Immersed Cross Sections
Immersed Cross Sections
Immersed Cross Sections
Christian Ingerslev
U.S.A.
Pergamon
INTRODUCTION
OVERVIEW OF IMMERSED
TUNNEL P R A C T I C E S
General Steel Shell Tunnels Concrete Tunnels The Expansion Joint Concept for Concrete Tunnels The Waterproofing Membrane Concept Protection of Concrete Against Chemical Attack Longitudinal Prestressing DESIGN OF T Y PI C A L TUNNEL SECTION Interior Geometry Typical Composite Steel Shell Cross-Section Typical Concrete Box Structure Weight and Balance L O N G I T U D I N A L A R T I C U L A T I O N AND J O I N T S Steel Shell Tunnels Concrete Tunnels Shear Transfer in Intermediate Joints Intermediate Flexible Rubber Joint Design Expansion Joint Design Final Joint of Concrete Tunnels
STRUCTURAL ANALYSIS FOR CONCRETE
TUNNELS
6.1 6.2
,
Stage 1: Fabrication and Launching Stage 2: Internal Outfitting with Concrete Stage 3: Final Condition after Backfilling in Place LOADINGS Loading Combinations and Allowable Stress Increments Accidental Loads TYPICAL MATERIAL SPECIFICATIONS
8.1 8.2
.
The purpose of this report is to develop an understanding of the state-of-the-art structural design for immersed road and railway tunnels. The report covers common and specific aspects of steel and ~ncrete tunnels. It is the first time that engineers from Europe, the U.S.A., and Japan have collaborated on a report of this kind for immersed tunnels. This report is very timely in that, in recent years, a longstanding tradition has begun to fade. No longer are steel shell tunnels the only types of tunnels constructed in the United States, and concrete box tunnels the only type of tunnels used in Europe. A steel shell tunnel was a serious contender in the competition for the Great Belt Tunnel in Denmark, while the concrete box will be used for the Fort Point Channel Tunnel in Boston (Massachusetts, U.S.A.), currently under design. The history of imtaersed tunnel practice began in 1910, with the construction of a two-track railroad tunnel across the Detroit River between the United States and Canada. For the next thirty years or so, virtually all immersed tunnels were constructed in the United States. During this time a rather specific steel shell technology e m e r g e d - - a technology that has continued largely unchanged up to the present. In 1941, construction of the Maas Tunnel in Rotterdam (the Netherlands) began, marking the beginning of the use of immersed tunnels in Europe. Shallow river crossings with multiple-lane requirements led naturally to the concrete box scheme, a very different method of design and construction from the steel shell. A considerable number of tunnels were constructed in Europe with this method; and, as a result, engineers in the Netherlands, Germany, Sweden, France, Belgium, and Denmark adopted the concrete box method exclusively. This tradition persists today. Meanwhile, Japan was building tunnels of both types. The authors believe that Europe and the United States, as well as other countries, would do well to follow Japan's example. It is hoped that this ITA Report will encourage engineers to make equal use of both of these basic technologies. Each has unique advantages and limitations, depending on the site. The selection of method should be made by engineers on the basis of job site conditions--not because of a lack of understanding of either method.
2.
Tunnel Practices
2.2 Fabrication
Most concrete tunnel elements are prefabricated inside drydocks or specially constructed casting basins. Sometimes the cofferdam for the approach ramp structure is first used as a basin for the fabrication of the tunnel elements. Occasionally elements have been fabricated on semisubmersibles or launched using a marine lift. Steel shell tunnel, elements are usually fabricated in a shipyard. After the element is launched, most of the interior concrete is installed while the element is floating. The element is then placed in the trench. Steel shell tunnels have also been fabricated and partly concreted in drydocks.
2.3 Joints
All joints are gasketted and tightly closed. The immersion joints between the tunnel elements can be permanently flexible rubber compression gaskets, as is often the case for concrete tunnels. These gaskets are pre-installed at one end of each tunnel element. Intermediate joints can also be made rigid. This is done at the inside of the temporary immersion seal by welding lap plates to the shells of steel tunnels or by placing concrete for concrete tunnels. The final joint must always be made in
1. Introduction
Even though immersed tunnels are designed and constructed worldwide, special codes for immersed tunnels do not exist. Standard codes for highway structures are often used, although these codes relate to structures designed for a different structural performance and generally more severe environmental exposure than immersed tunnels. The layout and design of an immersed tunnel is very much related to construction opportunities and site conditions. The state-of-the-art described in this report covers different practices that serve the same goal: namely, to produce watertight and durable immersed tunnels. Important issues for concrete tunnels are the differences in practices for providing watertightness and durability against aggressive groundwater. These practices cannot be evaluated properly without an understanding of the influence of the longitudinal performance of the tunnel structure. These topics are discussed in some depth in this chapter, using numerical examples. Both steel and concrete tunnels require that the towing, immersing, and final conditions be analysed. For the structural design of steel shell tunnels, two additional construction stages require separate and different design analyses. These are described and the assumptions are detailed for each case.
selecting concrete tunnels; and in Japan, where both concepts are applied). Recent international immersed tenders in Europe have included alternative options for steel tunnels. The structure of a steel shell tunnel consists of relatively thin-walled composite steel and concrete rings. The steel shell provides the water barrier. The ballast concrete is placed outside the shell in pockets formed between the structural diaphragms. Concrete tunnels are monolithic structures in which most of the final weight is incorporated in the structural components. There is a wide range of cross-sectional configurations, depending on the intended use of the tunnels. In determining the ultimate shape and size of the tunnel cross-section, the designers must consider, for example, whether the tunnel is to be used for railway or motor traffic; how many tracks or lanes altogether are required; whether it will be a single tube, double tube, or multiple tube; what the ventilation requirements will be; and what construction practices will be applied.
underneath the roadway or walkway. Often, however, a special services gallery is preferred or may be required by the fire department for emergency escape. Examples of cross-sections Of concrete immersed tunnels are shown in Figure 3-2.
2.5.3 Miscellaneous
Low-point drainage sumps have to be provided within the confines of the structure. In binocular double-steelshell tunnels, the sump can be placed between the tubes. In single-steel-shell and concrete tunnels, the sumps have to be placed underneath the roadway. The presence of service galleries is helpful in positioning the pumps. Generally, for given cross-sectional requirements for vehicular space and ventilation, the concrete box section can be made shallower than the steel section. However, for the same conditions, the steel section is generally a narrower section. For a given internal air volume, the final weight will be the same in both cases
3. Watertightness
3.1 Genera/
Although watertightness is one of the primary objectives of any immersed tunnel design, the design would be deficient if the consequences of incidental small leakage were ignored. On the large perimeter surface of a tunnel element, the possibility of an undetected pinhole in a steel weld or an undetected construction imperfection of the concrete or waterproofing membrane cannot be ruled out completely. Suitable repair methods exist and must be specified in the design. The buildup of water pressure within the tunnel wall system must be avoided by provision of proper drainage into the tunnel drainage system. Seepages of this kind are very small and do not require extra drainage and sump capacity.
1,s0p_LeLo,2ss0, r I ' I
DOUBLESlIE3~ SHELL
SECOND HAMPTON ROADS 11JNNEL Ig76,
71111O
)
i
BAU.ASTCONCRE'~ II--
12400 25200
DOUBLE STEEL SHELL FORT MC -
'-I _!3050~
HENRY TUNNEL
BALTIMORE ,1984
= N~ 111120 I S11~1. SHEtJ~ SAN-F/tANSISCO
Figure 3-1.
Unless proper measures are taken, vertical full-depth cracks can occur at about 5-m intervals. Satisfactory processes have been developed to avoid these construction cracks, namely by reducing heat of hydration by using concrete with relatively low cement content and forced cooling in the lower part of the walls (see Fig. 3-3). Sometimes this is done in combination with insulation and heating of the base slabs. If cracks nevertheless are found, remedial grouting appears to be effective. Effective control of differential heat development largely depends on the heat of hydration of the concrete. This process, in turn, is a direct function of the amount of cement in the mix. Therefore, the use of typical concrete mixes with high cement factors (taken from highway structures codes) can be counterproductive in this respect. There are two basic concepts regarding control of leakage for concrete tunnels: 1. The expansion joint concept involves avoiding longitudinal stresses that can cause cracks, thereby relying on the watertightness of the uncracked concrete. 2. The waterproofing membrane concept involves enveloping the concrete tunnel element in a waterproofing membrane. Each of these concepts is discussed in more detail below.
I--
I--
24770
-I
_1
29800
t,
VI..AKE "tUNNEL 1975 VUSSINGEN
I~ 26180
108(~D
I_
30000
12680
_ I
"ll II II II
~NGSTADT TUNNEL 1968 GOTHENBURG
39700
.=
10300
SPIJKENISSE (THEN T E L N S E H RA D )
.,=
241 O0
Sometimes coating of the steel plate is also specified, but it is practically impossible to provide it without some local d e f i c i e n c i e s (e.g., at splices). N e v e r t h e l e s s , the coating, although imperfect, will reduce the direct current demand. O b v i o u s l y , a w a t e r p r o o f i n g m e m b r a n e must be fully continuous around the whole of a tunnel element. The emphasis for steel plates is on the quality of the welds. For bituminous and plastic membranes, the emphasis must be on providing the proper application; observing the proper limitations for weather conditions; and providing good overlaps over the upstand of the steel membrane. Sometimes the transition area is also clamped. A small leak in a non-adhering membrane can allow water to flow underneath the membrane to weak spots in the concrete at other locations, thereby making the source of leakage difficult to trace. Such problems may be reduced by using membranes divided into compartments, the edges of which are attached or embedded in the concrete. A combination of these two m e t h o d s - - i . e . , using both waterproofing membranes and the expansion joint m e t h o d - is not considered cost-effective or advisable.
Figure 3-3.
or to cover the cracks ~hat may be inevitable with the selected method of construction. When a waterproofing membrane is used, the concrete beneath the membrane will be monolithic, using regular vertical construction joints with continuous longitudinal reinforcement. The waterproofing membrane is made continuous over the full length of the tunnel element. The following options are used for waterproofing membranes: Steel plate below the base, with bituminous or plastic membranes on the walls and roof. Steel plate below the base and on the sides, with bituminous or plastic membrane on the roof only. Steel plate all around. Plastic m e m b r a n e s all around. All b i t u m i n o u s and most plastic m e m b r a n e s must be protected against m e c h a n i c a l damage. This protection is usually provided by a concrete layer, particularly on the roof. P r o t e c t i o n must be p e r m a n e n t l y held against the waterproofing layer. On the sides, some form of protection board may be used rather than concrete. Unfortunately, c o n c r e t e protection requires anchors that penetrate the membrane. Other necessary penetrations are for bollard and fender connections, access shafts, and certain immersion equipment. Steel plate skins applied on the sides for waterproofing are also used as stay-in-place formwork for the walls of the element. For steel membranes across the roof area, holes can be left in the plates for placing and vibrating the concrete. The holes are closed up later. This method is more complicated for a fl~t roof than for an arched roof. An alternative method is to apply the steel roof plates after concreting the roof. Steel rails are partly embedded in the top of the concrete. Steel plates will be welded to these rails and any voids will be grouted afterwards. Steel plate membranes (typically 6 mm thick) are anchored into the concrete by welded studs, usually four studs per square meter. B e c a u s e the thin plates are rather sensitive to temperature deformations, much care is required to keep the plates fiat. The bond between the steel plate skin and the c o n c r e t e should not be relied upon, notwithstanding the use of the studs, because the steel plate membrane does not provide a structural contribution to the c o n c r e t e structure. To protect the steel plate membranes against corrosion, i m p r e s s e d current c a t h o d i c p r o t e c t i o n can be installed.
3. 7 Longitudinal Prestressing
To avoid uncontrollable full-depth cracks, the longitudinal working tensile stress should not exceed the tensile strength of the concrete. Normally this requirement can be satisfied in the design of concrete tunnels. If needed, a fairly small rate of longitudinal prestress can substantially increase the allowable range of loading that causes tensile stresses, a procedure that has been adopted for a number of tunnels. For example, for concrete with a tensile strength of 2 MPa and 1 MPa
longitudinal prestress, the range of loading causing longitudinal tensile stress could be 50% higher.
tie the ring concrete and its reinforcing steel to the shell to !6rovide the composite action. The steel assembly is usually built on land, provided with end bulkheads, and then launched sideways or longitudinally. It is usual for the bottom part of the exterior concrete to be in place. In most cases, this "keel" concrete provides the necessary additional weight to ensure stability, and perhaps also sufficient strength, during towing. The ring concrete inside and remaining outside concrete ballast are placed while the assembly is afloat. The structural loadings during the subsequent construction stages are quite complex. The stresses in the shell are more severe during launching, towing and outfitting with interior concrete than occur when the element is finally immersed. The interior shell is about 8 mm thick and is stiffened with external longitudinal stiffeners. The form plate is usually 6 mm thick. For double-shell design, the composite steel concrete tunnel is completed by covering the steel shell all around with a thick concrete cover, which provides mechanical and corrosion protection for the steel shell. The double-steelshell element end detail shown in Figure 3-5 illustrates the typical structural steel arrangement. The single-steel-shell type of tunnel is simpler in concept, although also subject to critical steel shell stresses in the launching stage, especially when the interior concrete is not yet installed. To stiffen and stabilise the element in the launching stage, the base of the interior concrete ring is installed prior to launching. The steel shell is stiffened with stiffening plates and temporary transverse spiderweb frames. For double-tube single-shell elements, a vertical longitudinal steel truss is constructed in the center between the shells. This truss will later be absorbed in the interior concrete and partly removed where cross-passages are required. Figure 3-1 shows an example of a single steelshell element. Cathodic protection is sometimes required for singlesteel-shell tunnels if the tunnel passes through a zone of stray currents, such as a subway system or an industrial facility. In such cases, provisions are made to measure the currents at test locations. These tests will determine the required level of impressed current needed to prevent loss of steel section.
12260
~87o,
KEEL CONCREll[
6,130
6130 ="~.
dam
The fixed permanent ballast concrete, inside or outside. The weight of protective membranes and cover concrete. The roadway pavement, suspended roadway slabs, or fixed-track support concrete.
valves " ~
I oliqnment
weoge
I,., form
Figure 3-5. Double-steel-shell element and detail, and typical structural steel arrangement. Formplate is only partially indicated.
tion if a special services gallery is used. The exterior walls in the latter case would not require large panel recesses, which can act as crack initiators. The same could apply if the size of the low point drainage sump would require deep recessing into the base slab.
4.4 WeightBalance
The design of the cross-sectional geometry is very sensitive to variations in the density of the water and the construction materials; dimensional inaccuracies; and the weights of temporary equipment needed for transportation, temporary installation and the permanent condition. A concrete tunnel element must be able to float with all temporary immersion equipment on board. The freeboard should be minimal to reduce the amount of permanent and temporary ballasting. The temporary on-bottom weight, with the water ballast tanks filled, must be sufficient. For the permanent condition, it must be guaranteed safe against uplift with the fixed ballast in place. The roof is usually covered with a protective concrete layer that is also used for trimming purposes. When bituminous waterproofing membranes are applied to the walls, protective wood cover or concrete cover is also applied to the walls, either as cast-in-place or precast panels. The roof edges are usually bevelled, to reduce the risk of hooking ship's anchors. The minimum factor of safety for the permanent condition of immersed tunnels is often specified as 1.10, based on the following conditions: 1" Uplift forces: Buoyancy by the water at the maximum expected density and according to the theoretical displacement. Hydraulic lag, if applicable in tidal waterways.
These factors are used to determine the required geometry. The actual safety factor may be slightly higher or lower, depending on the actual as-built dimensions. For example, a part of the stable type of roof protection, such as rip rap, may be allowed to be included in the safety factor of 1.10. Sometimes a minimum safety factor of 1.06 is applied for concrete tunnels with only the structural concrete and ballast concrete as stabilising factors. For steel shell tunnels, 1.07 is applied, excluding the contribution of the side walk concrete. The minimum temporary safety factor during installation is usually 1.03 after release of the immersion equipment. Sometimes allowance has to be made for uplift caused by hydraulic gradient. A hydraulic gradient can be caused by tidal lag of piezometric height underneath the tunnel, which can occur with silty or clayey backfill in tidal waterways. Another cause of hydraulic gradient is the suction caused by the squatting of ships passing over; however, this factor is only considered in the installation stage. For steel shell tunnels, the total amount of concrete needed for the weight balance amply exceeds that required for strength. The external ballast concrete is the variable factor for the weight balance. For concrete tunnels, the thickness of the structural concrete is generally sufficient for the strength. The determination of the final geometry is more complicated because the ballast concrete is on the inside. Variation of the internal ballast volume affects the internal geometry.
5.
Immersed tunnels are rigid structures in the longitudinal direction. The stresses with which the structure would respond to axial tensile strain (temperature) and longitudinal bending strain (unequal settlement or large surcharge discontinuities) depend on the material properties and the longitudinal articulation.
2. Stabilizing lo~ds:
The theoretical weight of the structural steel, concrete and reinforcement steel, assuming a realistic density for the concrete that will not exceed the actual density.
steel bulkhead
/ - JOINT S~c'noN OF CONCRETERfNG BULKHEADPLATE REMOVED "~ / / -CONCR~FE RING DRA,N11JBE \
~ i o ] n t . f i l l e d
wizh
'~l
/ /
s~m. ~.~.\
exhau~
air duc'li
, ~
plate
Figure 3-6. Typical tremie concrete joint for a double-shell steel tunnel.
Figure 3-7. Intermediate joint with sliding arrangement for a double-steel-shell tunnel . ment under expected long-term decompression and relaxation of the gasket. Nevertheless, a second flexible rubber water barrier is installed at the dewatered joint by bolting it to the inside faces of the two tunnel elements. The shape of this curved, slab-type rubber seal or gasket, often referred to as the "Omega" seal because of its shape, is sometimes considered to be the main seal. The space between the Gina gasket and the Omega seal is usually drained off to the inside of the tunnel, providing a direct indication of the performance of the outer gasket. An example of an intermediate flexible joint design for concrete tunnels is shown in Figure 3-8. This type of joint can also be used for steel tunnels, when flexible joints are required. The Gina-type gasket acts as a flexible joint under compression, and can practically be considered as a hinge in longitudinal moment transfer. Shear resistance can be ignored, because it has a tendency to slip along its base under shear deformation. For this reason, shear deformation across a joint is not limited by properties of the Gina g~isket, but rather by the allowable shear strain of the Omega gasket, especially with regard to its corner sections. The exterior rubber gasket should be placed as much as possible to the outside of the structure, in order to keep the recess between the side and the outside of the gasket shallow. This arrangement prevents backfill material from accumulating gradually and obstructing the proper movement of the joint.
"Gina" gasket, is used for practically all concrete tunnels. It is used as a temporary seal at the installation stage and remains as a flexible compression seal for the permanent stage. The facing tunnel element ends are lined with steel plates that are matched as parallel planes within +5 mm tolerance. The gasket is clamped at its backside. The specifications for material characteristics and geometry are usually based on the permanent sealing require-
between the opposing tunnel faces to maintain separation during dewatering. They are positioned to lie inside the future temporary watertight enclosure and will maintain the longitudinal compression. The joint can be made monolithic, or with shear transfer provision to the concrete on one side of the joint and with a flexible waterstop on the other side. In this way, the joint will resemble the action of the expansion joint described in Section 5.5, above. An example of a final joint for concrete tunnels is shown in Figure 3-10.
upper bond values to find extreme bending in different members of the frame. In numerical analysis, it is practical to model an elastic foundation for the base with a given spring constant. For the soft soil case, the effect on transverse moment distribution is practically the same as a uniform ground pressure distribution. In the case of hard subsoil, it is advisable to investigate the sensitivity to the spring constant, because the spring constant cannot precisely be determined and may vary with time.
6.
6. I Transverse Analysis
For the transverse analysis, a rectangular concrete tunnel can be considered as a series of plane frames. When the loads and soil reactions are constant in the longitudinal direction, or vary only gradually in that direction, the frames can be analysed with balanced loads. However, in areas of heavy surcharge, such as embankments, and, especially, near discontinuities of surcharge, as well as in areas of expected redistribution of soil reactions, the external vertical loads acting on the plane frame are not balanced. The shear forces between the adjacent frames need to be analysed for these conditions. An elastic beam analysis can be conducted to determine the longitudinal distribution of the vertical subsoil reaction, and the longitudinal shear distribution. The shear forces can then be entered as vertical loads acting along the vertical members of the frame. The application of hydrostatic pressure and surcharge loads is straightforward. The magnitude of the lateral soil pressure and possible wall friction caused by the backfill cannot easily be determined. It is best to assume lower and
r l
L_ . . . . -_ _ l
,xxxxxxxxx-
The understanding of the longitudinal performance of concrete tunnels is important in view of the relatively low tensile strength capacity of the concrete and the desire to avoid transverse cracks. It has been shown that these cracks can indeed be avoided. The effects of hydrostatic compression, temperature stresses and longitudinal bending on the longitudinal concrete stresses are explained below with numerical examples.
i/
L/m~"JIION JO~rT
MOADWAT ~JM~'ACm<O
1
I
I
Figure 3-9.
N
T ~IT ~
I
Figure 3-8. Intermediate flexible joint for concrete tunnels ( D u t c h solution). Typical Dutch expansion joint.
.~ov~i.E oorngtoAu
,,J
Temperature deformation caused by a temperature decrease of the tunnel by 10C would cause a decompression of about 10 mm of joints positioned at intervals of 100 m. A numerical example of concrete compression, based on the use of a rubber compression joint, is shown below: for D = 20 m: initially f (avg.) long-term f (avg.)
c
for D (avg.) = 6 m (near shallow end of a tunnel): long-term fc (avg.) = 0.5 x 6 x 0.024 = 0.072 Mpa*
ROADWAYSURrAC~O
I 'TP.AN~'~C~ .ONT
1
Figure 3-10. Example of the final joint for concrete tunnels.
concrete area is equal to the ratio of the densities of concrete and water, a simple expression for the average concrete stress is: f (avg.) = D e * 0,024 MPa, ( D in m) The longitudinal hydrostatic compression force will follow the changes in water depth (e.g., by the tide), as long as the immersed part of the tunnel has a free end and friction between the free end and the section under consideration can be ignored. The hydrostatic compression will be fixed as soon as the tunnel is closed between the terminal structures. The compression is assumed to act at the centroide of the structure when flexible intermediate compression joints are used. Losses of the fixed hydrostatic compression. Losses of the fixed hydrostatic compression are caused by timedependent relaxation of rubber compression joints and shortening of concrete, causing decompression of the joint. The degree of relaxation of the rubber compression joints varies with the characteristics of the selected type and the range of compression. Generally the loss will be about 50% of the initial compression. Typical performance rates for flexible rubber compression gaskets used as intermediate joints are shown in Figure 3-11. Shortening of the concrete can be caused by shrinkage and temperature deformation. Creep of the concrete can be practically ignored. The concrete codes do not provide proper parameters to determine the shrinkage strain for thick-walled concrete members of immersed concrete tunnels. If calculated according to the FIP-CEB model code, a very small strain is found. The shrinkage strain is not restrained when flexible intermediate joints are utilised, but the resulting decompression of these joints will cause a minor loss of the locked-in compression force. In the Netherlands, the shrinkage strain is ignored for tunnels in which the concrete is directly exposed at the outside.
The relevant temperature changes of the concrete are of a seasonal nature. At the time of installation, the whole of the structure can be assumed to be the same temperature as the surrounding water. The temperature changes of the concrete are usually specified in two parts: 1. An overall increase or decrease, relative to the ambient ground water, with a range from _+ 15C to + 10C. 2. A temperature "gradient" over the thickness of a wall or slab (e.g., a linear temperature variation from 0C on the outside to + 10C on the inside: this gradient can then be split into an overall increase or decrease of 5C, to be added to (1), above, and a linear variation over the thickness from -5C to +5C, or vice-versa). The overall increase or decrease of the temperature of the structure will lead to longitudinal deformation. For tunnels with monolithic intermediate joints between the terminals, this could lead to concrete stresses in the range of the tensile strength or yielding of the vertical joints. Tunnels with intermediate flexible compression joints are not restrained to longitudinal deformation of the concrete, but the flexible joints will react with a minor increase or decrease of compression. The linear temperature variation over the thickness of a wall or slab of +5C to -5C will cause bending stresses varying from -1.0 MPa to +1.0 MPa over the thickness of a concrete wall or slab. Although small, these stresses are a magnitude larger than the hydrostatic precompression. On the outside, there will be tension in the summer and compression in the winter. The stress diagrams of Figure 3-12 clearly show that the temperature variations will not cause transverse cracks.
Example calculation:
External cross-section:* Modulus of Inertia: Section modulus: Elastic modulus: Spring constant: h b I S E K = = = = = = = = 1o = 8 m 30m 1.200 m4 300 m 3 2 1@ MPa ko b 2.30 60 MN/m 2 35.5 m
Characteristic length: Undisturbed settlement: Maximum shear force (at fixed end):
Yo = 0.05 m
Vma* = 2 E1 Yo]lo 3
= 53 MN Maximum bending moment (at 3/4 1o from fixed end): M x = 0.322 1 V x = 612 MN/m Maximum bending stress: f (max)= 2.04 MPa
This simple calculation shows that a range of 50 mm of such differential settlement can cause monolithic tunnel elements to crack transversely. These cracks would be fulldepth over the thickness of the base slab and the walls. While it is preferable to prevent forced deformations of such magnitude, sometimes it cannot be avoided. Controlling the width of such cracks by longitudinal reinforcement will not provide ductility of the tunnel element as a whole; furthermore, it is not economical. However, transverse cracks in the concrete can be avoided by providing ductility at the construction joints. Because the vertical construction joint cannot transfer concrete tensile stresses, it will crack at an early stage. The longitudinal reinforcement through the joint should not yield within the expected bending range, which should be safely below the bending capacity of the uncracked concrete. Beyond that range, the joint should yield, and thereby provide sufficient rotation capacity to avoid an increase in tensile stresses in the uncracked concrete. The resistance of the longitudinal reinforcement will not be mobilised, except in the joints. The concrete will not crack. If the spring constant in the above example were four times higher, or if it were the same but the undisturbed settlement was limited to 25 ram, the maximum bending moment would only be 300 MN/m and the maximum concrete tensile stress would be 1.0 Mpa. The reinforcement through the construction joint could then be 0.25 % of the concrete area, without yielding of the joint. Monolithic concrete tunnel elements can be designed with a relatively low amount of longitudinal reinforcement (similar to concrete tunnel elements with expansion joints), without the risk of concrete cracks resulting from the in-
,!
Figure 3-11. Typical performance of flexible rubber compression gaskets used for intermediate joints,
"rOIRg~"ftJIE 14~
INTt~t~X~A ~IC
i
~ ~, ~ ~0
......
'"
P"=*'~
"" '7
I .......
v rnctM -
'1
i
'
M ma -
0u~J~
7t0"c F t o ~ ,
~ -I.0 ~
-0..o6 ,,
a,/PJICmlON
Figure 3-12. Diagrams of concrete temperature and stresses for summer and winter, valid for tunnels with flexible intermediate joints.
Figure 3:13. Examples of longitudinal analysis of the response of concrete tunnels to unequal settlement.
In some cases, interior "spiders" may be required to maintain the roundness of the shell plate. Accurate roundness must be maintained to assure a uniform thickness of the interior ring concrete. This is critical to the eventual structural integrity of the elements when it is backfilled. A portion of the exterior concrete about 1.5 m thick and called the "keel," is usually placed in the bottom of the formplate prior to launch. This provides protection for the bottom of the element during launch, as well as stability during towing. After this concrete is placed, the loads of the element are transferred to launching sleds or, in the case of end-launching, fore and aft poppets prior to launching. The shell and diaphragms may require additional reinforcements in either case. The keel concrete is considered to act as a longitudinal beam acting compositely with the stiffened shell plate and diaphragms. Both static and d y n a m i c water pressures that will occur during launch must also be determined and allowed for in the design. While the basic design of the elements takes into account some of the expected launching loads, the contractor must be made responsible for checking the details for his proposed method of supporting and launching, and must provide whatever additional reinforcement that may be r eq u i r ed . Towing may impose longitudinal stresses on the steel element. Often the practice has been to tow the element under its own flotation, because the draft of an element that has not been outfitted with interior concrete and the remainder of the exterior concrete is only about 2 to 3 m. If the elements are to be towed in ocean waters, the height of long periodic swells becomes important, because the swells impose significant longitudinal bending moments. Unless they are accounted for, these bending moments could cause buckling of the top shell. More recent practice, when ocean towing is required, has been to tow elements on large offshore construction barges in order to reduce this effect. Chapter 7 of this edition discusses the impact of tunnel element transportation on immersed tunnel design for both steel and concrete options.
sections of the interior concrete are placed, additional moments are imposed on the shell. The sequence of concrete placement is designed to counteract the hogging effect and reduce its deflection ,;o that at the end, the element is as straight as possible. As concrete weight is added, the element settles lower into the water, thereby imposing transverse moments on the shell plate and diaphragm. The formplate does not take any of this load, as water is allowed into the pockets between the shell and the form plates. The only exception occurs at the end pockets between the dam plate (extension of the bulkhead) and the first diaphragms. These pockets are filled with structural concrete, placed in the dry, to help resist the large forces that act on the joint during dewatering. Typically, the shell plate is designed to take the water pressure as the outfitting progresses. The placing sequence is determined by the engineer as a series of reasonable volumes that can be placed while limiting the moments and shears in the exposed shell plate. The concrete placement should be symmetrical about the transverse and longitudinal center-lines of the element. As the head of water increases on the shell plate, the circumferential moments must be resisted by the diaphragms and shells embedded and, therefore, fixed in the keel concrete. This condition must be checked for each stage of the placement sequence; it generally becomes most critical during the placement of the heavy haunch sections adjacent to the bulkheads. Longitudinal loads are resisted by the shell plate and the radial stiffeners. Stress analysis for critical compressive stress for buckling of curved panels under uniform compression are used to determine moment capacity after applying a suitable safety factor. The local buckling of the shell is investigated for loading in torsion and transverse shear. Buckling of the arch between stiffeners is investigated. The shell plate and longitudinal stiffeners are considered as a cylindrical shell spanning between diaphragms. The shell plate alone can resist water pressure as a ring structure and a cylindrical beam. Assumptions are made regarding how these loads are distributed between the shell and the stiffeners. The diaphragm is designed as a ring structure, with its ends fixed in the keel concrete. The effective width of the shell plate that acts with each diaphragm, is determined as a function of the centerline radius of the shell and its thickness.
Before computers became readily available, the tunnel configuration assumed for analysis was simplified to circular or elliptical shapes in order to facilitate otherwise very difficult, or perhaps even impossible, calculations. Thanks to the availability of relatively inexpensive, accurate, and easy-to-use computer software such as STAAD III, STRUDL and others, the structure may be modelled to the exact shape of the proposed tunnel. Because the location of the frame line, the cross-section area, and the moment of inertia are all interrelated and affect the results (moments, shears, axial loads and deformations), an iterative design approach is necessary. Using such an approach, a series of computer runs are made, each using information from the previous run, to modify frame line location and member properties, thereby converging on an accurate and economical solution. For the preliminary design, the total thickness of tunnel walls may be based on previous experience or very rough hand-calculations. The tunnel is drawn accurately and the frame line is laid out by "eye" or judgment, more or less along the mid-thickness of tunnel walls, discounting exterior tremie concrete. Coordinates of selected "joints" along the frame line can be very conveniently obtained if the work is done by CADD drafting. Using the results of this first run, the design engineer can proportion steel and concrete depths and thicknesses based on stress analysis, and obtain new section properties. These data are used in turn as input for a second, more refined computer analysis, wherein the frame line is located at the center of gravity of the gross composite section and member properties are based on the same. This process is repeated until the designer is satisfied that changes resulting from further refinement are negligible. Theoretically, the final computer run should be based on a frame line corresponding to the center of gravity of the cracked section, and member properties, area, and moment of inertia are based on the same. It should be kept in mind that the composite threedimensional tunnel structure with variations in loadings is extremely complex. The goal of the analysis should be to determine an envelope of maximum moments, shears and axial loads so that all possibilities of overstress are eliminated. For example, gross section properties may result from maximum effect at one location and minimum effect at another, whereas the cracked section may show the reverse. This approach, inconceivable in the past, is not at all difficult today with the help of clever computer programming. Future analysis methods may make use of three-dimensional finite element modeling. This will be particularly interesting in the analysis of loadings such as ship collision, anchor dragging, and internal explosion. The longitudinal distribution of these loads is otherwise very difficult to determine.
Table 3-1.
B. TOTAL STRESS INCREMENT FOR COMBINATION OF BASE LOADING WITH ANY OF THE FOLLOWING: B1. Extreme high water B2. Anchor dragging or dropping B3. Sunken ship load B4 Temperature restraints B5. Unequal settlements B6. Temperature restraints and unequal settlements B7. Internal explosion B8. Earthquake, unequal settlement B9. Earthquake, temperature restraints, unequal settlements B10. Erection condition
1.5 ***
1.00
1.30 ** 1.15 1.0 1.50 1.65 1.30
NOTE: A dash indicates that this aspect is known not to be reviewed, or is not critical. * Refers to Dutch practice: the load factor used for the ultimate limit state is 1.7, reduced for the material factor incorporated. ** The factor 1.30 also includes extremely high water. *** 1.4" A + 1.15 * B1.
8. L o a d i n g s 8.1 Loading Combinations and Allowable Stress Increments An indication of the types of loads and their combinations used for the design of immersed elements is given in Table 3-1. The table, based on data provided from the U.S.A., Japan, and the Netherlands, is not exhaustive. It should be noted that the factors given in the table are based entirely on the specific project conditions and requirements. The reference projects in the table relate to tunnels of widely differing structural nature. They are intended to provide an understanding of the range of values for loading conditions that are encountered in immersed tunnel design(s). The reference projects used in the table are: I: A steel shell traffic tunnel (Ted Williams Tunnel, Boston, Massachusetts, U.S.A.) IIa: A longitudinally prestressed reinforced concrete traffic tunnel with waterproofing membrane (Tama River Tunnel, Japan). lib: A reinforced concrete railway tunnel with waterproofing membrane (Keyo-Line Daiba Tunnel, Japan). III: A typical Dutch reinforced concrete traffic tunnel (Tunnel De Noord).
For reference II, only the ultimate-limit state factors are given. However, service-limit state verification is also done in view of watertightness requirements; for example, temperature restraint has to be considered.
8.2.4
The probability and extent of internal explosion loads depends very much on how the tunnel is to be used. An example is the rectangular Liefkenshoek Belgium, located near an industrial area, which has been designed for an internal explosion pressure of 4 bar. This requirement substantially increased the amount of transverse reinforcement needed.