figure 6.17 that crack which created in the corner of
the window is parallel to the embankments slope, marked as a yellow line on the wall.
Figure 6.17
Crack parallel to the slope.
01/04/2009
43
7 FLOODING EXPERIMENT
7.1
INTRODUCTION
After successful completion of desiccation experiment, flooding experiment has been performed in order to investigate how the cracked structure behaves when is subjected to
flooding conditions and to experimentally confirm the failure mode.
As it was described previously, studies carried out by Cooling and Marsland during their
Cofferdam test, have reveal that when the flood appears and rises water can flow through
the fissured zone, and than initiate the failure on the outward face of the embankment.
However, it is not very clear from their study what kind of failure they’ve actually seen
and how long it took for the structure to fail. That’s why these and the other unknowns had
to be tested and proved by the physical experiment and visual observations. The assumptions made in this project have considered the failure mode caused by the uplift mechanism of the clay blocks which can be due to the pressure building inside the embankment.
7.2
EXPERIMENTAL SET UP
The same experimental set up has been used in this experiment. All cavitated tensiometers
have been refilled with water before the experiment as they could work and measure again.
Two gaps which created along sides of the embankment as the effect of shrinking and can
be visible in Figure 7.1 were sealed using expanding foam in order to prevent any water
flowing on through the side ways. Data logger has been reset to zero and all the sensors
have been programmed to take the readings every 5 seconds to be able to measure any
changes immediately ie.: capillary wetting front or internal seepage.
Figure 7.1
Embankment model with the experimental set up.
01/04/2009
44
It can be also seen from the picture above that two miniature resistivity arrays have been
used during the flooding experiment in the same set up as during desiccation process in
order to map the changes inside the structure and to detect internal seepage and subsequent
capillary wetting following flood.
7.3
EXPERIMENT
Before the upstream side was filled with water last geophysical scan has been taken and
used as the reference map for further scans. Results shown in Figure 7.2 are representative
for the extremely dried structure where the resistivity 400-800 Ωm is a dominant value.
Figure 7.2
Inversed resistivity map taken before flooding.
After the last resistivity scan, inward side of the embankment has been gradually filled
with water to the level of the first tensiometer installed on the slope (as shown in Figure
7.3). The minimum available water flow has been set up on the pump to minimise any disturbance of the cracked structure as well as to see if there is no any potential side walls
leakage or seepage occurring on the outward side.
Figure 7.3
Channel filled with water to the level of first tensiometer.
01/04/2009
45
As it can be seen from the Figure 7.4 there was no seepage occuring on the downstream
side of the embankment which is due to the fact that the bottom part of the embankment
was not affected by desiccation cracking and lasted impermeable.
Figure 7.4
Downstream side of the embankment after one hour of flooding.
After 1 hour water level on the upstream side was gradually raised again to the level of
third tensiometer installed on the slope and left for another hour, as it can bee seen in
Figure 7.5.
Figure 7.5
Upstream side of the embankment with the water level 20 cm below the
crest.
01/04/2009
46
Once the water level raised and the water pressure behind embankment increased, first
leaks on the bottom downstream side of the embankment have been observed (as shown in
Figure 7.6). Two of them were due to the sideways seepage underneath the sealant, as the
water is always trying to find the easiest way to seep through the fractured structure. However, these leaks were to small to stop the experiment as they could enlarge and cause unexpected failure.
Figure 7.6
First observed leaks on the downstream side.
Figure 7.7 and Figure 7.8 show that the most interesting observed seepage appeared on the
bottom-centre of the slope.
Figure 7.7
Local seepage.
01/04/2009
47
Figure 7.8
Local seepage from the bottom of the slope.
This seepage has proved the theory about the cracks network which was proposed by
Cooling and Marsland15 and later by Dyer et al9. This network creates when the structure is
subjected to desiccation process and when the particular cracks are interconnecting.
During the visual observations, scans using geophysics have been taken at the same time
in order to map changes in the soil resistivity as well as to be able to detect the capillary
wetting front. Figure 7.9 shows the results from two scans taken during first and the second flood rise on the upstream side. It can be clearly seen from these two scans how the
resistivity values decreased during the water level increase and how the capillary wetting
has changed the water regime inside the embankment.
Figure 7.9
Two inversed resistivity maps taken during first stage of flooding.
The disappearance of the high resistivity places, can be explained as the water filling of the
cracks and the moisture content increase around the cracks which than increased conductivity of the soil.
01/04/2009
48
As it is presented in Figure 7.10 during the next step, water level has been raised till it
reached the crest of the embankment in order to observe the water activity through the surface fissures. It is very important to say at this stage, that some of the water which was distributed by the surface cracks was seeping inside them and some of the water was flowing
farther up and filling the other cracking paths.
Figure 7.10
Water penetrating cracks on the crest.
After the water has reached the edge of the downstream slope and filled up the crest cracks
than it started to flow through the slope cracks and finally out from the slope. There was
no failure observed until the flow rate behind the embankment has been increased up to
600 l/min and overflowed and kept for another 4 hours.
Figure 7.11
Overflowed crest.
01/04/2009
49
It can be seen from the Figure 7.11 how the flow rate increased in the place of the first
seepage and than stopped which might be due to the local collapse or internal sealing by
the fines brought by the water. There were another two additional seepages observed
which have created on the slope during this overtopping.
Observations made during the first day of the experiment have made very clear understanding of the behaviour when the cracked structure is subjected to flooding. It was
clearly seen that there is no uplift mechanism when the structure is subjected to the flooding. Erosion process and undercutting of the clay blocks were the major observed factors
which were causing local failures and internal collapses. Sediment transport which has
been very clearly seen could fill up the cracks, but only in the case when the drop in the
water level occurs which allows settle all the fines. In the case when there is a water
movement in the fractured structure, all of the sediment will be transported with the water
flow and the erosion will become larger as the soil is gets softer with the moisture content
increase. An example of the eroded crack is shown in Figure 7.12.
Figure 7.12
Eroded crack inside embankment close to the edge of the slope.
It has been also seen that the phreatic line of the seeping water is non-uniform as for the
non-cohesive soils, and it depends on the cracking network (see figure Figure 7.13).
Figure 7.13
Observed phreatic zone.
01/04/2009
50
It can be see from the Figure 7.14 and Figure 7.15 that erosion affected only half of the
crest and the downstream side of the embankment and the other half was sealed by the
sediment transport which is compatible with the water flow direction. A measured depth
of the erosion has shown that every day the erosion was progressing of about 5 cm.
Figure 7.14
Figure 7.15
Eroded slope and the half crest.
5 cm deep cracks on the crest after first day of flooding.
After one week of the overtopping the crest of the embankment it can be see in Figure 7.16
that two flow paths have created on the downstream slope where the most of the water and
most of the energy was directed to these paths. That’s why the most of the erosion was taking place only in the central part of the slope instead of its whole width. This erosion has
exposed another inside cracks which were not visible from the surface when the embankment was desiccated.
01/04/2009
51
Figure 7.16
Eroded slope with the two flow paths visible in the middle of the slope.
As it can be seen in Figure 7.16 erosion process was the main failure as the clay was softening. Thus, the water head above the embankment crest has been increased to about 10
cm by putting the wooden barrier across the channel, as it is shown in Figure 7.17. All the
gaps underneath the barrier have been sealed by expanding foam and bags with the sand in
order to stop the water seeping below the barrier. The idea was to increase the hydrodynamic forces. However, it only resulted in the increase of the erosion, but again there was
no uplift mechanism due to the increase of the pore water pressure above the crest. Once
the water reached the edge of the barrier it was distributed between the cracks and mainly
two flow paths where the erosion progressed. The cracks were getting deeper end wider. A
depth of 25 cm has been reached after two weeks of flooding which is a depth where the
horizontal crack has been found at the end of the desiccation experiment.
Figure 7.17
Increased water head with the wooden barrier.
01/04/2009
52
As it was mentioned before, undercutting of the clay blocks, erosion, softening of the clay
and dragging forces were the major factors which were affecting upstream slope during
flooding. Series of pictures presented below can prove visual observations made during
experiment. This is very well documented processes of one clay block which was previously undercut by the water and than pushed away after collapse. It can be seen in Figure
7.18, Figure 7.19 and Figure 7.20 that increase in water had above embankment resulted in
increase of hydrodynamic forces.
Figure 7.18
Figure 7.19
Undercutting of clay block.
Visible crack in collapsing block.
01/04/2009
53
Figure 7.20
Collapsing clay block.
After observed collapse presented block was pushed down through the flow path by dragging forces and than washed away, as shown in Figure 7.21 and Figure 7.22.
Figure 7.21
Figure 7.22
Clay block pushed away.
Visible small collapse in the clay block.
01/04/2009
54
Measured moisture content profiles presented in Figure 7.23 have shown how the capillary
wetting front has changed the moisture regime inside the structure. It can be clearly seen
that the bottom part as well as upper part of the embankment have increased their moisture
content as they were much more drier than the central part. Wetting of the bottom part
could be also forced by the water pressure behind the embankment as it was higher than in
the other parts. This may also explain why the moisture content on the bottom of the embankment was higher than in the other parts.
Daily moisture content profiles during flooding
Moisture content [%]
12
14
16
18
20
22
24
0
0.1
0.2
Initial
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
Day 8
Day 9
Day 10
Day 11
Day 12
Day 13
Day 14
Day 15
Day 16
Depth [m]
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Figure 7.23
Soil moisture content profiles during flooding experiment.
A different effect has been observed in the suction profiles shown in Figure 7.24. Observed over the flooding experiment suction has decreased quicker in the upper (3 days)
and middle part (5 days) of the embankment when the bottom part was still under suction
for 7 days. This can be explained as the upper, rubbilised part of the embankment, has become wetter as the water could penetrate the cracks faster and be faster sucked by the dry
soil. When the bottom part is still untouched by desiccation cracking and impermeable,
water needs more time to capillary penetrate this zone. After decrease in the suction, all
tensiometers have been working as piezometers and were recording piezometric water
level corresponding to flooding conditions and responding to any change in the water level
behind and above the embankment.
01/04/2009
55
Daily Suction/Piezometric water head Profile
Suction/Positive pore water pressure [kPa]
-110
-90
-70
-50
-30
-10
0
10
30
50
70
90
110
0.1
0.2
Depth [m]
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Figure 7.24
Initial data
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
Day 8
Day 9
Day 10
Day 11
Day 12
Day 13
Day 14
Day 15
Day 16
Suction/positive pore water pressure profiles during flooding experiment.
01/04/2009
56
8 DISCUSSION ABOUT USER FOCUS OUTCOMES
8.1
8.1.1
STATED OBJECTIVES AND OUTCOMES
FAILURE MECHANISM FOR BREACH INITIATION
The failure mechanism postulate in UR119 was not investigated in the model embankment
test which instead investigated a different crack pattern based on large widely spaced
cracks brought about by intense drying of a bare surface. Despite the large cracks pattern
observed some superficial fine fissuring was also observed.
No uplift mechanism of the upper crest was observed and we thought it would be useful to
remind the past research on this phenomenon in case other practitioners would observe
such a failure on the field.
Failure mechanism that was postulated in the interim UFMO UR119 has to be withdrawn.
Investigation done by Cooling and Marsland5 and later by Marsland14, which was cited in
the previous report as an example of the possible failure, was related to the uplift mechanism (Figure 1 (b)) and uplift pressures which are due to the excessive pore water pressures in the underlying sandy gravel which resulted in the flotation of the overlying alluvium in the marsh and complete collapse of the flood defence. This type of failure was
also proposed by Van et al.26 in the later studies as shown on Figure 2. Theoretically, if
the sand layer is incompressible and completely open to the river, and the overlying layer
on the landward side is impervious, then the pressure in the sandy gravel layer becomes
equal to the river and every rise in the water level in the river will result in the increase of
the pressure in the permeable layer.
The uplift mechanism mentioned above was observed on the landward side of the embankment close to the toe and was followed by the shallow slip of unsupported back slope.
Figure 1 Failure caused by seepage pressures in underlying coarse layers (after Cooling et
al5).
01/04/2009
57
Figure 2 Schematic process of uplift induced slope failure (after Van et al ).
26
Figure 1 (a) shows the second failure proposed by Cooling and Marsland5 and relates to
underseepage and excessive pore water pressures mobilising in the pervious layer which is
covered on the landward side by the less pervious layer of sand. Under high water heads,
“boils” and “springs” may develop near the landward toe of the bank. This can then lead to
complete failure of the embankment as the toe of the embankment may be undermined,
especially in case when it become violent and widespread.
The theory presented above is understandable and acceptable as the increase of the pore
water pressure is occurring in the permeable sandy or sandy gravel layer, where the water
flow is much more fluent than in the impermeable clays.
It has to be pointed out that the embankment model constructed in the concrete flume was
constructed on the impermeable clayey berm and additionally protected by the cut off leakproof membrane in order to stop any seepage through the embankments on its lower parts,
as well as to concentrate the water flow on the fissured zone.
It was stated in the UR119 that Cooling and Marsland5 observed the increase in pore water
pressure into the fissured zone in much the same way as if the core had been covered with
the gravel blanket. As they were using simple piezometric pipes to measure the hydrostatic
head, they possibly observed the increase of the pore pressure on the back slope which was
due to the velocity of the water flowing through the fissured layer which always increases
on the incline plan. This is contradictory to Marsland13 findings, later copied by Dyer et
al.9 in UR119. In this case, only dragging forces can be considered as the factors playing
major role in the embankments failure. It can be only agreed that a similar flow of water
occurred in the fissured zone on tested embankment model to the one which can be seen in
the gravel blanket.
The excess pore pressure would be caused by combination of changes in static and dynamic pressures.
01/04/2009
58
Water flow in the clayey embankments is always neglected as they are constructed as impermeable and homogenous structures. This was successfully tested and proved in the
laboratory studies. Figure 3 and Figure 4 show the measurements of the pore water pressure during flooding experiment. It can be seen from both figures that there was no significant increase in pore water pressure measured by piezometers and even after increase of
the water head above embankment and the flow rate there was no critical water head generating during the test.
P
o
rew
a
te
rp
re
s
s
u
re[k
P
a
]
Pore water pressure measurements during flooding at 20 cm
10
Positive pore water pressure
0
Negative pore water pressure
-10
10/19/08
10/23/08
10/27/08
10/31/08
11/4/08
11/8/08
Elapsed time [s]
Figure 3
Measurements of pore water pressure during flooding at 20 cm depth.
Pore water pressure measurements during flooding at 40 cm
P
o
rew
a
te
rp
re
s
s
u
re[k
P
a
]
10
Positive pore water pressure
0
Negative pore water pressure
-10
10/19/08
10/23/08
10/27/08
10/31/08
11/4/08
11/8/08
Elapsed time [s]
Figure 4
Measurements of pore water pressure during flooding at 40 cm depth.
The reduction in mass permeability caused by fissuring was observed immediately as a
piping appeared soon at the bottom of the embankment when the flooding was at its highest level but not overtopping. The flow of water through the fissured soil in the laboratory
didn’t lead to slope instability but more as a superficial erosive process rather than a real
breach initiation. There appeared to be no major disruptive forces and specifically differences in hydrostatic pressure across individual soil blocks that would have led to rapid
failure because the hydraulic gradient through a heavily-fissured embankment has a relatively steady slope. There was heavy seepage and as such a significant loss of the waterretaining function of the flood embankment.
It was observed that during flooding the flowing water progressively eroded the embankment to a depth where the clay was intact. This part remained resistant as it was resilient to
the water and still cohesive. A good example of comparative resistance of intact and fissured clay similar to this observed during model test is shown on Figure 5.
01/04/2009
59
Figure 5
Exposed undamaged clay core (Easington, E. Yorkshire), (after Marsland14)
It can be seen from the picture, that the intact clay core was exposed after overtopping
when the top (corresponding to shrinkage) layer was washed away. The extent of the deterioration seen on the picture above is strictly related to the soil properties which are explained in the next paragraph.
8.1.2
SOIL PROPERTIES
Another important issue which must be highlighted is the properties of the soil used in
laboratory studies and classified as stiff clay with the undrained shear strength varied from
121 kN/m2 to 152 kN/m2 and dry density 1.948 Mg/m3. These were different to the one investigated after the North Sea Flood, (Cooling and Marsland5,1953). The results of the soil
surveys along the Thames estuary in 1953 revealed the variable nature of marsh soils,
peats, peaty clays and silty clays overlaying the sand and gravel stratum. The dry density of
the marsh clay which was used for the embankments construction was found to be 1.6
Mg/m3 with the undrained shear strength varied between 9.576 kN/m2 to 14.36 kN/m2 and
which classifies the soil as very soft clay. Hence, a similar failure mode, such as a slipping
surface or the removal of the clay blocks by dragging forces, was impossible to replicate as
the undrained shear strength of the soil used in the laboratory model was greater then the
one measured during the Cooling’s5 investigation.
Marsh clays led to more catastrophic failures than the one observed during the laboratory
test. Such failures were due to the poor geomechanical soil properties as well as to the softening ratio. Soft soils are usually much more subjected to softening when they are in contact with water. Especially the deteriorated structures, investigated by Cooling and
Marsland5, which were affected by highly fissured upper layer and were more sensitive to
any changes in moisture content and pore water pressures. As a consequence, their
strength was reduced to the minimum.
The soil properties analysis such as plasticity index and linear shrinkage can give some indications about the extent of cracking which can occur in clayey embankments as the
cracking is directly linked with the shrinkage of the soil. However, this cannot be considered as determinant for the embankments assessment as the cracking is dependent on many
factors, such as type of the soil, weather conditions, size of the structure, water regime under and around embankment and vegetation.
01/04/2009
60
8.1.3
GEOFLOOD MODEL
Concerning the Data and geotechnical models from GeoFlood for use in PAMS, a geotechnical
stability model couldn’t be produced because the laboratory tests failed to find any failure
mode strongly related to fine fissuring and to the one proposed in UR119 so there was
nothing specific to model.
8.1.4
MONITORING AND MEASURING THE EXTEND OF FINE FISSURING IN THE FIELD
The monitoring and measuring of the extent of fine fissuring in the field using visual inspection analysis is of course the first step to identify possible weak sections of embankments. This needs to be carefully done and the difference between superficial and desiccation fissuring has to be recognised. Sometimes tension cracking can also be wrongly
assessed as a desiccation cracking. This type of cracking is mostly depending on the localised settlement of the embankment and also cannot be ignored.
The model embankment was subjected to sets of conditions including an intense period of
desiccation that resulted in large widely spaced cracks that could not close.
8.1.5 MONITORING AND MEASURING THE DEPTH OF CRACKING USING SUCTION AND
MOISTURE CONTENT PROBES
The direct measurements of soil suction can give an indication about the depth of the
cracking if the crack communicates directly with the sensor with an automatic cavitation to
the atmospheric pressure. The “live suction” monitored on the upper layer of the embankment can give some information on the extent of the cracking because of its fast changes
and fluctuations in comparison to the middle and bottom layers measurements of the
model.
Figure 6 shows the variations of the soil suction during the test in the upper layer of the
embankment model. It can be clearly seen that during the first stage of the drying experiment, just after the rainfall (red bars on the graph), the suction sharply dropped down to
start increasing again after the rainfall stopped. Obviously, as the soil was becoming drier,
the suction increased and the decrease was more accentuated. Physically it can be explained as an “easy access” for the percolating water into the fractured structure which was
quickly picked up by the measuring sensor.
Suction variation at 20 cm
100
300
Suction [kPa]
80
200
60
40
100
Precipitation [mm/m2]
Precipitation
Suction
20
0
0
6/21/08
7/11/08
7/31/08
8/20/08
9/9/08
9/29/08
Elapsed time [s]
Figure 6
Soil suction variations in the upper layer of the embankment model.
For the sensor installed at 40 cm depth shown on Figure 7 it can be seen that this sharpen
decrease in suction started appearing after four weeks which means that the progressive
01/04/2009
61
vertical cracking has reached this depth to allow the water to percolate deeper and spread
around.
Suction variation at 40 cm
100
300
Suction [kPa]
80
200
60
40
100
Precipitation [mm/m2]
Precipitation
Suction
20
0
0
6/21/08
7/11/08
7/31/08
8/20/08
9/9/08
9/29/08
10/19/08
Elapsed time [s]
Figure 7
Soil suction variations at 40 cm depth of the embankment model.
Investigation of the soil suction profiles (Figure 8) recorded before and after the rainfall
have shown that there is a similarity in suction variations in desiccated layer in agreement
with the interpretation given by Jennings12 (1961) (Figure 9).
Soil suction profiles - before and after rainfall
0
Before
rainfall
After rainfall
10 min
0.1
0.2
0.3
0.4
0.5
0.6
Depth [m]
Depth of cracking
0.7
0.8
0.9
1
-70
-60
-50
-40
-30
-20
-10
0
Suction [kPa]
Figure 8
Soil suction profiles before and after the rainfall indicating depth of cracking.
01/04/2009
62
Figure 9
Soil suction in desiccated clay under seasonal variations (after Jennings12 1961).
It can be clearly seen on Fig 8.9 that there is an evident point where the two suction profiles, before and after the rainfall, merge at the boundary between intact and desiccated
clay. This can give an indication about the depth of cracking.
The measurements of the moisture content are slightly different to those of the suction presented above. Figure 10 shows the response of the upper sensor to the moisture changes
due to the water entering the cracks. The lower sensor (Figure 11) wasn’t able to detect
any significant changes in the moisture content in agreement with the soil suction profiles
presented before.
Moisture Content variation at 20 cm
M oisture C ontent [% ]
22
Moisture Content
Precipitation
200
20
18
100
16
14
6/1/08
P recipitation [m m /m 2 ]
300
24
0
7/11/08
8/20/08
9/29/08
11/8/08
Elapsed time [s]
Figure 10
Moisture content variation in the upper layer of the embankment model.
01/04/2009
63
Moisture Content variation at 40 cm
300
Moisture Content
Precipitation
22
200
20
18
100
16
14
P re c ip ita tio n [m m /m 2 ]
M o istu re C o n te n t [% ]
24
0
6/1/08
7/11/08
8/20/08
9/29/08
11/8/08
Elapsed time [s]
Figure 11
Moisture content variation at 40 cm depth of the embankment model.
It can be concluded from the results presented above that suction probes are more sensitive
to the creation of fissured zones and are better than moisture probes to detect the changes
in the soil structure.
Another remark is that these measurements are local which highlight the need of a distributed system of suction monitoring on the field. A good use of this kind of measurements
relies on visual observations.
The interpretation given above is the author’s free interpretation, and needs to be proven
on a real embankment.
8.1.6
DATABASE OF FISSURING OF EMBANKMENTS
Deliverables of extended database of fissuring of embankments couldn’t be completed by
authors as there was very little information provided by EA about potentially fissured embankments and the network of coastal and river defences in Britain is to extensive to visit
all of them within 3 years project. Thus, authors relied on the existing and available
knowledge and were trying to cooperate with the locally based EA offices. A requesting
letter sent in 2007 around the EA offices resulted in only few answers identifying some
places. Therefore, since the project started in 2006, only 8 embankments were visited
around England and are presented on Figure 12. These places were indicated by EA as the
highly fissured embankments.
01/04/2009
64
Lancaster - 2007
York - 2007
Gwyneed - 2007
Thorngumbald - 2006
Gwyneed - 2007
Windsor - 2007
Bridport - 2007
Bridport - 2007
Figure 12
Investigated places of fissured embankments between 2006 and 2007.
The most of visited sites had only superficial cracking which was found only in the locations exposed to the sun and not covered by the grass (see Figure 13 a and b).
(a)
(b)
Figure 13
Superficial cracking found on embankments in 2007.
During the one of the investigations shown on Figure 14 it was found that the cracks identified by the EA engineers were due to the soil washed away from between the pebbles
used for the restoration after breaching of existing embankment in 1976 and then after
1997 overtopping when the embankment was topped up by the pebbles mixed with soil to
increase its height.
01/04/2009
65
Figure 14
Pebbles mixed with top soil with some visible loose material.
In many cases, cracking of the embankments is neglected by the engineers as it is difficult
to visually observe it from the surface covered by rough grass or vegetation. Also some
other potential and dangerous symptoms of embankments instability can be ignored i.e.:
tension cracking along the slope due to the settlement or bulges on the landward side due
to the increase of pore water pressure in underlying strata. As the first example can be imperceptible when the embankment is well covered by grass and vegetation, the second example can became ignored and interpreted as a naturally created relief.
Adequate EA departments should pay more attention to the training of the specialists who
are responsible for the condition assessment and somehow for human lives. To conclude
this point it is strongly recommended to run high quality and intensive training course,
where experienced engineers will give an overview about different conditions of soil structures as well as train the EA personnel how to assess the structure. Extensive Magnetic
Conductivity survey and Resistivity arrays survey for a more detailed assessment can be a
useful tool in terms of cracking and anomalies detection along the structure of the embankment.
8.1.7
CHARACTERISTICS OF DESICCATION FISSURING
Considering the results presented in UR119 and UR14 it was decided to accept some of the
comments made in UR119 about characteristics of desiccation fissuring. Points presented
in the section about user focussed outcomes, have been revised and are listed as follows.
• Desiccation fissuring can occur after a relatively short period of time within 2 years
of construction.
• If the embankment is constructed during hot summers and not covered by
polythene sheet, fissuring can occur after few hours during construction.
• Desiccation fissuring is more likely to occur in areas of poor grass cover but not
exclusively.
• Desiccation fissuring typically extends to a depth of 60 cm below the surface (crest
or side slopes) but can penetrate to a depth of 100 cm.
• Desiccation fissures generally propagate perpendicular from the drying surface
(crest surface or side slopes) and can bifurcate into lateral fissuring at a depth of 30
01/04/2009
66
cm that can results in an orthogonal network of desiccation fissures. The
orthogonal network of fissures would allow lateral internal seepage beneath the
surface of the embankment, which according to infiltrometer tests data results in a
permeability similar to coarse sand or gravel.
• It should be pointed out that the desiccation cracks observed in the embankment
model were wider and coarse that those previously observed at or near the surface
of flood embankments (see UR119) with the exception of one major crack observed
in 2006 during studies at Thorngumbald. This can be due to the lack of vegetation.
Observations made by Black and Croney2 (1958) in their large studies shows that
the increase in negative pore water pressure during the summer is mainly due to
transpiration from the vegetation. Where the vegetation had been removed, or
where it is naturally sparse, the seasonal change in pore-water pressure is small.
This shows that evaporation from the surface plays a limited part in drying the soil
during the summer.
• Even so the model test provided a valuable insight into the onset and development
of desiccation cracks for an embankment, which could be used in future to inform
researchers for larger scale embankment studies such as those carried out in the
USA (via HR Walllingford).
• Observation made during model tests have shown that desiccation cracking started
appearing after 4 hours of drying at a relatively low suction (around 5 kPa). Even,
after one week of drying, when the embankment was extensionally cracked, the
suction was oscillating around 10 kPa.
8.1.8
USE OF GEOPHYSICS IN CRACKING DETECTION
This experimental study has shown that a miniature geo-electrical method using resistivity
arrays can be used as non-invasive method for the detection of desiccation cracks. All the
major vertical cracks, which have been visually observed at the surface of the clay model,
have been recorded by the resistivity equipment and displayed using the 2 dimensional
contour model. Despite the limitations of the method explained before, the vertical cracking network detected by the miniature resistivity arrays has been identified and validated
by visual observations.
Assumptions were made about the detection of horizontal cracks that could be hampered
by the insulating property of air related to crack continuity. It is also very important to remember that electrical resistivity images are the outcome of data processing (i.e. they are
based on apparent resistivity values) and for this reason they must not be interpreted as a
direct representation of the field situation, but rather as a guide for qualitative estimation
of the electrical resistivity distribution in the soil model. The limitation of the method was
the geometry of the electrodes and boundary effects that will be necessary to investigate
further to fully map the structure of the cracks also along the horizontal pattern in the future experiments on the scaled embankment built in our laboratory.
9 POSSIBLE NOVEL SOLUTIONS
Despite the fact that some of the remedial measures have been proposed in UR119, it was
decided to extend and summarize possible novel solutions in the final UR14 and to add
some innovative techniques which arose during laboratory studies.
01/04/2009
67
To limit, prevent and cure fissuring, revised remedial novel solutions were elaborated,
and are listed below:
•
Injections of polymers – once the cracking of the embankment is identified, it is
considered to inject the polymers (i.e. this supplied by URETEC) inside
embankment body, to decrease the permeability of the structure, prevent any
further cracking and to increase the strength of the structure. However, only “soft”
injection has to be considered in case of the potential destruction of the
construction during injection. It is proposed to use the polymers which can
penetrate inside embankment body, spread inside cracking network and expand to
fill in the small and the big gaps.
•
Spraying of the lime/cement suspension or gels on the embankment surface – the
purpose is the same as above. It is just a different technique which may be use for
curing the embankment cracking.
•
Rubblised layer mixing with polymers, lime or cement – when the depth of the
cracking is not bigger than available for mixing heavy machines, it can be
considered to mix the upper layer with the polymers, lime or cement to decrease
permeability of the cracked layer.
•
Use of the conductive gels or tracers – to improve the scanning technique to be
able to detect whole network of cracks as well as horizontal cracks. The grout of
gel used would be even better if it is conductive and hence allow detection using
Geophyscis techniques. The rehology of the gel must be suitable to cure between
20 minutes and 1 hour, take the shape of the cracks and not disperse into the
surrounding media too much.
•
Berm – the construction of a berm on the landward side of an embankment is only
an effective measure if the slip circle is contained within the embankment or
extends only a shallow distance below the embankment. The construction of a
berm may significantly reduce the risk of failures associated with water filled wide
fissures. The design of a berm should be carried out by an experienced
geotechnical engineer. The height of a berm can be 40 % to 50% of the
embankment height.
•
Replacement with clay or clayey – this method is based on the removing of
deteriorated layer and replacing it with the clay or clayey silt. It is suggested to
remove the top soil along with fissured clay to a min 50 cm below the top of the
intact clay. During dry seasons, exposed soil must be covered by waterproof sheet
in order to protect it against drying. Material with the proper consistency can be
then placed and compacted in layers. At least 60 cm of top soil is required on the
top of restored structure. Top soil has to be placed as soon as the compaction
works are finished. The best would be to carry out these works in 10 m sections.
•
Replacement with Hoggin – the procedure consist in removing all the top soil from
the fissured area, then digging out the fissured zone of the embankment, extending
the excavation work to a minimum of 200 mm below the zone. Then the removed
01/04/2009
68
material is replaced by hoggin. In order to place and compact the hoggin, it is
suggested to follow the Specifications for Highway Works11. The removed fissured
material can be used to form a berm on the landward face.
•
Granular crest – if granular material is poured into clay, this reduces the plasticity
of the clay and therefore its tendency to fissure. The treatment consists of rolling
and harrowing a thick layer of graded granular material into the crest of the
embankment. Layers are added until a dense granular surface is produced. Clay
below the treated layer will still be subject to fissuring, but there will be little or no
wide fissures at the surface.
•
Geotextiles and Geogrids – fissuring occurs when the negative water pressure
induced by desiccation exceeds the strength of the material. Under normal
circumstances the negative porewater pressure will not exceed the strength of a
geotextile or geogrid since the selected geotextile or geogrid depends on the system
adopted and on the dimensions of the embankment. Wide fissures are not able to
transfer across the geotextile or geogrid and their extent is therefore limited. On the
contrary, fine fissures are unaffected by the presence of a geotextile or geogrid and
are likely to occur to the same extent.
Before considering one of the proposed possible novel solution, the best available techniques for cracking detection must be considered to localise the potential cracking inside
embankment body and to specify how fissuring has influenced the structure of the embankment.
From the experience, which has been investigated during this project, the best technique is
non-invasive geophysical scanning which is able to give very quick view of the embankment body and localize the extend and the depth of the cracking.
10 FURTHER STUDIES
10.1
LABORATORY STUDIES
In order to understand efficiency of one of the proposed remedial novel solutions,
macro scale tests are planned on fissured clay model. In particular the model tests would
aim to investigate the following issues:
•
•
•
•
•
•
•
How an injected polymer fills the gaps in the cracked samples?
Which application is the best for cracked samples?
How injection may affect the samples (observations of any destruction during
injection)?
Investigation of use of any other available chemicals/suspensions?
How, to improve horizontal cracking detection?
How does the method increase/decrease the strength of the cracked samples?
How does the method affects the mineralogy of the soil?
01/04/2009
69
10.2
FIELD STUDIES
Once proposed remedial solution is satisfactory tested in the laboratory on a scaled
model it can be applied to the field based cracked embankment. In particular the field tests
would aim to investigate the following issues:
•
•
•
•
Investigation of the available method on the naturally cracked embankment?
Adaptation of the method for different soils.
Excavations of the tested sites in order to investigate the efficiency of the used
methods.
How does remediation technique can influence water regime inside the structure, if
so?
01/04/2009
70
11 CONCLUSION
11.1
MEASURING THE EXTENT OF DESICCATION CRACKING
Laboratory results confirm that desiccation fissuring can transform homogenous material into interconnected network of cracks. This can occur to a depth of typically 40-60 cm
within the outer surface of a flood embankment and can be initiated within relatively short
period after construction. This is directly dependent on the soil density, weather conditions
and water regime inside the structure. It has been proved during laboratory studies that
horizontal cracking forms horizontal discontinuity which is perpendicular to the embankment shape. The embankment model that has been tested in the laboratory wasn’t covered
by the grass and the nature of cracking can be different when there is the top soil and vegetation. In the soft clays cracking process can occur faster than in the stiff clays, as the soft
soil is prone to loose its strength and retaining water. Soil suction probe was found to be
the useful tool in monitoring desiccation fissuring, but it is a local measurement and needs
extended network of sensors to cover the embankment.
As a proof of concept, the laboratory model data obtained in this study must now be
used with a geotechnical stability model taking into account the 3 zones of different suction profiles and a head of water corresponding to the height of the embankment.
11.2
GEOPHYSICAL MEASUREMENTS OF THE EXTENT OF FINE
FISSURING
The potential benefits provided by restivity surveys needs to be fully acknowledged. It
proved to be a useful tool for detecting the presence of cracks and the extent of cracking
below the surface. This merits further research in the field. An ICE R&D and Scottish Executive Enabling Fund project is already underway.
11.3
FAILURE MECHANISM OF THE EMBANKMENT MODEL
The failure mechanism was different from that observed by Cooling and Marsland5.
The soil used in the model test was much more strengthened than the one investigated by
Cooling and Marsland15, and later by Dyer et al9. This led to the resistance of the clay
blocks, which were still in contact with the intact clay.
The evidence from the laboratory tests is that, because the hydraulic gradient through a
heavily-fissured embankment has a relatively steady slope, there appeared to be no major
disruptive forces (and specifically differences in hydrostatic pressure) across individual
soil blocks that would have led to rapid failure. Obviously there was heavy seepage and as
such a significant loss of the water-retaining function of the flood embankment. The hydro-dynamic forces need to be better understood and need some further investigation.
In general some of the limitations of the test can be recognised (i.e.: larger desiccation
cracking due to absence of vegetation, soil/concrete interface on the sides). Positive outcomes need to be highlighted (detection of cracks by restivity, accurate soil suction/moisture content profiles, almost unique observations of a the breaching of desiccated
01/04/2009
71
clay embankment with numerous: –time, rate, physical features, flow rate, etc). In future
other researchers can make use of these results to model breach growth.
Even so the results clearly demonstrate that breaching of a flood embankment constructed from clay soil is controlled by the weakness in the embankment construction
(such as highly permeable underlying strata) or deterioration (such as desiccation) and not
the classical Bishop’s slope failure. The classical slope instability is rarely recorded for
clay flood embankments except during the construction phase when the underlying soil
may be too weak to initially support the embankment (as it occurred for the new embankment at Thorngumbald). In the future fault trees and corresponding fragility curves need to
focus on potential weakness that could lead to breach initiation and avoid the misconception that classical slope failure is a primary failure mode during a flood event.
These findings as part of the ongoing research into flood embankments since the mid
1990’s should also feed into an improved asset condition survey methodology. Flood engineers should survey with a geotechnical and geophysical check list (soil properties, suction and moisture sensing, resistivity/conductivity) and attribute meaningful ratings that
really indicate the contribution to a likely failure taking place, not only visually. This result
alone justifies the research project.
01/04/2009
72
12 REFERENCES
1 Barker, R.D. (1997) Electrical imaging and its application in engineering
applications. Modern physics in Engineering Geology, Geological Society
Engineering Special Publication, 12, pp. 37-43.
2 Black, W., P., M., and Croney, D., (1958), Field Studies of the Movement of Soil
Moisture. Road Research Technical Report No. 41, London
3 British Standards (1986), BS 8004:1986, Code of practice for foundations., p. 24
4 British Standards (1999), BS 5930:1999, Code of practice for site investigations.,
p.114
5 Cooling, L.F and Marsland, A. (1954) Soil Mechanics of Failures in the Sea Defence
Banks of Essex and Kent. ICE Conference on the North Sea Floods of 31 January / 1
February 1953.
6 Croney, D., Coleman, J.D. and Black, W.P. (1958) Movement and distribution of
water in soil in relation to highway design and performance. Highway Research
Board Special Report, Transport Research Board, Issue: 40, pp. 226-252
7 Dane, J.H. and Topp, G.C. (2002) Methods of soil analysis. Part 4, Soil Science
Society of America, Madison, Wis.
8 Depountis, N., Harris, C., Davies, M.C.R. (1999). The application of miniaturised
electrical imaging in scaled centrifuge modelling of pollution plume migration.
Proceedings 2nd BGS International Geoenvironmental Engineering Conference,
London, pp. 214-221.
9 Dyer, M., Utili, S., and Zielinski, M. (2007) Influence of the desiccation fine
fissuring on the stability of flood embankments. UFMO UR11, FRMRC Research
Report
10 Fredlund, G.D. (2006) Unsaturated Soil Mechanics in Engineering Practice. Journal
of Geotechnical and Geoenvironmental Engineering., ASCE, pp. 286-321.
11 Highway Agency (2006) Specification for highway works. Earthworks. Series 600
12 Jennings, J. E. (1961) A Revised Effective Stress Law for use in the Prediction of the
Behaviour of Unsaturated Soils. Pore Pressure and Suction in Soils. Conference
organized by the British National Society of the International Society of Soil
Mechanics and Foundation Engineering at the Institution of Civil Engineers held on
March 30th and 31st, 1960, pp. 26-30.
13 Marsland, A. (1968) The shrinkage and fissuring of clay in flood banks. Building
Research Station, Internal report No. 39/68.
14 Marsland, A. (1957) The design and construction of earthen flood banks. Journal
Institution of Water Engineers, Volume 11, No. 3, pp. 236-258.
15 Marsland, A. and Cooling L.F. (1958) Tests on Full Scale Clay Flood Bank to Study
Seepage and the Effects of Overtopping. Building Research Station, Internal report
No. C562.
16 Marsland, A. and Randolph M.F. (1978) A study of the variations and effects of pore
water pressures in the pervious strata at Crayford Marshes. Géotechnique, 28, No.
4, pp. 435-464.
17 Nguyen, F., et al. (2005) Image processing of 2D resistivity data for imaging faults.
Journal of Applied Geophysics, Volume 57, Issue 4, pp. 260-277.
18 Peck, R.B. (1969) Advantages and limitations of the observational method in applied
soil mechanics. Géotechnique, 19, No. 2, pp. 171-187.
01/04/2009
73
19 Ridley, A., Brady, K.C. and Vaughan, P.R. (2003) Field measurement of pore water
pressures. TRL Report, Highways Agency.
20 Samouëlian, A., et al., (2003) Electrical resistivity imaging for detecting soil
cracking at the centimetric scale. Soil Science Society of America Journal ,67, (5),
pp. 1319–1326.
21 Sentenac, P., Zielinski, (2009) Clay fine fissuring monitoring using miniature geoelectrical resistivity arrays. Journal of Environmental Geology.
22 Smethurst, J.A., Clarke, D. and Powrie, W. (2006) Seasonal changes in pore water
pressure in a grass-covered cut slope in London Clay. Géotechnique, 56, No. 8, pp.
523-537.
23 Take, W.A., and Bolton, M.D. 2002. A new device for the measurement of negative
pore water pressures in centrifuge models. Int. Conf. Physical Modelling
Geotechnics, St. John's. Rotterdam: Balkema, pp.89-94
24 Tabbagh, J., Samouëlian, A. and Cousin I. (2007) Numerical modelling of direct
current electrical resistivity for the characterisation of cracks in soils. Journal of
Applied Geophysics, Volume 62, Issue 4, August 2007, pp. 313-323
25 UMS Manual.
26 Van, M.A., Koelewijn, A.R. and Barends, F.B.J. (2005) Uplift Phenomenon: Model,
Validation, and Design. International Journal of Geomechanics, Volume 5, No. 2,
June 2005, pp. 98-106.
01/04/2009
74
13 ANNEX
13.1
FIRST DAY OF DESICCATION PHASE
Figure 13.1
Downstream side of the embankment.
Figure 13.2
Crest of the embankment.
01/04/2009
75
Figure 13.3
13.2
Upstream side of the embankment.
SECOND DAY OF DESICCATION PHASE
Figure 13.4
01/04/2009
76
Figure 13.5
Figure 13.6
13.3
THIRD DAY OF DESICCATION PHASE
Figure 13.7
01/04/2009
77
Figure 13.8
Figure 13.9
13.4
FOURTH DAY OF DESICCATION PHASE
Figure 13.10
01/04/2009
78
Figure 13.11
Figure 13.12
13.5
FIFTH DAY OF DESICCATION PHASE
Figure 13.13
01/04/2009
79
Figure 13.14
Figure 13.15
Figure 13.16
01/04/2009
80
13.6
SIXTH DAY OF DESICCATION PHASE
Figure 13.17
Figure 13.18
Figure 13.19
01/04/2009
81
13.7
SEVENTH DAY OF DESICCATION PHASE
Figure 13.20
Figure 13.21
Figure 13.22
01/04/2009
82
Figure 13.23
13.8
RESISTIVITY SCANS
Figure 13.24
Figure 13.25
Inversed resistivity map taken before experiment .
Inversed resistivity map from the second day.
01/04/2009
83
Figure 133.26
Inversed resistivity map from the third day.
Figure 133.27
Inversed resistivity map from the fourth day.
Figure 13.28
Inversed resistivity map from the fifth day.
Figure 13.29
Inversed resistivity map from the sixth day.
Figure 13.30
Inversed resistivity map from the seventh day.
01/04/2009
84

UR14_2009 Final version_nso - School of the Built Environment

Transcription

Similar documents

Angiosperm Latin Name: Cyperus involucratus Common Name

Technical Sheet - Microbes Biosciences

Angiosperm Latin Name: Euphorbia ingens Common Name

rattler pre-harvest desiccation: cereals and oilseed rape

Project Profile: Horasan Kurtozu Sarıkamış Highway

Prologue

Revel Casino News 2013

Angiosperm Latin Name: Callisia fragrans Common Name: Basket

AZOMITE granular - Down To Earth Fertilizer

Pro-Grow