Railway Bridges on the Iron Ore Line in Northern Sweden – From Axle Loads of 14 to 32,5 ton
Ibrahim Coric,
Trafikverket, Luleå, Sweden
Björn Täljsten, Thomas Blanksvärd, Gabriel Sas, Ulf Ohlsson, Lennart Elfgren
Luleå University of Technology, Luleå, SwedenContact: lennart.elfgren@ltu.se
Abstract
The Iron Ore Railway Line was built around 1900 and has more than 100 bridges. It has a length of ca 500 km and runs from Kiruna and Malmberget in northern Sweden to the ice-free harbour in Narvik in Norway on the Atlantic and to Luleå in Sweden on the Baltic. The original axle load was 14 ton. The axle load has gradually been increased to 25 ton in 1955, to 30 ton in 1998 and to 32,5 ton in 2017.
The increases in axle loads have been preceded by monitoring and assessment studies of the bridges. The capacity and need for strengthening or replacement of the bridges have been evaluated. Many of the bridges could carry a higher load than what it was designed for.
Experiences from studies before the axle load increases in 1998 and 2017 are presented and discussed.
Keywords: Railway bridges, foundations, steel, reinforced and prestressed concrete. Assessment,
Strengthening, Fatigue
1 Introduction
The Iron Ore Railway Line was built 1883-1904 and has a length of ca 500 km. The line runs from Kiruna and Malmberget in northern Sweden to the ice free harbour in Narvik, Norway, on the Atlantic and to Luleå on the Baltic in Sweden, see Figure 1. There are 144 bridges (20 long concrete, 72 short concrete, 12 steel, 2 composite and 8 rock tunnels).
There has been a steady increase of the amount of iron ore to be transported. In 2017 the northern route (Kiruna –Narvik) carried
about 25 million ton and the southern route (Kiruna – Luleå) about 10 million ton.
The original axle load was 14 ton (140 kN).
The axle load has gradually been increased to 25 ton (250 kN) in 1955, to 30 ton (300 kN) in 1998 and to 32,5 ton (325 kN) in 2017.
The increases in axle loads have been
preceded by monitoring and assessment
studies of the bridges. The capacity and need
for strengthening or replacement of the
bridges have been evaluated.
2 Axle Load 14 ton in 1900
The soil conditions were in some places problematic with peat, bogs and mosses [1]. Up- side down tree stumps with roots were sometimes used to carry the track. Stone foundations where used for most of the bridges, see Figure 6. and some of them are still in use, see Figure 7.
Figure 1. Map of the Iron Ore Line between the harbours in Narvik in Norway and Luleå in Sweden.
3 Axle Load 25 ton in 1955
The track was successively strengthened and many bridges were changed and in 1955 the allowable axle load could be raised to 250 kN [1].
Figure 2. An ore train passes a concrete bridge.
4 Axle Load 30 ton in 1998
4.1 Investigations
The demands continued to grow and an investigation was carried out on how to increase the axle load to 300 kN, Paulsson & Töyrä [2], [3], Figure 2. Four bridges were monitored and a decommissioned 7m long concrete trough bridge from Lautajokk, close to the Arctic Circle, was transported to Luleå University of Technology and tested for fatigue, Figures 3-4. According to the codes, the fatigue capacity of the slabs was too low in many of the bridges on the line, Paulsson et al. [4], Thun et al. [5].
4.2 Fatigue Test
The bridge was loaded with 6 million cycles with an axle load of 1.2×300 = 360 kN (including a code dynamic load factor of 20%). The mid-point deflection is given in Figure 5. The increase with time is mostly due to creep in the concrete. No notable damages were observed and only hair line cracks appeared in the bottom of the slab. Finally the bridge was loaded to the maximum capacity of the jacks, 875 kN. A beginning of yielding in the reinforcement was noted but the ultimate load capacity was probably slightly higher due to strain hardening in the reinforcement.
Figure 3. Full scale fatigue test of a 29 year old railway trough bridge at Luleå University of Technology, Paulsson et al. [4], Thun et al.[5].
The design concrete compressive strength was 40 MPa. However, due to coarse grinding of the cement the concrete strength increased over time. 16 concrete cores were drilled out, tested and found to have a compressive strength fcc = 72.6 MPa (mean of 6 tests) in the slab and fcc = 81.2 MPa (2 tests) in the beams and a tensile splitting strength (4 cylinders) fcspl = 4.4 MPa and a uniaxial tension strength (4 cylinders) fct = 2.9 MPa.
Figure 4. Cross section A-A (top) and elevations B- B and C-C (bottom) of a tested railway trough
bridge [4], [5]. In the slab there is no vertical reinforcement so the shear transfer to the beams has to be taken by the concrete. A feared – but not
materialized - shear crack is indicated in red close to the right support of the slab in section A-A.
Figure 5. Mid-point deflection. At 0 and 250 kN load for 6·106 load cycles with a maximum
deflection of 4,8 mm [5].
4.3 Assessment versus test results
The tests showed that the fatigue capacity of the bridge was much higher than what was predicted by the codes, Thun et al. [5]. Critical was the shear capacity in the connection of the slab to the longitudinal beams with no shear reinforcement in the slab. According to the fib Model Code[6] the number of possible shear stress cycles, N, can be written as a function of the ratio of the maximum shear force Vmax (under relevant representative values of permanent loads including prestressing and maximum cyclic loading) and the design shear resistance attributed to the concrete Vref = VRd,c as:
Log N = 10 (1 - Vmax/Vref) Eq. (1) For Vmax / Vref = 0.5 we get N = 105 load cycles and for Vmax / Vref = 0.3 we obtain N = 107 load cycles.
Monitoring has given that a bridge of this type experiences four axles (two bogies) as one load cycle. During a year with 8 trains per day, 68 wagons (each with 2 bogies) per train and 365 days we obtain 198,560 cycles ≈ 200 kc.
The shear load effect VE = Vmax at the support of the slab consists of dead load VEg and train load VEq
with values VE =VEg+ VEq = 34 + 113 = 147 kN/m.
For a concrete with a compressive strength of 40 MPa, we obtained with the Swedish code BBK94, Thun [5], a shear resistance of Vref ≈ 215 kN/m and Vmax / Vref = 147/215 = 0.68 which gives N = 103.2 = 1.5 kc. If the concrete capacity is increased to 80 MPa, we obtain Vmax / Vref = 147/292 ≈ 0.5 and N = 105 = 100 kc which corresponds to about half a year of traffic. The value of Vref varies in different codes depending on traditions and amount of longitudinal reinforcement. The test showed that the bridge could stand more than 6,000 kc without other than hairline cracks. A fatigue shear crack that might be detrimental is indicated at the right support of the slab in Figure 4 (top). The crack did not materialize. Probably, the longitudinal reinforcement had a positive influence by dowel action and by keeping the section tight together. The fatigue capacity of concrete is often estimated in a rather conservative way, while the reinforcement is more prone to fail in fatigue and is also modelled with better accuracy; see further discussion in Thun et al. [7] and Elfgren [8].
4.4 Further investigations and summary
Further investigations with monitoring and probabilistic methods where made on a bridge over Luossajokk close to Kiruna, Enochsson et al.[8], Figure 6. They showed that the bridge could carry the higher load.
Figure 6. Luossajokk - Continuous RC single-trough railway bridge with spans of 10.2 m and 6.3 m.
Built in 1965 and tested in 2001, Enochsson et al.
[8].
In general it can be said that the static load carrying capacity of the bridges was surpassed in some local sections in many bridges when the axle load is increased to 300 kN. This applied mostly to the transverse direction capacities in trough bridges, bottom slabs and bearing constructions of steel bridges. Main beams of steel or concrete in the longitudinal direction often showed enough capacity to handle the increased loads.
Foundations showed enough load capacity for 30 tons with some exceptions.
In summery it was found that about one half of the 144 bridges could carry the increased axle load 300 kN in their existing states after a standard assessment. The other half had to go through an extended assessment process. Out of them, 60 bridges were cleared, 10 bridges were strengthened, and only 11 bridges had to be replaced.
The results from the investigations were also used to check the load-carrying capacity on a parallel Finnish railway line, Elfgren et al. [10]
5 Axle Load 32,5 ton in 2017
5.1 General
All bridges have again been assessed-for the increased axle load. Bridges with problems have been further investigated and some of them are followed by monitoring. The original stone foundations are checked for movements and settlements (Number of different kind of bridges, age and span lengths, Capacity, Assessments, Tests)
5.2 South Rautasjokk Bridge
The steel truss railway bridge at Rautasjokk, about 20 km NE of Kiruna, was built in 1962 on an old stone foundation from 1902, see Figure 7. It is the twin to a bridge over Åby River some 45 km west of the city of Piteå, see Figure 8. Both bridges have a length of 33 m and were built in the transition time from riveted to welded bridges. The Åby Bridge was monitored and then moved and placed on new foundations beside the railway line and tested to failure, Häggström [11]. An extensive monitoring program was carried out including some 140 sensors for loads, displacements, strains, temperatures and accelerations. At the loading to failure two hydraulic jacks was used with cables anchored in the bed rock. In order to estimate the ultimate capacity a detailed FEM model was developed with the Abaqus software.
The model consisted of shell elements considering all connections as rigid.
Figure 7. South Rautasjokk Bridge built in1962 on a foundation from 1902, Häggström [11]
Figure 8. The Åby Bridge, a twin to the Rautasjokk bridge in Figure 7, placed beside the rail track for
testing to failure after being moved from its original position, Häggström [11].
5.3 Åby bridge test results
At the final testing non-linear deformations started at about 8 MN and continued up to 11 MN when the top cord buckled, see Figure 9.
Initially the FEM was carried out using design values as input for the material properties. Later on the input was updated and a new analysis was performed. The real behavior of the structure fits somewhere in between the two FE models. The FEM-results with updated material parameters have approximately the same peak load as the one from the tests. Nevertheless the non-linear
behavior recorded during the tests was not accurately described by FEM calculations.
Since the load added to the bridge only corresponded to one wagon, the load level for a whole train set is compared for a certain number of points. The loading is compared for the normal observed from a train set F46 (which is the train set it was designed for) and the loading according to current standards (Eurocode) for new bridges along the line subjected to the heaviest axle loads in Sweden (Load model 71 with =1,6). Safety or dynamic amplification factors are disregarded. The comparison shows that the bridge could withstand loading that substantially exceeds both the load it was designed for as well as the load the model in use today before failing.
The influence of damages to the bridge has been studied in the EU-project MAINLINE [12]-[14]
6 Conclusions
The gradual increase of the load on a railway line has been presented. Assessment, monitoring and tests have been carried out in order to check weather existing bridges could carry increased loads.
Figure 9. Force-deflection diagram with inserted
modeled and real deformations for the buckling of the top cord,
Häggström [11].
Numerical tools are proven to be reliable instruments for assessment especially when combined with material testing and monitoring to calibrate the models. The tested structures had a considerable “hidden” capacity which is little reflected during ordinary assessment processes and which is accounted for neither in standards nor in design guidelines [14]. Perhaps some of these differences arise from redistribution of loads during the testing in the statically indeterminate structures. Another reason is the high safety factors that are used both for loads and materials. Methods to strengthen structures with insufficient capacity are given and discussed in e.g. [11]-[14].
7 Acknowledgements
The authors gratefully acknowledge financial support from the European Union, Trafikverket, LKAB/HLRC, SBUF and LTU. They also thank colleagues and collaborators who have worked in the projects Sustainable Bridges and Mainline, and the Swedish Universities of the Built Environment The experimental work and monitoring campaigns were carried out in cooperation with staff of the Mining and Civil Engineering (MCE) Laboratory (formerly Complab) at LTU.
8 References
[1] Persson, E. Bertil and Sten, Rolf (2003):
Malmbanan Luleå – Riksgränsen – (Narvik), (The Iron Ore Line. In Swedish).
http://www.historiskt.nu/normalsp/staten/
malmbanan/malmbanan_main.html
[2] Paulsson, Björn and Töyrä, Björn, project leaders (1996): 30 ton på Malmbanan. (30 ton on the Iron Ore Line. In Swedish). In the project several studies were carried out regarding different parts of the infra- structure. The studies are presented in a summary report and in seven detailed reports. A similar study was carried out by Jernbaneverket in Norway for the line Riksgränsen – Narvik.
3.0 Infrastruktur - Broar och Geoteknik (Bridges and Geotechnology. Summary Report), 34 pp
3.1 Inventering broar (Inventory of Bridges), 22 pp + 8 app.
3.2 Beräkningar och konsekvenser – broar (Bridges – Calculations), 48 pp + 10 app.
3.3 Forsknings- och utvecklingsprojekt avseende betongbroars bärighet (Research Project on Bridges), 51 pp + 5 app.
3.4 Geoteknisk inventering (Geotechnical Inventory), 53 pp + 14 app.
3.5 Stabilitetsutredning (Stability of Embank- ments), 19 pp + 5 app.
3.6 FoU Beräkningsmodell för grund- läggning på torv (Foundations on Peat), 53 pp + 11 app.
3.7 Geotekniska åtgärder (Geotechnical Actions) 50 pp + 5 app.
[3] Paulsson, Björn (1998): Assessing the track costs of 30 tonne axle loads. Railway Gazette International, Vol 154, No 11, pp 785-788.
[4] Paulsson, Björn, Töyrä, B., Elfgren, L., Ohlsson, U., Danielsson, G., Johansson, H., and Åström, L. (1996): 30 ton på Malmbanan. Rapport 3.3 Infrastruktur.
Forsknings- och utvecklings-projekt avseende betongbroars bärighet. (Static field tests on four trough bridges and a laboratory fatigue test on one trough bridge. In Swedish), Banverket & Luleå tekniska universitet, 51 pp + 5 App.
Available at http://ltu.diva-portal.org/
[5] Thun, Håkan, Ohlsson, U., Elfgren, L. (2000).
Fatigue Capacity of Small Railway Concrete Bridges: Prevision of the Results of Swedish Full-scale Tests. Comparison and Analyses.
Final Report to the European Rail Research Institute, ERRI D216, Structural Engineering, Luleå University of Technology, 99pp.
http://ltu.diva-portal.org/
[6] fib Model Code 2010 (2013): International Federation of Structural Concrete.
Hardcover Ed. 2013, 434 pp,, ISBN 978-3- 433-03061-5.
[7] Thun, Håkan, Ohlsson, U., Elfgren, L. (2011):
A deformation criterion for fatigue of concrete in tension. Structural Concrete, Journal of fib, Vol 12, Issue 3, pp 187-197.
[8] Elfgren, Lennart (2015): Fatigue Capacity of Concrete Structures: Assessment of Railway Bridges. Research Report, Luleå University of Technology, 103 pp. Available at http://ltu.diva-portal.org/
[9] Enochsson, Ola; Hejll, A; Nilsson, M; Thun, H; Olofsson, T; Elfgren, L. 2002. Bro över
Luossajokk. Beräkning med
säkerhetsindexmetod (Luossajokk Bridge.
Probability modelling. In Swedish). Luleå University of Technology, Report 2002:06, 93 pp, http://ltu.diva-portal.org/
[10] Elfgren, Lennart, Enochsson, O., Puurula, A., Nilimaa, J., Töyrä, B. (2009): Preliminary Assessment of Finnish Railway Bridges.
Railway Infrastructure Upgrading with Increase of Axle loads from 25 to 30 tonnes on the Line Tornio - Kolari. A Comparison with the Swedish Railway Bridges on the lines Luleå – Narvik and Haparanda – Boden. Luleå University of Technology, 40 pp. Available at http://ltu.diva-portal.org/
[11] Häggström, Jens (2016). Evaluation of the Load Carrying Capacity of a Steel Truss Railway Bridge: Testing, Theory and Evaluation. Licentiate Thesis, Luleå University of Technology, 2016, 142 pp.
ISBN: 978-91-7583-740-6, see http://ltu.diva-portal.org/
[12] MAINLINE (2014). MAINtenance, renewal and Improvement of rail transport Infra- structure to reduce Economic and environmental impacts. A European FP7 Research Project during 2011-2014. Some 20 reports are available, see e.g. D1.2 at http://www.mainline-project.eu/
[13] Sustainable Bridges (2007). Assessment for Future Traffic Demands and Longer Lives. A European FP 6 Integrated Research Project during 2003-2007. Four guidelines and 35 background documents are available at www.sustainablebridges.net
[14] Paulsson, Björn, Bell, B., Schewe, B., Jensen, J. S., Carolin, A., and Elfgren, L. (2016).
Results and Experiences from European Research Projects on Railway Bridges. 19th IABSE Congr. Stockholm, 21-23 Sept. 2016:
Challenges in Design and Construction…
Zürich, 2016, pp. 2570 – 2578. ISBN 978-3- 85748-144-4. Available at http://ltu.diva- portal.org/
