Development of earthquake bracing systems for multi-storey buildings using slender shear wall elements in cross-laminated timber (CLT)

Kamyar Tavoussi1, Wolfgang Winter2, Tamir Pixner3

1 Assistant Professor, Dr. DI,

2 Prof. DDI, Head of Department, 3 Research Assistant,

Vienna University of Technology, Department of Structural Design and Timber Engineering, Austria

Summary In modern timber buildings big openings in the façade are getting more common. The aim of the presented research project conducted together with a producer of CLT and the Austrian forest products laboratory (Holzforschung Austria) was to minimize the number and size of the shear walls provided to resist seismic loads by using stiff massive panels and non-conventional high performance anchorage systems in steel. The panels out of several layers of boards were glued using vacuum pressing equipment. These elements were optimized by the variation of the number of layers and their geometry. Several full size tests with different configuration of boards were carried out. The tests showed good results even for high horizontal cyclic loadings.

1. Introduction Europe has a long tradition of multi-storey timber based urban buildings. In the last century cement based buildings dominated completely the market in central Europe but for several years modern timber constructions are developed. CLT is a new product invented 15 years ago in Europe. Produced with heavy hydraulic presses under 5-7 kg/cm² it is used for structural elements. Early 2006 the company Mölltaler Ökohaus (1) and the Department of Structural Design and Timber Engineering at the Vienna University of Technology started a research project co-financed by the Austrian Found for applied Research (FFG). The aim was to develop an innovative earthquake-resistant timber structure using a small number of slender shear walls to make big openings possible. These shear walls should be produced in an economic production process using vacuum pressing equipment under a pressure of 1-2 kg/cm². One objective was the optimal board configuration of CLT walls for resistance against lateral forces.

Fig. 1 Production hall with vacuum press

The seismic design was based on a case study of a two-storey single-family house with asymmetrical bracing system and low torsion stiffness as worst case (Fig.2). A “propeller-system” out of horizontal arranged CLT panels for transmission of horizontal loads to the stiffening shear walls was developed (Fig.2,3). The remaining structure (ceiling and columns) which is not part of the bracing system could be built conventionally.

2. Design approach Worst case assumptions were used to determine the maximum load the panels should resist. The load-set-up for the seismic design refers to the design code EN 1990 (2). The seismic calculation was based on the Italian design code (3) (similar to EN 1998 (4)). According to the seismic zone 1 of Italy a horizontal ground-acceleration of 0,35 g was applied. A behaviour factor of q = 2 was assumed. For the time history analyses of the building seven artificial earthquakes according to EN 1998 were generated. Static and dynamic analyses were computed with framework software (RSTAB, Fa. Dlubal (5)) frequently used by structural engineers. The structural behaviour of the shear walls were modelled using a special configuration of quadratic frames with eccentric connections.

Fig. 4 Quadratic frame

The maximum representative lateral load under earthquake for one shear wall was approx. 140 kN, leading to vertical anchorage of approx. 400 kN. These anchor forces could not be transmitted by conventional methods. The design of a new anchor system became necessary. For estimation and comparison of static and dynamic resistance of different wall types against lateral loading following tests were planed and carried out.

Fig. 2 Plan view of ground floor and first floor

Fig. 3 3D – model of stiffening walls and “propeller-system”

3. Development and monotonic testing of 12 different panel types

3.1 Configuration of boards By combining criteria’s of optimal structural behaviour and production preferences 12 wall types were produced. Each wall type differed in the thickness, orientation and the angle of the laminated boards. The dimension of all test samples was 120 by 280 cm. The quality of the boards is C24 (W5 and W6 different) according to EN 338 (6).

W1

W2

W3

W4

W5

W6

W7

W8 (produced with hydraulic press)

W9

W10

W11

W12

Fig. 5 12 Panel types with thickness (mm) and angel of inclined laminated boards

[mm]

32 32

32

32

32

Σ 160

[mm] 32 32(66,8°) 32 32(66,8°)

32 Σ 160

[mm] 32 32(45°) 32 32(45°)

32 Σ 160

[mm] 32 32(±49,4°) 32 32(±49,4°)

32 Σ 160

[mm]

32

32

32 (C16)

32

32

Σ 160

[mm] 15 (OSB)

60/120 Joists

15 (OSB) Σ 150

[mm]

32

32(45°)

32

32(-45°)

32 Σ 160

[mm]

32(-45°)

32(45°)

40

32(45°)

32(-45°)

Σ 168

[mm]

32(-45°)

70(45°)

32(-45°) Σ 134

[mm]

32

32

32

32

32

Σ 160

[mm]

32(-45°)

77(45°)

32(-45°) Σ 141

[mm]

32(-45°)

32(45°)

47

32(45°)

32(-45°)

Σ 175

3.2 Testing setup The lateral force for the most stressed slender shear wall resulting from dynamic analysis was approx. 140 kN. The standard test equipment for shear walls used for timber frame houses are carried out with significant lower loads (10-20 kN). For the extreme high load a special test setup had to be developed.

Fig. 6 Testing apparatus

The hydraulic cylinder (green) which put the pressure on the end of the wall (brown) could reach a maximum of 200 kN of compression and 150 kN of tension. These maximum loads were applied. The existing frame (black) was modified by additional elements (red) for such high loading. The high tensile forces were anchored for the monotonic tests with steel bars (A=14 cm2) (blue). These bars were fixed with steel plates (blue) at the top of the wall and at the vertical additional steel element (red) on the other end.

Because of the high load the deformation of the test apparatus and the anchorage system had to be considered to determine the displacement of the wall itself.

Fig. 7 Contributions to total measured displacement

The measured displacement at the top of the wall was the result of 3 components: u total = u frame + u steel + u wood

Fig. 8 Points of measurement

01…Measurement of total Displacement 02…Measurement of Steel bar elongation 03+04…Measurement of Steel frame rotation

03

01 02 04

3.3 Test Results The tests were carried out referring to EN 594 (7).

Fig. 9 “F - u total” Diagram (simplified, phase of decreasing loading not shown)

Fig. 10a “F - u frame“Diagram (deformation due to steel frame rotation, mean value)

Fig.10b “F - u steel” Diagram (deformation due to steel bar elongation, mean value)

No fracture could be observed for the majority of tested walls. The best results (Fig. 9) were supplied by wall 2 and 3 with two inclined layers oriented in the same direction. These walls were asymmetric assembled and tested with the inclined layers in the direction of compression. The better performance of wall 2 demonstrates the efficiency of layers inclined in the direction of the diagonal of the wall (66,8°). Nevertheless the advantage of wall 2 compared to wall 3 was not that significant. In consideration of economic efficiency an angle of 45 degrees for inclined layers was chosen for further tests. Furthermore the asymmetric configuration leaded to the legitimate assumption of a weaker performance of the wall in the other direction. In the following tests inclined layers for walls were arranged in both directions. Wall 1 (vacuum pressed) and 8 (hydraulic pressed) with identical crossed board configuration but with different production methods showed comparable results, remarkably lower than the best panels with inclined boards (W2, W3). The poor performance of wall 10 and 11 due to the fracture of the overloaded diagonals pointed out the necessity of a vertical layer. Wall 9 and 12 with inclined layers in both directions did not perform better than the conventional cross laminated walls during monotonic tests. If the theoretically better performance of inclined layers would really exist, this had to be proved by cyclic testing.

4. Cyclic testing of 2 selected panel types

4.1 Configuration of boards

2 configuration types were selected for further cyclic tests. Wall 13 was identical with wall 1 except a thicker vertikal middle layer (40 mm instead of 32 mm). The board configuration of wall 14 and wall 15 was identical with wall 9.

W13

W14/15

Fig. 11 2 Panel types with thickness (mm) and angel of inclined laminated boards 4.2 Anchorage system The anchorage system consists of two steel channels (each: A=23cm2) placed on both sides of the shear wall. The idea was to transmit the tensile forces with a steel element to the top of the wall to avoid anchoring high tensile loads directly in the wood. For the steel to steel connection of the tension element to the anchor in the foundation special steel nails as driving fasteners (8) were used. This non-detachable connection executable on the construction site can provide high ductility in case of earthquake (Fig. 12).

Fig.12 Assembly of wall and anchorage system

[mm] 32 32 40 32

32 Σ 168

[mm] 32(-45°) 32(45°) 40 32(45°) 32(-45°) Σ 168

4.3 Test Results To describe and estimate the behaviour of such shear walls for an earth-quake scenario tests were carried out referring to ISO 16670 (9). Wall 14 and wall 15 were tested mirror-inverted (Fig. 14b, 15b).

Fig. 13a Wall 13: “F - u total” Diagram Fig. 13b Wall 13

Fig. 14a Wall 14: “F - u total” Diagram Fig. 14b Wall 14

Fig. 15a Wall 15: “F - u total” Diagram Fig. 15b Wall 15

The asymmetric hysteresis is due to the characteristics of the hydraulic cylinder which could not apply the same tension force as compression force. The hysteresis shows a decreasing stiffness with increasing cycling loading for wall 13. The stiffness of wall 14 and 15 remains almost similar with increasing cycling loading.

5. Conclusion

5.1 Shear wall The tests showed the similar efficiency of multilayer shear walls produced by vacuum presses in comparison to walls produced under high pressure with hydraulic presses. No fracture could be observed for the majority of tested walls. A vertical layer increases the load carrying capacity considerably and will be necessary for such high loads. Under cyclic loading perpendicular laminated timber panels decrease their stiffness more significant compared to inclined laminated timber panels.

5.2 Anchorage System

The damping capacity of the whole element (Shear wall in combination with the tension elements and the anchorage system) had been significant influenced by the plastification of the steel plates and the local plastification of the compressed wood (approx. 2,5 kN/cm² under max. lateral load of 200 kN).

Fig. 16 Plastification of the steel plates The tested anchorage system proved to be reliable for high tensile forces. The ductility could be increased by optimising the tension elements and the number of driven steel nails. This is advantageous in case of earthquakes.

6. References 1…Ökohaus Systembau GesmbH, Latzendorf 100, A - 9832 Stall 2…EN 1990: Eurocode – Basis of structural design, CEN, European Committee for

Standardization, Brussels 3…Normativa sismica Edifici-bozza aggiornata al 25/03/03 4…EN 1998-1: EUROCODE 8 – Design of Structures for Earthquake Resistance, - Part 1: General

Rules, seismic Action and Rules for Buildings, CEN, European Committee for Standardization, Brussels

5…Ing.-Software Dlubal GmbH, Am Zellweg 2, D-93464 Tiefenbach 6…EN 338: Structural timber–strength classes, CEN, European Committee for Standardization,

Brussels 7…EN 594:1995 Timber Structures – Test Methods – Racking Strength and Stiffness of Timber

Frame Wall Panels, CEN, European Committee for Standardization, Brussels 8...Hilti Austria Ges.m.b.H. Altmannsdorferstr. 165, Postfach 316, A-1231 Wien 9…ISO 16670:2003 International Standard, Timber Structures – Joints made with Mechanical

Fasteners – Quasi-static Reversed-cyclic Test Method, First Edition 2003-12-15.