ORIGINAL PAPER
Significance of Geological Parameters for Predicting Water Inflow in Hard Rock Tunnels
K.H.Holmøy •B.Niln
Received:29June 2012/Accepted:31January 2013ÓSpringer-Verlag Wien 2013
Abstract One of the most challenging aspects of tun-nelling is prognostication of water inflows.More reliable prediction of groundwater inflow may give considerable economical saving for future tunnel projects and may also prevent damage of environment and installations on the surface.This paper is discussing the significance of eight hypothes regarding geological parameters for predicting water inflow in tunnels.The respective hypothes have been tested as part of a recent rearch project in Norway.Six Norwegian tunnels with different geological conditions were lected for the rearch;the Romeriksporten,Frøya,T-baneringen,Lunner,Skaugum,and Storsand tunnels.Bad on detailed study of the tunnels,the hypothes are tested by comparing water inflow with geological param-eters and factors such as Q value,faulting,rock stress orientation,rock cover,thickness of permeable soil or depth of lake/a above the tunnel,rock type,and width of weakness zones.It is found that four out of the eight tested hypothes are supported,two have low to medium support and two are not supported.On
e unexpected result is that for the tunnels covered by this study,the water inflow was found to increa with rock cover.
Keywords Water inflow ÁHard rock ÁTunnelling ÁEngineering geology
1Introduction and Background
Groundwater inflow may cau veral problems for underground excavations,such as:•
Risk of lowering the groundwater level,which may cau ttlements and damage to buildings and instal-lations on the surface.
•
Drainage of surface areas and lakes,which may damage vegetation and recreational areas.Sometimes the nat-ural discharge conditions can be disturbed,causing dry out of springs and streams.
•
Difficulties in estimating costs due to high degree of uncertainty connected to location and quantity of groundwater inflow.
•Reduced rock mass stability.
•
Expensive and time-consuming grouting (pre-and post-excavation).
Reliable prediction of groundwater inflow may give considerable saving in construction costs by optimizing construction techniques,including pregrouting.Also,a better understanding and more reliable prognostication of groundwater inflow may improve the possibility of pre-venting damage to the environment and nearby installations.
2Tunnel Projects with Problems due to Water Inflow:Some Examples Several underground projects around the world have experienced major delays,large unforeen costs,and negative attention from media,politicians and the public due to large water inflows.The partly excavated railway
tunnel at Hallandsa
˚s in southern Sweden (Swedish Rail Administration 2008)is one project with a long and
K.H.Holmøy (&)
SINTEF Building and Infrastructure,P.O.Box 4760Sluppen,7465Trondheim,Norway
e-mail:Kristin.H.
B.Niln
Department of Geology and Mineral Resources Engineering,Norwegian University of Science and Technology,Sem Sælands veg 1,7491Trondheim,Norway
Rock Mech Rock Eng
DOI 10.1007/s00603-013-0384-9
troublesome history due to difficulties caud by water inflows.The Hallandsa˚s project consists of two8.6km long single-track tunnels through a structural horst formation, with challenging geological conditions and highly varying rock mass quality.The main rock types are Precambrian gneiss and amphibolite.Tunnelling started in1992,but in 1997the project was stopped due to environmental pollu-tion of nearby streams caud by chemical grouting that was ud in the Hallandsa˚s tunnel.In
2003,excavation was resumed,and in January2013only9%of the main tunnel remained to be excavated(Trafikverket2013).
In Iceland,veral tunnels in basalt have experienced vere difficulties due to large water inflow.In the north-west of the country a9km long road tunnel was built between1991and1996.The bedrock was of late Tertiary age,and the total thickness of rock cover300–500m. During excavation water inflow at the face of up to3,000l/s was encountered(Steingrı´msson and Hardarson2000). The extremely large water inflow was encountered in connection with a fault and basalt dyke.The excavation was stopped for half a year before it was possible to pro-ceed.At that point,the water inflow had stabilized at 1,000l/s,and it decread further to400l/s.The tunnel profile was enlarged to make room for water supply pipes, and I´safjo¨rdur municipality now us this as its main water supply(Hardarson and Haraldson1998).The groundwater table was lowered between30and100m.
Another example is the Ka´rahnju´kar hydro power pro-ject,also located in Iceland,which experienced problems with difficult geological conditions and large inflow of water in the40km long headrace tunnel(Peturson2007). Excavation at Ka´rahnju´kar started in2003and was com-pleted in December2006.For excavation three tunnel boring machines(TBMs)were ud.One of the TBM ctions had particularly large problems due to extremely large water inflow and conventional drill an
d blast tun-nelling had to be ud on this ction instead of TBM.
Outside the Nordic countries,veral projects have experienced extensive problems due to groundwater inflow. The34.6km long Løtschberg ba tunnel in Switzerland (BLS AlpTransit2008),a two-tube single-track rail tunnel project is one example.The rock cover here was up to 2,000m,and the bedrock consisted of sandstone,marble, limestone,flysch,gneiss,granite,and amphibolite.Large water inflow with water pressure of up to120bar was encountered in dimentary bedrock,and extensive pre-grouting had to be carried out.In a ction with limestone, water inflow lowered the groundwater table below the vil-lage St.German,resulting in drying of the source for drinking water and subsidence in the centre of St.German.
In Italy,nine railway tunnels with total length of73km were excavated between Bologna and Florence between 1996and2005.The tunnels run through silicoclastic turbidites(marls,limestone,sandstone,and shale).Water drainage into the tunnels caud disturbance of natural springs and streams,a total of31springs were impacted by the drainage(Gargini et al.2008).Experience from the railway tunnels showed that most of the impacted springs were located not more than150–200m from the tunnel axis.
One relevant example from Asia is the12.9km long Hsuehshan road tunnel in Taiwan,which is perhaps one of the most difficult and time-consuming tunnel projects in the world(Wallis2006).Tunnelling started in July1991 and thefinal breakthrough was in April2004.The slightly metamorphod Tertiary rocks of dimentary origin here were highly folded and jointed with veral major faults. Both conventional drill and blast and TBM were ud for excavation.The TBM were forced to stop13times due to collap of the tunnel face and extremely large water inflow of up to5,000l/min.大班户外活动
From South America,one relevant example is the Alf-alfal hydropower project in Chile,where challenging geological conditions,high stress(stress spalling)and water inflow with pressure of up to100bar were encoun-tered(Buen et al.1994).In granitic rock mass at Alfalfal,a concentrated water inflow of about6,000l/min was encountered.A bypass tunnel had to be excavated,and the abandoned tunnel was ud as drainage to abate the high ground water pressure.Another relevant project in Chile is Andina,where two tunnels were excavated in connection with treatment of waste water from the copper mine (Stefanusn2000).In granitic rock mass of the Cretaceous Andes mountain range high stress combined with major joints gave large water inflow,and in February1997a major cave-in combined with water inflow of up to600l/s stopped the tunnelling.Also in this ca a bypass tunnel was ud as a solution.
Finally,the Arrowhead water supply tunnels in southern California,USA may be mentioned among projects with considerable challenges and delays due to large water inflow(Burke2004).The two Arrowhead tunnels are located in the San Bernardino Mountains,which are characterized by late Tertiary faulting and folding.The bedrock mainly consists of granite and gneiss.In early 2000,after excavating2.4km,large water inflow with pressures up to17bar was encountered.Pregrouting was carried out,which reduced the water inflow to946l/min. In the fall of2000,it was decided to terminate the contract, and engineering studies were initiated to determine the best excavation method,including a plan for pregrouting for the remaining excavation.New contractors started the detailed planning in January2002,and two new TBM’s arrived on site in July2003for completion of the project.
As illustrated by the cas described above water inflow is definitely one of the biggest challenges for the tunnel industry.
K.H.Holmøy,B.Niln
3Hypothes and Selected Tunnels
As part of the major Norwegian rearch programme ‘‘Tunnels for the citizens’’(Lindstrøm and Kvee
n2005),a PhD study was carried out by thefirst author of this paper (Holmøy2008)in order to increa the knowledge of factors affecting water inflow in hard rock tunnels.In this study,eight hypothes regarding significance of relevant geological factors for predicting water inflow were tested. The hypothes were lected bad on literature review and experience from tunnelling,both clearly indicating that the permeability in hard rock is controlled mainly by discontinuities.
The hypothes which were lected,and will be dis-cusd in this paper,are as follows:
1.The water inflow is smaller in rock mass with Q values
lower than0.1,than in rock mass with Q values between0.1and10.This is bad on the assumption that the central part of a weakness zone/fault normally consists of clay material,which will al the rock mass,and that the largest water inflow will occur in the damage zone marginal to the central part,where the rock mass is denly fractured but without clayfilling.
2.Water-bearing joints are oriented with an angle of
45±15relatively to nearby major faults(bad on Selmer-Oln’s1981theory).
3.Water-bearing discontinuities are sub-parallel(±30°)
with the major principal stress(r1).
4.Water inflow will decrea with increasing rock cover
due to higher gravitational stress causing closing of discontinuities.
5.A lake/a above the tunnel gives large water inflow,
due to high rervoir capacity.
6.Igneous rocks give larger water inflow than dimen-
tary and metamorphic rocks due to their brittle character.
7.Major rock type boundaries(including dimentary
layers with different compositions)give large water inflow due to incread fracturing.
8.Large weakness zones with Q values lower than0.1
give larger relative water inflow than minor weakness zones.
Bad on analys of six Norwegian tunnels,the degree of support for the hypothes has been evaluated.Sec-tions with large water inflow in all six tunnels have been analyzed,and the degree of support have been rated as‘‘no support’’,‘‘low to medium support’’and‘‘support’’.
The following six Norwegian hard rock tunnels have been lected for the analys:
•Romeriksporten(excavation completed1996).
•Frøya(excavation completed1999).•T-baneringen,stage one Ulleva˚l-Nydalen(excavation completed early2002).
•Lunner(excavation completed2002).
•Skaugum(excavation completed2004).
•Storsand(excavation completed2004).手抄报边框图片
The aim for the lection of the six tunnels was to obtain a basis for studying tunnels with different g
eological con-ditions reprentative of Norwegian projects.It should be emphasized that the analys described in this paper are bad on six Norwegian tunnels only,and the results of the study therefore should be ud with care and only for projects in hard rock of similar geological conditions.
Four of the six lected tunnels were included also in the rearch project‘‘Tunnels for the citizens’’(Lindstrøm and Kveen2005):T-baneringen,Lunner,Skaugum,and Stor-sand.For the tunnels,extraordinary site investigations were carried out,and data were easily available for rearch.The other two tunnels were lected due to their particular relevancy,and becau thefirst author of this paper was involved as engineering geologist on site at Romeriksporten as well as at the Frøya tunnel.The Rom-eriksporten tunnel is also particularly interesting for this study becau extremely large water inflow was encoun-tered.This caud considerable extra cost and time delay. For the Frøya suba tunnel,the geological conditions were very challenging and large water inflow was encountered. The locations of the lected tunnels are shown in Fig.1.
The mapping of the six tunnels was done by various geologists,and the extent and quality of the mapping thus varied.In many cas,the weakness zones were not described in detail with correct structural geological defi-nitions.Since the term‘‘weakness zone’’may be inter-preted in different way
s,it was difficult in some cas to interpret if the mapped‘‘weakness zone’’is a fault zone,a ction of weak rock mass or a fractured zone.In this paper,the term weakness zone means a zone in the rock mass where the mechanical properties are significantly poorer than in the surrounding rock mass,and the zone may be either of veral types as described by Niln and Palmstrøm(2000).In cas where a more exact description exists,this will be ud.
4Rearch Methodology
In the analys that are discusd in this paper,main emphasis has been placed onfinding potential correlations between water inflow and geological parameters.For the lected cas,either a part of the tunnel or the entire tunnel has been analyzed.Water inflow has been found bad on measuring water inflow in probe drilling holes and pregrouting rounds.Probe holes are drilled normally
玉米面做法Significance of Geological Parameters for Predicting Water Inflow
by percussion drilling for checking the rock mass quality and amount of water inflow ahead of the tunnel face.The water from the hole is collected in a bucket,and the inflow measured as liter per minute.If the measured water inflow is above a given threshold value pregrouting is carried out.In the analys,the respective tunnels have been divided into 25-m long ctions.Section length 25m was
chon becau the length of the probe drilling and pregrouting holes usually are between 21and 26m.A problem with measuring water inflow in drill holes is lack of control regarding where the water enters the drill hole,and dividing the tunnel into shorter ctions than 25m there-fore would not be correct.
Upper part of Fig.2shows a plan view of a tunnel with two pregrouting rounds of 23m length.The cond pre-grouting round starts 15m ahead of the first ,after three blasting rounds (one blasting round is typically 5m long).The grouting holes as shown in the figure are angled outwards approximately 15°,and the normal over-lap between pregrouting rounds is 6–8m.The cross ction in the lower part of Fig.2shows a typical tunnel face with 6probedrilling holes and 20pregrouting holes,including 7pregrouting holes at the tunnel face.Both probe drilling and pregrouting holes are grouted.
肺气肿吃什么药好The six tunnels of this study were excavated during a relatively long time span (1995–2005).Therefore,the grouting techniques differ considerably.The number of pregrouting holes in one pregrouting round varies from 21to 48,and the length of the holes from 18to 24m.
Bad on measurement of water inflow in probe drilling holes and pregrouting rounds,calculation of 五年级学生
water inflow into the tunnel has been done for every 25m.The water inflow was measured before grouting of the respective drill holes.Therefore,the measured inflow was high,and
is
Fig.1Locations and geology of the six lected tunnels
K.H.Holmøy,B.Niln
believed to reprent quite well the amount of water that would have flowed into the tunnel in ca no grouting was done.
The water inflows encountered in the tunnels have been carefully evaluated versus relevant geological parameters in order to check whether the hypothes are supported or not.Spreadsheets and the add-in StatPro from Palisade (Albright et al.2003)have been ud for analysis.Dia-grams showing scatterplots with correlation,trend line with mathematical function,coefficient of determination (R 2),and boxplots have been extensively ud.
Boxplots have been ud since the clearly illustrate the scattering of a data t.For StatPro this is explained by Albright et al.(2003)as follows:‘‘The right and left of the box are the third and first quartiles.Therefore,the length of the box equals the interquartile range (IQR),and the box itlf will contain 50%of the obrvations.The height of the box has no significance.The vertical line inside the
工作个性签名box indicates the location of the median.The point inside the box indicates the location of the mean.Horizontal lines are drawn from each side of the box.They extend to the most extreme obrvations that are no farther than 1.5IQRs from the box.They are uful for indicating variability and skewness.Obrvations farther than 1.5IQRs from the box are shown as individual points.If they are between 1.5IQRs and 3IQRs from the box,they are called mild out-liers and are hollow.Otherwi they are called extreme outliers and are solid.’’
In the scatterplots,the two parameters correlation and coefficient of determination both have been ud.In some scatterplots,also a trend line is given.Correlation is defined as follows (Albright et al.2003):
Corr ðX ;Y Þ¼
cov ðX ;Y Þ
stdev ðX ÞÂstdev ðY Þ
where the covariance,cov(X ,Y ),is given by (Albright et al.2003)as:
Cov ðX ;Y Þ¼P
i ¼1
ðX i À"X
ÞðY i À"Y Þn À1Standard deviation for X or Y are the square root of the variance of X and Y .
‘‘The correlation is unaffected by the units of mea-surement of the two variables,and it is always between -1and ?1.The clor it is to either of the two extremes,the clor the points in a scatterplot are to some straight line,either in the negative or positive direction.On the other hand,if the correlation is clo to 0,then the scatterplot is typically a ‘cloud’of points with no apparent relationship.However,it is also possible that the points are clo to a curve and have a correlation clo to 0.This is becau correlation is relevant only for measuring linear relation-ships’’(Albright et al.2003).
The trend line with mathematical function and the coefficient of determination R 2is drawn and calculated in Excel and StatPro.R 2measures the quality of a linear fit.The better the linear fit is,the clor R 2is to 1.The coefficient of determination is defined as (Albright et al.2003):
顺溜溜
R 2¼1ÀP e 2
i
P
ðY i ÀY Þwhere e i is the difference between the obrved value Y and the fitted value b Y
,and "Y is the respon variable’s mean.‘‘In simple linear regression,R 2is the square of the cor-relation between the respon variable and the explanatory variable’’(Albright et al.2003).
邓先芙5Geological Conditions and Water Inflows of the Selected Cas
In this Section,the lected cas reprenting six Nor-wegian hard rock tunnels will be described with main emphasis on geological conditions and water inflows encountered during excavation.5.1Romeriksporten Tunnel
The 13.8km long Romeriksporten tunnel is a part of the 42km long high speed railway link between Oslo and Gardermoen airport.Two adits were excavated;one between Bryn and Godlia and one southwest of Langvann,e Fig.3.In the ction between the two adits,large water inflow was encountered,and the contractor experienced
considerable
Fig.2Principle drawing showing a plane view of a tunnel with probe drilling and pregrouting holes (upper part of figure ),and a cross ction of the same tunnel (A–A)(bottom )
Significance of Geological Parameters for Predicting Water Inflow
