Embed Size (px)
Citation preview
10th International Conference on Short and Medium Span Bridges
Quebec City, Quebec, Canada, July 31 – August 3, 2018
Performance evaluation and analysis of the cross laminated timberfor the bridge decks
Atsushi Toyoda1, 5, Takanobu Sasaki2 and Tomoyuki Hayashi2,Humihiko Goto3,Shogo Araki4 and Shingo Kato1
1 SUNCOH Consultants Co., Ltd, Japan2 Inst.of Wood Tech., Akita Pref. University, Japan3 Akita University,Japan4 Hattori Engineering Co.,Ltd,Japan5 [email protected]
Abstract: Cross-laminated timber (CLT) is a light-weight construction material, superior in terms of transportability and workability. Applications of CLT slabs include those in civil engineering e.g., soundproofing and bridge decks; however there are no such practical implementations in Japan. This study focuses on the bridge deck application of CLT slabs and examines its potential as an alternative material for concrete bridge decks used in repair work of existing bridges. Repeated load tests and dynamic wheel load tests were conducted to evaluate the fundamental performance of CLT slabs for bridge deck applications of road bridges. The deterioration of wood due to decay is of the highest concern in using CLT slabs for this application. If this deterioration can be prevented, high long-term durability is expected. CLT slabs with improved durability owing to waterproofing treatment were used for a test construction of a forestry road bridge in March 2017.
1 Introduction
There is a total of approximately 700,000 bridges with a minimum length of two meters in Japan. A little over 20% of these are 50 years or older, with approximately 70% estimated to reach this range over the next 20 years. Approximately 70% of the 700,000 bridges are currently managed by municipal governments. The economic burden on municipal governments will be considerable. The development of economically superior construction methods has become a matter of urgency, since an enormous number of bridges will require repairs in the future. Bridge decks made of cross-laminated timber (CLT), which is a new wooden material, are lighter, stronger, and offer superior workability. This material enables reductions in the cost of steel girder enforcements, and early liberation of traffic lanes owing to improved workability by reducing the load on the main girder. It is expected to form part of the construction method for repair and replacement of bridge decks. It has been reported that CLT slab have been adopted as bridge decks in Scandinavia. Fatigue tests of the form of presumed car loads driving over the bridge repeatedly must be performed prior to the use of CLT slabs as an alternative bridge deck material to concrete. Wheel load tests, with presumed loads of T-14 load (wheel load: 56 kN) and T-25 load (wheel load: 100 kN), were performed to verify the fatigue durability of the material. The CLT bridge deck was
32-1
adopted on a trial basis for a newly constructed forestry road bridge, since sufficient performance of the CLT bridge deck was verified for the repeated loads.
2 Test method
2.1 Repeated-load tests
Japanese cedar CLT slabs, comprising 5-layer 7-ply boards of 3,600 mm length, 900 mm width, and 180 mm thickness were used for the repeated load tests (Figure 1). A resorcinol resin adhesive was used for the laminate adhesion of CLT slab. The repeated load tests were performed using a method involving repeated loading of a concentrated load at the center of a 3-m span between supports. A hydraulic servo fatigue tester (with capacity of 100 kN) was used, as shown in Figure 2. The loads and displacement of the center were measured. During testing, the load was applied in stages of 17% (20 kN), 25% (30 kN), 33% (40 kN), and 42% (50 kN) of the maximum bending failure load (118.8 kN) derived from a bending failure test using center-concentrated loading on 5-layer 7-ply Japanese cedar CLT slab of the same dimensions, while observing the extent of the damage on the specimen. Furthermore, the targeted load was loaded during the tests using displacement control, since the tester used was not capable of controlling the load. The frequency for the repeated loading was set to 1.0 Hz.
Figure 1: Laminate composition of CLT slab Figure 2: Repeated loading test of CLT slabs
2.2 Dynamic wheel-load tests
The load position was relocated for the case where the design traffic load was applied continuously to perform dynamic wheel load tests to reproduce actual drive-overs by wheel loads. This was for the purpose of verifying the fatigue durability with respect to the traffic loads, which is one of the required performance capabilities of the bridge decks. The design calculation for the actual bridge was performed for the assumed case of replacing steel bridge decks (with a distance between bridge girder centers of approximately 2 m) with CLT bridge decks. Japanese cedar CLT 7-layer 8-ply slabs with a length of 4,000 mm, width of 2,000 mm, and thickness of 240 mm were fabricated for the purpose of these tests. The specimens were secured by bolting CLT bridge decks to the steel H-girders, as shown in Figure 3 and Figure 4. The bridge decks were linked by metal joints at the center in the direction of the bridge axis. The wheel load was set to 56 kN (load of one rear wheel for the design load of 140 kN) and 100 kN (load of one rear wheel for the design load of 250N), and each of these loads was applied by driving 100,000 round trips to take measurements of the displacements and the strain behaviors of the CLT bridge decks.
32-2
690 690H300 L=3380
1000 100030
024
020560
690 690H300 L=3380
1000 1000
▽ Test floor
Specimen length 4000
1823
73
1950 100 1950
CLT bridge deckCLT bridge deck Metal joint
Figure 3: Specimen for dynamic wheel load test Figure 4: State of dynamic wheel load test
3 Test results
3.1 Repeated-load tests
The relationship between the number of repetitions and the loads, as well as center displacements, is shown in Figure 5. Loads of 17% (20 kN) and 25% (30 kN) of the maximum load, which is the presumed design load level, were applied repeatedly in stages of three million times and four million times, respectively; however, no cracks or adhesion peels were observed in the test specimen. Because the tests were conducted using the displacement control method, it is evident that while the center displacement remained constant, the load declined as the number of repetitions increased. This is surmised to have been due to the bearing stress generated on the specimens at the supporting and
loading points, the displacements of which accumulated and led to the same load not being applied even when the input displacement was given; as a consequence, the load gradually decreased. However, sufficient fatigue durability of the CLT bridge deck is considered to be indicated since no significant damage was observed on the test specimens that were subjected to repeated loads of several million times. The test continued with elevated load levels of 33% (40 kN) and 42% (50 kN) and a total of 8.5 million repeated loadings was performed; however, no damage to the test specimens was
observed.
3.2 Dynamic wheel-load tests
The values of displacements are highest for displacement 1 and displacement 6, as shown in Figure 6. The relationship between the number of drives with displacement 1 and displacement 6 are shown in Figure 7. These tests were conducted by varying the wheel load between 56 and 100 kN and the number of drives exceeded 200,000 round trips, but no damage that can trigger fatigue was observed on CLT slabs. The tensile stress of the slab calculated based on the maximum value of the tensile strain in the bridge width direction (approximately 1,000μ) was approximately 3.5 N/mm2, which correlated well with the design value of 4.2 N/mm2. These results can be considered to have indicated sufficient fatigue durability of CLT bridge decks, which is similar to the results obtained from the repeated load tests.
Figure 5: Results of repeated load tests
32-3
4 Analysis Method
4.1 Analysis Model
For comparison with experimental results, the slabs used in the experiments that had dimensions of 4,000mm length, 2,000mm width and 240mm thickness (7-layer 8-ply) were modeled for analysis. The elasticity modulus of each lamination was modeled in two different ways: an averaged 1-layer of orthotropic material (1-layer model), and multi-layer material composed of 8 individual layers of crossed orthotropic materials (8-layer model). The load was applied in the model on an area of 200 x 500 mm in order to simulate the wheel load in the experiment. For materials elements, the values obtained from experiment were adopted in the model.The elasticity moduli adopted for the 1-layer model and the 8-layer model are shown in Table 1. In the
Table, z denotes direction perpendicular to bridge axis (strong axis), x is the bridge axis direction (weak axis), and y is the load application direction. Shear modulus is obtained from the following equation.
[1] G xy=G yz=G zx=E15
where E is the elasticity modulus in the fiber direction. For the 8-layer model E = 5.0 GPa, and for the averaged 1-layer model E = 3.5 GPa.Loading method, shown in Figure 8, followed 3
patterns: -1/4 loading (-1,000mm from slab center), central loading (slab center), and +1/4 loading (+1,000mm from slab center), with a load of 80kN applied in each case. Displacement calculation was made at the same location were displacement were measured in the experiment at distances of -1,750mm, -1,000mm, -217mm, 0mm, 217mm, 1,000mm, 1750mm from the center.
4.2 Analysis Result
(1)Elastic Analysis under Static LoadingIn order to confirm the suitability of the model, 2
finite element method (FEM) analysis tools, Salome-Meca and CalculiX, were used to perform elastic analysis of the model under static loading condition.
Specimen length 4000
CLT bridge deck width 2000
平面図
590475 955 570 565 845
10
250 750 783 217 217 783 750 250
60
Figure 6: Positions of measurement devices.
0
1
2
3
4
5
6
7
0 50000 100000 150000 200000
Disp
lace
men
t （m
m）
Number of drives
Displacement 1
Displacement 6
Figure 7: Example of relationship between number of drives and displacement of CLT slab
Figure 8: Simple plan view of CLT slab
32-4
Firstly the open source FEM tool Salome-Meca was used for analysis. Salome-Meca is equipped with pre-processor and post-processor that can be graphically operated. Therefore, it is a relatively easy to use analysis tool with not much need for special arrangement of entry and output data or difficult command-based operation. Using these graphical functions, the above-mentioned analysis model was 3D modeled and mesh divided using tetrahedral elements.Results of analysis with Salome-Meca for -1/4,
central, and +1/4 loading points, are shown in Figures 9, 10, and 11, respectively, together with the experimental results.
The relative difference of the analytical value with the experimental value for the maximum displacement at the center was as follows: For -1/4 loading point, -0.1% for the 1-layer model and -4.1% for the 8-layer model. For central loading point, 2.5% for 1-layer model and -2.4% for 8-layer model. For +1/4 loading point, -2.7% for 1-layer model and -6.7% for 8-layer model. Therefore, generally the analytical results were similar to the experimental ones, and the 1-layer model and 8-layer model showed almost similar results. Regarding displacement of points away from the loading points, the analysis yielded the same symmetry as in the experimental results under center loading (Figure 10). Under ±1/4 loading points (Figure 9, and 11), although there was some divergence at the edge away from the loading points, the analysis showed identical distribution as the experimental results. Movement at the edges in the cases of ±1/4 loading points were not restricted. Therefore, the analytical results showing that displacement is getting smaller as we move away from the loading points towards the edges are logical. However, in the experiment, this movement of the edges was larger than anticipated, especially in the case of -1/4 loading point where the displacement at one edge exceeded the value of the displacement under the loading point. The reason for this can be attributed to the experiment arrangement where at the edges the CLT slab was supported by an H-beam using bolts with a rubber plate of 20mm thickness placed between the slab and the H-beam. The large displacement at the edge may then be due to effect of excessive deformation of the rubber.In comparison to the 8-layer model, the relative difference of the results for 1-layer model was a
maximum of 4.9%. Therefore, within elastic range of static loading, the 1-layer model is sufficient to model the slabs.
Figure 9: Comparison of 1-layer and 8-layerSalome 1/4 loading point
Figure 10: Comparison of 1-layer and 8-layerSalome center loading point
Figure 11: Comparison of 1-layer and 8-layerSalome 1/4 loading point
32-5
Furthermore, as the analysis model was symmetrical, analysis results of -1/4 and +1/4 loading points were identical. The experimental results also showed near symmetry. Therefore, comparison between only center loading point and -1/4 loading point will be considered in the analysis.
5 Test Constructions
Construction work of the experimental bridge was performed in the Ishikura-stream Prefectural Forest at Tazawa lake in Senboku City of Akita Prefecture. The bridge was constructed using a rough-terrain crane, capable of suspending 50 kN, to install steel H-beams and CLT bridge decks, where each erection was completed in one day. CLT bridge decks were brought to the site after wooden curbs (CLT) of 180 x 180 mm attached to both sides. The weights of bridge decks varied according to the implemented waterproofing treatment, where the heaviest CLT bridge decks that were wrapped with FRP sheets (Figure 12) weighed approximately 6 kN. This was lower than the 7 kN for the steel H-beams used as main girders, which confirms the superiority of CLT slabs with regards to transportation and installation work. The CLT bridge decks were secured to the main girders and cross-beams using lug screws (φ16 x 150 mm) on the lower surface of the CLT slabs from the bottom surface of the upper flange of the steel H-beams, as shown in Figure 13. Holes were drilled in the CLT slabs to enable penetration of the stud bolts (φ16 x 180 mm) welded on the upper surface of the flange of the steel H-beam, where driving lug screws from above the abutment was difficult. Then, the bolts were secured with washers and nuts. Furthermore, the bolt holes and the seat-bored sections in the vicinity of the nuts were filled with resin adhesive. The completed experimental bridge with asphalt pavement is shown in Figure 13.
Figure 13: Paved experimental bridge
6 Conclusions
CLT bridge decks are confirmed to have sufficient fatigue durability, based on the repeated load tests and the dynamic wheel-load tests.A simple analytical method was proposed to easily analyze the failure behavior of CLT slabs. The
elastic-plastic analysis of failure tests was done by simulating the timber compression strength and tensile strength. This was achieved in the analysis through giving yielding points to the elastic-plastic elements of the compressed side and tensioned side of the CLT panels model. The yielding points were estimated from experimental values of actual failure tests. The model was successful in simulating the experimental results of the load-displacement curve because the curve in the elastic plastic analysis showed gradual transformation to rigidity zone in a similar manner to the experimental curve. Future work includes the implementation of the investigative research intended for practical implementations, such as the verification of the effectiveness of the waterproofing for the CLT bridge decks.
7 References
Figure 12: Conditions prior to paving
32-6
Abrahamsen, Rune B., Nyløkken, Trond E., “Bridge deck rehabilitation using cross-laminated timber”, Proceedings of the International Conference Timber Bridges,T.Sasaki,Y.Ariyama.,”Development of waterproofing treatments for cross laminated timber slabs”, Paper Reports
on Use of Wood for Civil Engineering 16 (in Japanese).
32-7
