Seismic performance of CLT buildings
SISMO project – Collaboration between Stora Enso, Simpson Strong Tie and the University of Minho
As CLT by Stora Enso is becoming a more and more popular building material in southern Europe and also in Asia, its seismic performance gets more relevant. Stora Enso has been participating in many different European research projects concerning seismic performance of buildings using CLT by Stora Enso. Parallel to these European research activities, Stora Enso also collaborates with universities and other companies to get further understanding of the issue. One of these collaborations was a SISMO project, started with Prof. Jorge Branco and Hélder Sousa from the University of Minho in Portugal.
The SISMO project consisted of several parts, where first a “state-of-the-art” study was conducted to find out which information was already available in terms of properties of the material, design and construction processes. Thereafter the project went over to a more “practical” part where both theoretical modelling of seismic design as well as laboratory tests for connector behaviour were conducted, the latter part in collaboration with Simpson Strong Tie, a company specialized in connectors for timber structures.
In the laboratory, both the angle brackets and hold-downs were tested for their static and cyclic performance. Angle brackets were tested on both lateral and uplift directions, whereas hold-downs were only tested in uplift direction. Tests were made with two different boundary conditions: a) rigid, with a steel metal plate; b) with a timber support. The tested connectors correspond to AE116 angle brackets and HTT22 hold-downs from Simpson Strong Tie.
Prior to this experimental campaign, a numerical model was calibrated with a smaller sample data. Regarding the modelling of the static tests, it was found that the SAWS model allowed for a suitable prediction of the maximum load of the load-slip behaviour compared to the experimental results.
Considering the variation of length of the wall, the analysis was divided into two categories — type W1 and type W2 depending on the type of connectors used in the wall models. In the first walls (W1), only angle brackets were used as connectors between the wall and the foundation, while in the second (W2), hold-downs were used in the corners and angle brackets were used in the interior. In both wall models, the corner connections are located at 0.08 m from the edges of the wall panel and the interior connectors are almost equally spaced. A vertical loading of 20.8 kN/m was applied on all models.
The results of the models evidenced that, for walls of different lengths, the ductility of walls tends to decrease in the presence of hold-downs, while the lateral load capacity increases significantly.
For walls with a length of 1 m, the displacement at which maximum load was attained was around 147 mm for walls W1 and 90 mm in the case of W2. This shows that these walls were overdesigned. Up to a length of 3 m, the optimum number of connectors between the wall and foundation was found to be four. The increase in the capacities of walls with a length of 3 m with five connectors compared with the corresponding walls with four connectors for walls W1 and W2 was 6.5% and 4% respectively. For walls of 4 m and 5 m, the capacity of walls with five connectors showed a significant increase in the maximum load, initial stiffness as well as an improvement in the post peak behaviour. For walls with a length of 6 m, the use of at least five connectors is recommended.
Considering the variation of position of the connectors, the analysis was divided into two categories—according to the type of connectors being used (type A: 5 angle brackets; type B: 2 hold-downs with 3 angle brackets)—and subdivided in three configurations. The configurations take into account the positions of the angle brackets such that:
- Configuration E: the spacing between the connectors is approximately equal;
- Configuration C: the angle brackets in the interior of the wall are concentrated in the central portion of the wall;
- Configuration S: there is only one angle bracket in the central portion of the wall and the other interior connectors are closer to the sides of the wall.
Two different wall lengths have been considered (4 m and 5 m), with a constant height of 2.5 m and a vertical loading of 20.8 kN/m acting on the top of the wall.
The results of the parametric analysis on the location of the connectors evidence that walls of Type B have a mean increase of load capacity of 16.9% compared to walls of Type A. Also it is seen that for both types (A and B), there is a mean increase of approximately 23% between walls lengths of 4 m and 5 m. It is also seen that the connectors should not be concentrated in the central portion of the wall, whereas, for this kind of loading, the wall capacity is maximized when the connectors are located closer to the sides of the wall.
The connector “experiment” also provides a great amount of data available which should help optimise the use of connectors in a CLT by Stora Enso building. This is very important especially, when designing multi-storey timber buildings in seismic areas. A seismic design guideline has been developed within the context of this project. However, it should be aimed at enabling the engineers to design their CLT by Stora Enso building with an online-tool such as CLTengineer for instance, which is being provided by Stora Enso for basic structural design.
In order to get the theoretical models as close to reality as possible, more tests should be carried out—especially in a larger scale. As far as modelling analysis is concerned, the tests that have been made have proven to produce substantial information as they allow us to obtain significant damage in a controlled laboratory environment and enable us to assess the damage path. These tests would also allow us to analyse the influence of different combinations of connectors (type, number and location) as well as the influence of different CLT by Stora Enso panels (thickness, cross-section layout and direction).
The work in seismic design of timber structures in Europe is still in a very early phase, partly due to the fact that the timber industry and most timber buildings are located in North and Central Europe, where seismic actions do not affect the structures. However, as southern Europe and other countries with high probability of earthquakes are starting to design higher buildings in CLT by Stora Enso, the seismic performance will become more relevant. Stora Enso will continue the work and invites different stakeholders to take part in research and exchange of information.