Construction work on Düsseldorf’s subway system was started in1973 and the local authorities always made a point of minimising disturbance to local inhabitants, traffic and trade in the vicinity of the tunnels.
In 1997, a tender was issued for a 1123m-long shield-driven tunnel to carry a double-track metro line, to link the southern suburbs more directly and reliably with the Central Station and the downtown area. A joint venture was formed by the companies Bilfinger & Berger BauAG, Bouygues, Hochtief AG, Philipp Holzmann AG and HBG / Wayss & Freytag AG who, as partners, constructed the double track tunnel with an outer diameter of 9.15m. There were two stations constructed between the shield-launching and destination points.
Geotechnical conditions and tunnelling technology
A slurry shield with an outer diameter of 9.44m was launched at the shaft in Siegburger Strasse in October 1998 and arrived at Station Oberbilker Markt in April 1999.
The ground cover above the crown was between 4.7m and 11m, with a maximum groundwater level of 14m above invert.
The soils encountered at the shield drive were, quaternary gravels and gravely sands and dense tertiary fine sands. The tunnel was constructed through a densely built area and passed directly beneath numerous buildings, a railway bridge, a street car line and numerous utilities. The foundations of a railway bridge, located some 70m behind the start shaft were crossed under with a ground cover of about 5m without disturbance to the ongoing train traffic. A total of 79 000m³ sand and gravel were excavated by the shield. The shield was equipped with 15 double thrust jacks to give a total thrust force of 5400tonne. The shield reached its deepest point along the tunnel alignment at 19m below ground level. A total of 240 000m3 of spoil were excavated for the tunnel and the stations.
No special measures needed
After evaluting the results of the soil investigations and structural analysis the client decided to drive the shield under the bridge without any stabilisation grouting or other special measures. A survey and settlement control cross-section was installed prior in front of the bridge.
In order to ensure tunnelling face stability support pressure had to be applied. The support pressure was transmitted by the pressurised bentonite slurry in the working chamber. Soil was cut with a rotating cutting wheel and hydraulically transported to a separation plant at the surface where the soil and slurry were separated. Separated slurry was returned into the circulation.
A particular construction method was chosen for the stations ‘Oberbilk’ and ‘Ellerstrasse’. The roofs of the stations were built conventionally by the cut-and-cover method with diaphragm walls as the vertical elements. In order to prevent an interruption of the shield drive, the station locations were tunnelled through, prior to further excavation of the stations. Locally the ground cover between the shield crown and the ceiling was only 0.5m. The diaphragm walls at the station entrance and exit were cut through four times.
At the end of the shield drive construction of the stations continued with the soil beneath the roof being excavated and the tunnel tube being dismantled inside the station. In co-ordination the excavation sequences, anchors were placed to tie back the diaphragm walls, and subsequently , the underground stations were completed. The TBM was not recovered intact at its destination at ‘Oberbilker Markt’ as the necessary shaft would have been located at an busy road intersection. Thus, the shield jacket remained in the ground and is lined with an in situ concrete shell.
Support pressure analysis
The support pressures were calculated with regard to the water pressure, the soil properties and the loads transferred to the tunnel face from the buildings and structures. These loads had also been to take into account to ensure both stability at the tunnel face and minimum settlement.
Based on a limit equilibrium failure model for non-cohesive soils, the support pressures could be calculated according to Mohr-Coulomb theory. The model consists of a soil wedge in front of the shield and a silo above. Partial factors of safety of hv=1.1 on the water pressure and he=1.75 on the active three-dimensional earth pressure were considered in the design.
To account for the variations in ground water levels over time, the calculations were made for minimum and maximum groundwater level established for the duration of construction. Thus, the support pressure was adjusted to the encountered ground water level on site during construction.
The main operational modes considered in the support pressure analysis are:
a) slurry supported tunnelling face – standard operational mode;
b) partial air/slurry support – for maintenance work;
c) compressed air support over the full height of the tunnel face – for maintenance work.
Within the design process of the support a blow-out criterion needs to be observed. The support pressure in the tunnel crown must fulfil equilibrium of forces in vertical direction, otherwise, a sudden loss of support pressure, known a ‘blow-out’ of slurry or air may occur. The pressure relief reflects the influence of varying overburden, the ground water table and that of the stresses induced by adjacent structures.
Under the railway bridge
The highest pressure was to be applied for the crossing under the railway bridge due to the large stresses the foundation induces into the ground. Minimum support pressures were found for the passages of the station where the ceiling slabs and the diaphragm walls transferred stresses partly below the tunnel and away from the tunnelling face.
Due to a precise pressure regulation to changing conditions and an optimal transmittal of the design into reality on site, the support pressures were very close to those calculated in advance. Minor deviations are mostly due to adjustment of the support pressure to the actual water level.
In order to avoid possible flow of tail void grout in tunnelling direction, the support pressure was increased in the station areas in conjunction with the tunnel site management after experiences made in the first station box. This contributed to a tunnel drive with only small settlements.
For the railway bridge, tender documents restricted the tolerable settlements to 24mm and for the buildings adjacent to the tunnel the allowable slope induced by differential settlements was defined at 1:350. The maximum settlement measured at ground level in tunnel axis was less than 15mm and on average mostly negligible, i.e. 4.4 mm.
Settlements of the railway bridge were far below the 24mm as fixed in the contract and there was not the slightest effect on the structure with a settlement of less than 4mm. Prior to passing under the bridge, the quality of the tunnelling was successful, as proven in the survey section, with special attention to the magnitude and distribution of settlements. A major source of settlement was found to be related to tail void grouting.
Design of segmental tunnel lining
In respect to the vicinity of structures sensitive to settlements and the alignment of the tunnel close to the ground surface, a tunnel lining quasi-resistant to bending composed by reinforced concrete segments was put up for tender by the client.
The resistance to bending of the lining had to be enabled by rotation of neighbouring rings by half the length of a segment and the transmittal of shear forces in the circumferential joints by coupling of the rings.
Conventionally, coupling of neighbouring rings is achieved in the design by means of a groove and tongue system in the circumferential joints. The groove and tongue system serves three vital purposes:
1.centring aid for the geometrically correct placement of the segments;
2. transmission of shear forces between neighbouring rings of a tunnel tube quasi-resistant to bending;
3. transmission of the jacking forces introduced at the tongue.
Experience over time showed that the combination of the purposes as above may result in concrete flaking off the edges of the groove. While these possible over-strains are purely cosmetic, if located inside the tunnel they are easy to repair, but the failures located at the outer circumference cause permeable fractures in the tunnel lining and are vey very expensive to treat.
In order to improve this deficiency and to produce a watertight tunnel lining without any overstrains at the outside or inside of the tube, a cam and pocket coupling system was introduced to Germany by our project in Düsseldorf. With this system, the functions 1 and 2 as above are served by the cam and pocket system. The transfer of the jacking forces during tunnelling will be carried out by load transfer zones to the side of the cam and pocket system.
The jacking forces are not tranmitted by the coupling element any more. Due to this separation of purposes, it was possible to design a far better geometry for the cam and pocket compared with the conventional groove and tongue coupling system. The prime objective of the geometrical optimisation of the cam and pocket system is to ensure that in the event of overdue strain being exercised on the coupling system during ring assembly, the cam always fails first, as a kind of predetermined breaking point.
The outer pocket edge remains undamaged so that in this way, an important prior condition for the watertight segmental lining is ensured.
The tunnel tube is created by means of left and right conical rings with a conicity of 60mm. In this way it is generally possible to place the key-stone above the tunnel horizontal axis. Neighbouring rings are rotated half a segment in a tangential. So, instead of an unfavourable cross joint, T-joints are created.
The distribution of segments within a ring is governed by the set-up of the 15 pairs of jacks around the periphery of the selected TBM for the tunnel drive. As a result, a ring comprises seven equally large segments and a key-stone of half the size of a standard segment. A segment is 45cm thick and 1.50m wide.
The ring position is selected in such a way that the shoes of the jacks in each case are applied to the longitudinal joints and to the centre of a segment. The cam and pocket system for coupling segmental rings was originally developed by Wayss & Freytag to stengthen and stiffent the tunnel lining. An advanced system with a cam and pocket was also initiated by Wayss & Freytag, as a member of the BTC Group JV (Boortunnel Combinatie), during the construction of the Second Heinenoord Tunnel south of Rotterdam. This then served as a basic solution for the tunnel lining in Düsseldorf.
The cam design chosen for Düsseldorf measaured 136mm by 336mm. This gives a theoretically determined ultimate load for the cam and pocket system of 430 kN.
The maximum coupling force of 60kN determined in the statistical analysis thus results in a high safety factor for its final state as well as unintended constraints when the rings are being assembled.
A Wayss & Freytag first
For the first time, Wayss & Freytag installed gasket profiles with an integrated gliding layer at the longitudinal joint of the key-stone and the neighbouring segments.
The key stone is placed conventionally without any special lubricating of the profile surfaces. The selected gasket profile features a thin hard rubber layer extruded onto the surface of the profile. This functions as a gliding layer due to a comparatively low coefficient of friction.
The gasket design ensures water-tightness of the lining, as a rubber accumulation of gasket material in the corner zones during segment placement is avoided. Under most adverse circumstances this material accumulation could result in concrete flaking off due to unfavourable stress distribution.
Related Files
Longitudal section
Distribution of the ground loss along the tunnelling drive
