Shotcrete use in the Southern Link tunnel, Stockholm

B.I.Karlsson Swedish National Road Administration T.P.Ellison Betongsprutnings AB BESAB ABSTRACT: The Southern Link is a major road tunnel project, on the southern outskirts of Stockholm. It covers a distance of about 6 kilometres, of which 4.5 run through rock tunnel. The project is estimated to cost approximately SEK 7 billion. This project includes tunnels with imposing dimensions co nstructed by a drill and blast method through hard rock. Steel Fibre Reinforced Shotcrete (SFRS) and bolts were used as primary rock support. The performance requirements for the shotcrete were very stringent with respect to strength, durability, and environmental considerations. In order to find a mix design that could satisfy these requirements, several pre-construction trials were carried out. To date (October 2001) approximately 95% of the shotcrete work has successfully been completed. Due to blas ting, shotcreting could not take place closer than 50 metres from the working face. To determine the effect of peak particle velocities arising from the blasting, a test programme was carried out. Results from these tests showed that blasting-induced vibrations decline very quickly with distance from the face. 1 INTRODUCTION The Southern Link (Sdra Lnken) is a road tunnel project to the south of Stockholm city. The primary aim of the project is to improve the urban environment for the citys inhabitants by freeing the central urban area of through-traffic. The project will solve a major traffic congestion problem by diverting over 100 000 cars daily from the area. According to the plans the project will be completed in 2004. governed by maintenance considerations rather than geological factors. 1.2 Design The Southern Link consists of a 6 km road of motorway standard. There are several underground grade-separated junctions, installation rooms, shafts, and emergency exits. The total length of rock tunnel is about 17 kilometres, and the rock volume to be excavated is well over 2 million cubic metres.

Figure 1. Map of Southern Link route.

1.1 Geology The rock tunnels in the project are located in the Precambrian Scandinavian gneiss formation. Rock quality is fairly good, typical Q-values (Barton et al. 1974) are between 1-100. Some fractured zones occur, and cracks are more intense in some areas. Rock support requirements in this project were
Figure 2. Main tunnel section.

Most of the tunnels are located quite close to the ground surface. Soil depth varies from 0 to 20 m, and the groundwater level is near the ground su r-

face. The dimensions of the rock tunnels are quite imposing with spans up to 32 metres. The project includes 12 tunnel portals, and due to the rock topography there are underground connections between rock tunnels and concrete-lined tunnels. One of the tunnel portals has a span of 32 metres and an overburden of less than 5 m. This section was excavated very carefully by two pilot tunnels and 28 stope rounds (see Figures 3 and 4). Longitudinal rock bolts up to 10 m long were installed outside the tunnel profile, and the rock mass, excluding the tunnel face, was shotcreted before excavation started. Shotcreting and bolting were carried out after each blasting round. The required shotcrete thickness was 0.5 m, and cavities were also filled in places behind the lining. In e xtreme cases the total thickness of lining was up to roughly one metre. Approximately 500 rock bolts with a length of 2-4 m were installed over this 20metre length of tunnel. This method places high demands on the shotcreting process, especially in order to satisfy the requirement for spraying layers up to 200 mm thick that remain in place. This was possible due to the use of alkali-free accelerators. The excavation and rock support in this section took six months to complete. Deformations were checked during the whole operation, and so far no alarming settlements have occurred.

32 m

Figure 4. Blasting sequence, plan view.

1.3 Environment It was specified that the groundwater level was not to be affected during construction or after the tunnel was completed. This was due to the pre sence of sensitive structures, streets, railways, pipes, and buildings founded on wooden piles above the tunnels. Co nsiderable effort was made to limit noise, vibration, and traffic congestion at road works. All buildings in the area were inspected before excavation started, and on -line vibration recorders were used during excavation. Drilling and blasting were not permitted between 10 pm and 7 am. 1.4 Information

10 6 5 2 1 13 9

12

8 7 4 3 12 m

11

32 m

Figure 3. Blasting sequence, cross-sectional view.

It has been essential to keep the general public well informed as to the progress of the works. Two public relation officers have been employed basically full time for this purpose. Information h a s b e e n d ispersed through notices in local newspapers, flyers, brochures, open house evenings, etc. A bus has been specially fitted and designed as a mobile exhibition, travelling around to the various residential areas affected by the works. Tours down into the tunnels have also been arranged for the general public, and have proved very popular. 1.5 Sealing Due to the location of tunnels in an urban env ironment, extensive sealing works are necessary to avoid lowering the ground water table, which could cause settlement and damage to the buildings above.

The permissible groundwater seepage along the tunnel alignment varies from between 1.5 to 4 litres/min per 100 metres of tunnel. These values are guidelines established by the water rights court. Although cement grouting is the primary method used, chemical grouting using enviro nmentally-friendly polyurethane products is required in certain places. 1.6 Rock reinforcement Rock bolts, and fibre reinforced shotcrete are the primary means of rock reinforcement. Rock reinforced shotcrete is used in areas where the spans are long, and the rock is of inferior quality. 1.7 Tunnelling The excavation method used was traditional drill and blast. All the tunnels were pre-grouted, and some post-grouting is being done in places as well. In some areas with no or very little rock overburden, the ground was frozen before excavation. In these areas, temporary support was provided by shotcreting, and the final support consists of a pre-cast lining with a thickness of 0.8 m. To prevent frost problems associat ed with ingressing water in the tunnels, isolated wall drains were used. The rock support mainly co nsisted of untensioned rock bolts and shotcrete. The crowns of all tunnels were supported with fibre-reinforced shotcrete, while most of the tunnel walls were sprayed with plain shotcrete. 1.8 The shotcrete works The volume of shotcrete amounts to a total of about 66 000 m3, of which 75% (49 500 m3) is fibre reinforced shotcrete and 25% (16 500 m3) is plain shotcrete. Two specialists in the field, Betongsprutnings AB, BESAB, Gothenburg Sweden (referred to as Contractor no.1 below) and Byggs Sprutbetong AB, Malung Sweden, a subsidiary of NCC and referred to as Contractor no.2 below were contracted to carry out the works. Contra ctor no.1 was responsible for 40% and no. 2 for 60% of the shotcrete volume. As most of the bedrock in the Stockholm area is high quality, the normal procedure was to use 40 mm fibre reinforced shotcrete covered by 20 mm plain shotcrete. Each company designed and d eveloped its own materials and machinery. Even if the shotcrete design differed, both were able to meet the clients specifications in the preconstruction trials and at the quality controls during the implementation.

2 DEVELOPMENT OF MATERIALS AND EQUIPMENT 2.1 Requirements There were several clauses within the specifications affecting shotcrete properties, and these stipulated the following requirements: - The tunnels were to be constructed to last at least 120 years. - For environmental reasons, use of admixtures that could pollute the ground water was not permitted. - Adhesion between rock and shotcrete was not to be affected by admixtures. - The concrete was to be resistant to freezethaw cycling. - Strength requirements were considerably higher than normal, both with regard to modulus of ruptu r e a n d p o s t-cracking toughness. Apart from these specifications, there were additional issues to consider: - Regard was to be paid to the health and well-being of those working in the tunnels. - It was considered beneficial to be able to place thick layers of shotcrete in one pass. - Rebound must be low. Pumpability and sprayability must be good. The shotcrete mix must be cost-competitive to be acceptable. 2.2 Pre-construction trials, Contractor no.1 The designers and contractors had no prior experience of any project where the shotcrete properties were as stringent as for these tunnels. For example, frost durability has usually not been specified in other tunnelling projects in Sweden. It was therefore necessary to conduct preconstruction trials under site conditions to demonstrate that the required FRS properties could be achieved. An initial mix design was determined from available literature on materials. See Table 1.
Table 1. Initial mix design for shotcrete Ingredient Aggregate (0-8 mm) Portland Cement (SR) Silica Fume Water/cement ratio Quantity (kg/m 3) 1600 480 10 0.45

It was also decided that Dramix RC 65/35 hookedend steel fibres would be used at a dosage rate of 55 kg/m3. Superplasticizers and alkali-free accelerators from Rescon, Sika, and Master Builders were tested. Test spraying was performed in a t u n n e l u n d e r c onstruction in Stockholm. Betongindustri AB, Stockholm, who were later contracted to deliver ready-mix concrete for shotcrete use during construction, supplied the concrete. The pre-construction trials started in 1997 and were completed in 1998. 2.3 Pre-construction trial results, Contractor no.1 Vattenfall Utveckling AB, lvkarleby, undertook laboratory testing of shotcrete properties. All the requirements were fulfilled after only two rounds of trials. It was especially satisfying that freezethaw tests showed acceptable results. The final mix included Rescon Superflow 2000 as superplasticizer and Rescon AF 2000 as accelerator. The results from laboratory -tests for this mix design are shown in Table 2.
Table 2. Test results for trial mix shotcrete Property Compressive strength (MPa) Post-crack flexural strength f5.10 (MPa) Post-crack flexural strength f10.30 (MPa) Frost resistance (kg/m 3) Method SS 13 72 20 ASTM C1018 ASTM C1018 SS 13 72 44 Specified 40 4.0 3.0 0.5 Result 60 4.5 4.0 0.15

Betongindustri AB, Stockholm and Swerock were contracted to deliver ready-mix concrete for use during construction. Internal Curing was als o added to the shotcrete. The results for this mix design are shown in Table 4.
Table 4. Tests results for trial mix shotcrete_______________ Property Method Specified Result_____ Compressive SS 13 22 20 40 52 strength (MPa) Post-crack flexural ASTM 4.0 5.3 strength f5,10 MPa C1018 Post-crack flexural ASTM 3.0 4.7 strength f10.30 MPa C101 8 Frost resistance SS 13 72 44 0.5 0.2 (kg/m3)_

2.6 Quality control A number of tests had to be carried out on the in situ shotcrete for quality assurance purposes during construction. These were all included in the project specifications. The tests included: 1. Fibre content. 2. Thickness, measured in 25 mm diameter drilled holes. 3. Compressive strength, based on cubes sawed from panels sprayed during construction. 4. Flexural strength of beams sawed from panels sprayed during construction. 5. Adhesion, based on cores drilled and pulled off insitu. 6. Freeze-thaw resistance. Frequency of testing depended on risk estimations and geological conditions. The compressive strength tests were normally carried out once per 1000 m2 of in-place shotcrete, and flexural tests once per 2000 m2. Adhesion tests were done once per 1000 m2. Freeze-thaw tests were only necessary in zones where frost was expected. 2.7 Machine development, Contractor no.1 In conjunction with the pre -construction trials to develop the shotcrete mix for this project, machines were developed to suit the prevalent co nditions. Aliva AG, Switzerland, was contracted to supply concrete pumps, robotic arms, and the additive pump for shotcreting. Contractor no.1 su pplemented this with the carrier, compressor, and electrical equipment. The maximum capacity of the concrete pump was 20 m3per hour. However, this was reduced to 10-

2.4 Mix design, Contractor no.2

2.5 Preconstruction trial results, Contractor no.2 Vattenfall Utveckling AB, lvkarleby, undertook the testing of shotcrete properties on behalf of Contractor no.2. All the requirements were met after three trial rounds. The freeze-thaw tests also showed acceptable results in these cases. The final mix included Rescon Superflow as the superplasticizer and Rescon AF 2000 as the accelerator.

15 m3/h during practical spraying. The total vertical reach of the robot arm was 15 metres, and the unit could move five metres along the tunnel du ring spraying before re-location of the equipment was necessary. 2.8 Machine development, Contractor no.2 Contractor no.2 developed its own shotcrete robots for use in the project. Built on articulated Volvo trucks, they have a vertical reach of 14.5 m, and a horizontal reach of 6 m. The machines are equipped with accelerator tanks, admixture tanks and pumps for the accelerators and admixtures, robot arms, Putzmeister concrete pumps with a maximum capacity of 20 m3/h with compressors, hydraulic hose and cable reels, and the required work lighting. An extremely accelerated concrete was used at weak spots where freezing had proven necessary for passage. This was mixed in-situ in a Robomix with a capacity of 10 m3 per hour. 2.9 Construction To date, approximately 95% of the contract has been completed, which is equivalent to about 62 700 m3 shotcrete. Some changes in the mix design were necessary during construction, the most important involved changing the superplasticizer to Master Builders Glenium 51. This was done due to some unexpected variations in viscosity in the concrete that influenced pumpability. More than 400 strength tests, including both compressive and flexural strength, have been completed during construction to date, and all show satisfactory and uniform results. 3 VIBRATION TESTS ON APPLIED SHOTCRETE 3.1 Background Technical specifications for this project required that the shotcrete should reach a compressive strength of 6 MPa before loading, and that the maximum allowable Peak Particle Velocity (PPV) arising from blasting-induced vibrations was 150 mm/s after 24 hours. It was therefore necessary that spraying not be allowed closer than 50 metres from the blasting face. These restrictions tended to interrupt the construction cycle during this project. Similar projects have been succes sfully completed without such restrictions, e.g. the Arlanda Express railway tunnel north of Stock-

holm, where shotcreting was carried out much closer to the blasting face. It was therefore desirable to determine how blas ting vibrations affected the shotcrete. Ansell (2000) has studied how blasting-induced vibrations affect young shotcrete, and in particular has determined the maximum allowable PPVs associated with stress waves close to th e location of a blast. It was therefore necessary to determine how far from a blast the Peak Particle Velocity decreases to the critical level. Swedish Rock Engineering Research (SveBeFo), the Swedish National Road Administration, and the contractors Selmer/BESAB have assessed the last issue in the Southern Link project. 3.2 Blasting tests The blast holes were charged with Site Sensitised Emulsion (SSE) explosives. The holes were was 5.2 metres long, and each was charged with approxi-mately seven kilograms of explosive. All tests were performed without shotcrete on the tunnel walls. Accelerometers were installed in 0.30 m deep boreholes in the tunnel walls at a distance of between 5 and 50 metres from the blasting face. Each installation included two accelerometers, one parallel with the tunnel axis and one perpendicular to the tunnel wall. The accelerometers were grouted into the boreholes and were co nnected to a measuring unit with one channel dedicated to each accelerometer. Seven blasting rounds were recorded, each involving 14 devices. 3.3 Data analysis Peak Particle Velocity (PPV) data from four rounds was used for the final analysis (Reidarman & Nyberg 2000). The measured vibration directions at each station were both parallel and perpendicular to the tunnel axis. Altogether 69 values across and 98 values along the tunnel axis were used from the four rounds. For construction work it is necessary to know the maximum vibration level, and the most interesting may be the level that no vibrations exceed. 3.4 Results The tests showed that the first arrival Peak Particle Velocities (PPVs) were usually moderate in magnitude, even as close as five metres from the blasting location. This was particularly true of the registrations from those accelerometers located

perpendicular to the tunnel wall. Recorded peak particle velocities (PPVs) for the tests are shown in Figures 5 and 6. It is apparent that peak vibration levels parallel to the walls are higher than the values perpendicular to the surfaces. For simplicity, a line representing the maximum vibration intensities (PPV=80-1.5 R) was used for both directions. Based on the report by Reidarman & Nyberg (2000) it was concluded that the measured vibration levels were far below the maximum permitted 150 mm/s required by the client. Comparing this study with Ansell (2000) also indicates that shotcreting might be performed close to the blasting face without causing damage to the concrete. However, more studies are needed to confirm this. Such a study can provide important input to future tunnelling projects.
PPV (mm/s) vs distance (m)
100

REFERENCES Ansell, A. 2000. Dynamically loaded rock reinforcement. Doctoral Thesis, Bulletin 52, Dept. of Structural Engineering, Royal Institute of Technology, Stoc kholm. Nyberg U. & Reidarman L. 2000. Vibrationsmtning fr nyutfrd sprutbetong nra front i Sdra Lnken, Swedish Rock Engineering Research SveBeFo, Stockholm. Reidarman L. & Nyberg U., 1999. Vibrationer bakom front vid tunneldrivning, SveBeFo Report 51 (in print), Stockholm, Sweden.

50

10

20

30

40

50

Distance (m) Figure 5. Measured Peak Particle Velocity (PPV) parallel to tunnel axis, as a function of distance from blasting face.

PPV (mm/s) vs distance (m)

50

0 0

10

20

30

40

50 50

Distance (m)

Figure 6. Measured Peak Particle Velocity, PPV, perpendicular to the tunnel axis.