Remco Lamoré (Van Oord, Rotterdam, The Netherlands), Iran Tisheh (Boskalis, Papendrecht, The Netherlands), Ruud Steenbakkers (Van Oord, Rotterdam, The Netherlands)
For the D&C-works of the Fehmarnbelt Fixed Link connection, a new 18 km submerged tunnel connecting Germany and Denmark, a set of well interpreted geotechnical design parameters for both in-situ conditions as dredged and re-handled soils had to be established. With the challenge of re-using all dredged tunnel trench material within the project, a soil testing program was conducted for the determination of the geotechnical parameters of the main rehandled soil unit. Sustainable construction works can only be achieved by understanding the geological and geotechnical conditions in detail. Laboratory tests and a full-scale trial are performed to determine the parameters of the key construction material.
This paper presents the approach, the findings and the observations of the investigations, and presents the encountered characteristics of the re-handled glacial till. The conclusions show that robust design and construction with dredged and re-handed material is economically possible with a risk-driven investigation approach.
INTRODUCTION AND PROJECT DESCRIPTION
The Fehmarnbelt Fixed Link Project consists of a motorway and rail tunnel connection of about 18 km long between Fehmarn in Germany and Lolland in Denmark . The Project location is shown in Figure 1.
The seabed in the center of the tunnel lies at approximately 30m below mean sea level. The bottom of the tunnel exceeds depths of 16m below the seabed.
The Tunnel Dredging and Reclamation Contract (TDR), which is one of four principal D&C civil work contracts for the Fehmarnbelt Fixed Link Project, has been awarded to Fehmarn Belt Contractors I/S (FBC); a joint operations of marine contractors Boskalis and Van Oord. Briefly described, the TDR contract comprises of construction of a trench for the submerged tunnel elements, a work harbour including access channel at the Danish side and a work harbour at the German side. Furthermore, large reclamations, terminal areas, retaining dikes, and stockpiles are constructed for the placement of dredged materials and sea dikes and noise barriers for protection of the Danish populations.
Due to the sustainable and durable way of working in large dredging and reclamation projects, re-handling of materials gets more and more common. The geotechnical parametrization of re-handled materials is an important challenge for geotechnical engineers. Availability of large quantities of the best construction materials is not the key driver anymore for these works. Re-use of dredged materials was also one of the primary requirements for the Fehmarnbelt Link Project.
It was a project requirement that all mechanically and hydraulically dredged materials of the tunnel trench and the work harbour must be re-used as construction material and fill. Besides sand, locally mined Upper Till (mainly in the dry) is also used in Denmark for fill material under shallow foundations of buildings and roads. When sufficiently dry, excavated Upper Till can be constructed in layers allowing for compaction. There is however much less experience with dredged Upper Till (FBC will use the world´s largest backhoe’s for dredging of Upper Till to -25 MSL), transported by barges and placed back below the water table.
During the preparation and design phase of the project several offshore and nearshore geotechnical investigations have been performed in and around the Tunnel trench alignment.
In the tunnel trench route various soils and weak rock are encountered that have to be dredged. On the Danish side, directly from seabed level, Quaternary pre-consolidated glacial deposits of hard Upper Till (U5) are found, occasionally with a layer of post-glacial Sand on top. Soil Unit 5: Upper Till (Lower Quaternary), is predominantly found on the Danish side and is considered to be the most promising construction material for the project’s structures.
This Upper Till deposit is typically a very hard, (very) silty (very) sandy Clay Till with locally small to large boulders. The Upper Till Unit is a low plasticity clay (plasticity index of 6-10%), with initially a very high strength. The investigations included boreholes with alternating coring and down-the-hole cone penetration tests (CPT). A typical CPT is presented in Figure 3. The average q_net>20 MPa. The deposit is highly over-consolidated with a pre-overburden pressure exceeding 1,000 kPa (ice-loading). The q_net (cone resistance corrected for in-situ vertical stress) varies due to the varying composition of the Upper Till.
On the cores from various depths triaxial tests have been performed to derive undrained shear strength parameters. From a total of 29 triaxial tests a mean value of s_u=1,009 kPa is derived. A characteristic value of s_(u,k)=740 kPa is derived (statistical methods in accordance with EN1997-1).
The results of the soil interpretation and parameter determination for Upper Till is summarized in Table 1. These in situ Upper Till parameters are used (as base layer) in the various geotechnical designs of structures that are built on top of the initial Upper Till.
Table 1 Geotechnical characteristic parameters (index and strength) for in-situ Upper Till
FBC’S INVESTIGATION OF REHANDLED UPPER TILL (UNDRAINED SHEAR) STRENGTH
Since the strength characteristics after dredging and rehandling were key for the construction of many structures Fehmarn Belt Contractors (Boskalis and Van Oord) investigated the strength characteristic of the dredged and rehandled Upper Till. This was done in a staged way from course to fine, from small scale laboratory tests to full-scale embankments, see Figure 4. The latter to overcome scaling effects of the laboratory tests and avoid an eventual trial-and-error approach during construction.
Initially a Desk Study was performed, followed by testing thereafter (from laboratory testing to a trial embankment).
The results of Phase 1 “Standard Scale Laboratory Testing” will provide a first indication of the conditions since relatively small scale is tested. Phase 2 “Large Diameter Laboratory Tests” reduce the inaccuracy, compared to Phase 1, since the lumps sizes are more realistic compared to the lumps during execution. A full-scale Trial dredging and embankment construction of re-handled Upper Till, Phase 3, is close to the construction methodology and will establish good insights and best lesson’s learned for further construction (which will be monitored).
Phase 0: DESK STUDY
As a starting point FBC performed a desk study of available publications, literature, reports, etc.
Mechanically dredged and handled Upper Till was applied for Øresund REF . These projects are located more towards the north and have slightly other in situ characteristics, see Table 2, however very useful information can be obtained from these projects.
Literature information is available, as Clay Till was also encountered and used at the Storebælt and Øresund projects . In Table 2 in situ index parameters of the Fehmarnbelt Upper Till are compared to the Storebælt Clay Till and the Øresund Clay Till. As will be demonstrated later in this article, the strength of rehandled Upper Till largely dependents on the moisture content, hence the initial conditions are important.
Table 2 Characteristics of in-situ Upper Till in other projects
|Parameter||Fehmarnbelt Upper Till||Storebælt Clay Till||Øresund
|Plastic limit (%)||10||10||9|
|Liquid limit (%)||19||16||22|
|Natural moisture content (%)||9||12||12|
|Clay content (%)||15||15||15|
|Void ratio (-)||0.26||0.30||0.33|
For the re-handled soils and the re-handled strength, key parameters are the liquid limit, the plastic limit and resulting plasticity index. When soils are dredged and handled the moisture content increases due to the process. For the Øresund Clay Till the Liquid Limit is 22% (with a liquidity index LI=(w_n- w_p)/(w_l – w_p ) of 0.13). The Femarnbelt Upper Till is close to a Semi-Solid State with a liquidity index of about 0. It is noted that Atterberg Limits tests are conducted after sieving the material and discarding gravel. The consistency limits therefore only provide an indication of the relatively fine fraction of the Upper Till. Although the liquidity indices from the in-situ soils are low, a relatively small addition of water leads to a significant increase and significantly different behavior.
Based on the available information and likely sensitive properties of the material to be dredged, an extensive study into the characteristics of this material was required.
Phase 1: STANDARD SCALE TRIAXIAL TEST RESULTS
With the knowledge from reference projects, it is known that the behavior of the remoulded Upper Till depends highly on its moisture content. For all tests performed this was taken into account by testing at various moisture contents. The main aim of the tests was not only to derive engineering parameters, but also to determine the variation of those parameters depending on moisture content.
The specimens have been prepared with a level of compaction of 92% standard Proctor. Based on general experience with reclamations it is expected that this state will be reached in the reclamations.
To investigate the behavior after re-handling, a number of standard scale laboratory tests were performed. The performed tests are:
Unconsolidated, undrained triaxial tests (100×200 mm);
Consolidated, isotropic undrained triaxial tests (100×200 mm);
For the standard scale triaxial tests 100×200 mm dimensions were preferred rather than the 35×70 mm as the material contains gravel. It is expected that removing the gravel results in a less representative sample considering the relatively high gravel content of the Upper Till.
The first series of tests consisted of triaxial tests on 100×200 mm (diameter x height) specimens. In advance of testing the material was remoulded from intact cores (obtained during the geotechnical investigation for the tunnel trench) to small lumps (approx. 2-4 cm). Moisture content loss during the remoulding works are minimized, later testing indicated that the moisture content had not decreased significantly. Because of the size of the triaxial test, all lumps larger than 16 mm were dry sieved. The triaxial tests were executed as Unconsolidated Undrained (UU) and as Consolidated Isotropic Undrained (CIU) tests. Both series of tests were performed at target moisture contents of 7%, 9%, 12% and 14%. To this extent the moisture content of the remoulded was tested, subsequently the required volume of water to reach the target moisture content was determined. Tests were executed with a confining pressure of 50 kPa, which is a reasonable average estimate of the expected conditions after placement of dredged Upper Till in the project.
After the CIU tests moisture content was measured on the specimen, it was observed that the moisture contents were relatively high compared to the target moisture content, leading to lower undrained shear strengths than expected based on the prepared moisture content. It is concluded that the moisture content increased during the saturation phase of the CIU test. This has led to undrained shear strength data mainly on the higher end of the tested moisture content interval.
A moisture content dependency of undrained shear strength can be seen in the test results;
Phase 2: LARGE DIAMETER TRIAXIAL TEST RESULTS
After the execution of the standard scale triaxial tests the potential influence of lumps (and lump removal) was investigated. Although very useful, the standard scale laboratory tests could potentially not incorporate ‘macro’ behavior of the fill. For this purpose, large diameter tests were conducted, having dimensions 500×500 mm (diameter x height). The tests were again executed as Unconsolidated Undrained and Consolidated Isotropic Undrained triaxial tests.
Preparation of the large diameter triaxial test is shown in Figure 7, where the white membrane is supported by a temporary casing cylinder. The material is installed at a standard Proctor density of approximately 92%, to this extent the material is placed in layers of 10 cm. Based on visual observations and compaction effort, the resulting specimen can be considered dense. On the top-right side of Figure 8 the specimen can be seen after the test. This specimen was built in with an intended moisture content of approx. 9% (at 92% Standard Proctor Density), resulting in a degree of saturation of approximately 60%. Sampling after the test indicated that the actual average moisture content equaled 9.6%. In Figure 8 the resulting load-strain data is shown by the right line in the graph.
Another specimen with an intended moisture content of approximately 13% can be seen in the bottom-right of Figure 8. After the test it was found that the actual average moisture content equaled 10.8%. No further saturated occurred after installation of the large diameter specimens. Figure 8 shows the resulting load-strain data measured during the tests.
Based on combining the triaxial test data (standard and large diameter), Figure 9 be drawn.
In Figure 9 the moisture content dependency of undrained shear strength can be seen. It can be concluded that there is a strong relation between the undrained shear strength and moisture content for the material tested. Furthermore, the large diameter tests produce generally (slightly) higher strengths for similar moisture content compared to the standard scale tests. It is noted that large diameter testing was not possible with higher moisture contents then tested above.
Phase 3: FULL-SCALE EMBANKMENT TRIAL
To verify the results of the laboratory tests and subsequent parameter derivation and furtherly investigate the material behaviour and slope stability, an embankment trial was carried out. The embankment was built at the project site against the existing seadike, in relatively shallow waters. After construction of the trial embankment, according the proposed methods of construction, a geotechnical investigation was carried out employing in-situ testing, boreholes and subsequently a laboratory program.
The intended construction sequence consisted of reclaiming a soil volume to a level of approximately +0.5 m MSL. On top of this soil, a relatively thin layer of sand was placed (sourced from a local onshore location) to a level of approx. +1.0 m MSL. On top of this sand layer another dredged volume was placed to a level of +2.5/+3.0 m MSL.
An aerial photograph of the full-scale embankment trial is shown in Figure 10.
The full-scale embankment trial was constructed using locally mechanically dredged Upper Till. Dredging works employed a CAT390 excavator on a barge. During dredging works moisture content testing was performed on the barge. The initial moisture content is inevitably increased due to excavation and movement through the water column. This moisture content is likely further increased as the over-consolidated material experienced bulking after dredging and is re-saturated when placed below water.
After construction of the embankment (trial), a geotechnical investigation was performed. A total of two boreholes were conducted on the trial embankment. During advancing, undisturbed samples were taken and down-hole vane tests were performed. On the samples retrieved in the boreholes UU triaxial tests and Oedometer tests were performed. A total of five CPT were been performed, including at one location a dissipation test. Two CPT were located next to the boreholes in order to be able to derive a cone factor N_k based on the tests on samples taken in the boreholes.
The CPTs identified the different layers of the embankment trial, as shown in Figure 11.
The UU triaxial test results on the samples taken in the boreholes and in-situ vane test results are shown in Figure 11 for BH01 and for BH02. Based on a N_kt=15, the correlated undrained shear strength of material placed below the waterline is also shown in the figures, from nearby CPT. The grey dots show s_u values based on individual q_net measurements. The orange dots indicate the averaged correlated undrained shear strength over 0.25 m intervals. On the right side of Figure 11 the interpreted soil layers are described.
It is noted that the shallow measurements are not considered reliable for the mass behaviour of the embankment as this material is very fresh material and highly disturbed by the tracks of excavators and CPT truck.
Tests on the dredged material showed that the consistency limits of the in-situ soil and dredged soil are essentially identical.
Based on a combination of all tests performed (laboratory and embankment), the following figure is compiled.
A best-fit strength relation of s_u=4200e^(-0.4W(%)) is derived.
GEOTECHNICAL FINDINGS AND CONCLUSIONS
Based on the available information from other major projects in Denmark, like Storebælt and Øresund, and the complex work that had to be executed, an extensive testing plan was set up to be able to control the risks during the execution. A testing approach from standard scale, to large diameter, to a full trial embankment is executed, whereby the material lump sizes are increasing and the later construction methods are approached closer and closer. Based on the executed study and investigations it can be concluded that:
- The Upper Till is quite sensitive for external (marine) processes. In terms of (undrained shear) strength, the dredged and rehandled conditions of Upper Till placed below the water show a large difference compared to the in-situ (over-consolidated) conditions; from hard clay initially to soft/firm clay after dredging, transport and placement below water.
- Both standard scale and large diameter triaxial tests are executed and confirm the moisture content dependency of undrained shear strength and give valuable information for the works. Large diameter testing is limited by the moisture content, large diameter testing of remoulded Upper Till with higher moisture contents than tested for the current program was not possible. A scale factor between the standard scale and large diameter tests cannot unambiguously be derived, however results appear to suggest a factor near unity.
- Field investigations performed immediately after placement of the material show s_u values of Upper Till placed below water varying between 10 and 20 kPa. This is based on three different tests conducted (CPTs, vane tests and triaxial testing) for direct comparison and for correlations.
- It is expected that, within the possibilities in the laboratory, the standard diameter and large diameter tests provide useful results of the undrained shear strength compared to the actual material handling and lump sizes expected on site for the Fehmarnbelt development. On the project site the material will be less remoulded compared to the large diameter tests and particularly the standard scale triaxial test, as the material will be dredged with large backhoe dredgers (>30m3 bucket, Figure 14). Therefore, it was expected that the results of the laboratory tests are a lower boundary, however they seem to be quite realistic.
- FBC’s geotechnical investigation approach of rehandled Upper Till placed below the water table, and the relations developed for this, performed well for D&C purposes. In combination with monitoring during construction risks and opportunities can be applied where possible, based on the additional results of the monitoring campaign.
- With FBC’s approach sufficient confidence is gained to start the construction works, see Figure 15. With risks & opportunities and mitigating measures in place for the Upper Till below the water table the design is sufficiently robust to cope with variations on site.
Reclamation Area Sprongø, Clay Till as Engineered Fill, DGI, 24.02.1989
Fehmarnbelt Fixed Link – Geotechncial investigations, Jens Kammer, John K. Frederiksen, Gert L. Hansen, Rami Hammami, Paul Morrison, Niels Mortensen, Peter Skjellerup.
The Øresund Technical Publications – Dredging and reclamation, edited by Niels J. Gimsing and Claus Iversen