Seven-storey case study

The example is a seven-storey superstructure which is an actual CLT plate building in the Ital-ian Alpine city of Bolzano (Fig. 7.14). The SFRS of that building was tested full-scale on a shake-table in Miki, Japan (Fig. 7.15). During the shake-table tests, the SFRS was subjected to 12 historical earthquake records scaled to between 50% and 100% of the observed peak ground accelerations. The real earthquake records were interspersed with artificial, stepped earthquake

Fig. 7.14: Seven-storey building in Bolzano

records of relatively low intensity to detect whether the system’s response changed during “real”

earthquakes. The artificial tests indicated limited changes in the lowest order, natural frequen-cies of the system, throughout the complete series of earthquakes. This finding was taken as  prima facie evidence of robustness and that any structural damage was localized. For the X-z and Y-z planes (as defined in Fig. 7.15) the lowest natural frequencies were 2.34 Hz and 3.32 Hz prior to application of earthquake records. After application of 10 earthquake records the fun-damental natural frequencies had changed to 1.95 Hz and 2.93 Hz respectively. The system also remained self-centring throughout (i.e. free from residual distortions). Application of seismic events much stronger than those believed to be credible for Bolzano or other parts of Italy dem-onstrates the practicality of designing and constructing systems capable of sustaining no more than superficial damage during design level earthquakes ( Recommendation 3 from Chapter 2).

Combining Eqs. (7.3) and (7.4):

F _b = g   _I a_gS 2.5___

q W   (7.5)

Application of Eq. (7.5) to the seven-storey SFRS is based on:

g  _I= 1.5, reflecting that the size and usual height for a timber building, W  = 290 tonne,

a_g = 0.82g = 0.82 × 9.81 = 8.04 m/s for X-axis direction (perpendicular to primary plan axis),

= 0.60g = 0.60 × 9.81 = 5.89 m/s for Y-axis direction (parallel to primary plan axis),

     1

Fig. 7.15: Seven-storey SFRS test 

7.6 EXAMPLE OF SEISMIC DESIGN PRACTICES  103 S  = 1.25, which corresponds to type B soil (very dense sand or very stiff clay),

q = 3.0.

The values of a_g used here correspond to peak values used during shake-table tests (i.e. not those applicable to design of the building in Bolzano).

The mode shapes for lowest order beam modes are assumed to result in linear increases in horizontal displacement amplitude from the bottom to the top of the SFRS, thus resulting in the horizontal equivalent lumped-static-forces values as follows:

 j

i = 1, 6 (i.e. six lumped masses represent the SFRS)

where i signifies the level of a lumped mass, and z_i is the height of a lumped mass of magnitude m_i above the base of the SFRS. Figure 7.16  shows the lumped mass values, associated z_i values, and resulting individual and accumulated F _i values at various storeys. The accumulated F _i values (^∑F _i values) at any level are the shear flows to be resisted by the SFRS below each lumped mass position.

As can be deduced from horizontal shear flows (^∑F _i values) in Fig. 7.16 , the largest structural demand in terms of horizontal shear to be resisted occurs at the base of the SFRS, in the nar-rower plane direction (X-z plane). This is a consequence of both the relative slenderness of the building in that plane, which leads to the longest actual beam mode period and the relatively low lengths of CLT walls that are available to act as shear walls in the X direction. Practices for determining force flows to individual wall panels and connections in the lower and other storeys follow the principle outlined in Subsection 7.5.1.4. Because floor plans were quite simple and reasonably close to symmetric (Fig. 7.15), the design results were not very sensitive to the type of force flow analysis used to convert F _i values into force flows in wall elements and their con-nections.

Figure 7.17  shows the types of shear connectors and anchors used to resist sliding and over-turning at elevated floor levels and at the base of the test system. The hardware used was a mixture of commercially available products and specialty anchors. The building in Bolzano has less heavy-duty connections, because the design peak ground acceleration is less. Detailed analysis of test information suggests that connections in the tested SFRS were overdesigned relative to actual demands on their capacities by around 30%. As is normal for timber build-ing superstructures, the connections were primary in definbuild-ing the behaviour of the SFRS, and their selection and detailed design was crucial to obtaining and exceeding the desired structural performance.

The maximum lateral drift observed during tests was in the order of  H  /80 for the X-z plane and H  /130 for the Y-z plane, where H was the total height of 23.5 m. However, as noted previ-ously, the earthquake records used in these tests included 100% of the ground accelerations observed during the actual earthquakes with one being the highly destructive Kobe earth-quake on January 17, 1995. The maximum peak ground acceleration for Italy is only 0.35 g

[113], and drift levels that might occur in a building like the one in Bolzano are in the order of  H  /200 or less. Therefore, lateral drift levels would actually be within the range of target limits typically specified by design codes. Drift predictions for the tested system obtained via

F _i

7.6 EXAMPLE OF SEISMIC DESIGN PRACTICES  105

relatively complex dynamic analysis were accurate, but a pushover analysis would also result in a satisfactory SFRS. Pushover analyses should account for flexural and shear deformation in CLT wall elements and any flexibility of connections that might lead to horizontal rela-tive sliding or vertical opening (relarela-tive rotations of storeys because of uplift). Although not elucidated in detail here, the design of this seven-storey SFRS’ elements was otherwise quite straightforward.

Finally, the admissibility of equivalent static force design depends upon the estimation of T ₁. Therefore, engineers are required to estimate the fundamental natural periods associated with each design orientation (e.g. planes X-z and Y-z in the discussed building). Usually this cannot be done with high accuracy, and opinions differ on acceptable methods. Consequently, given that experience with design and performance of CLT plate superstructure is so far quite limited, the surest fall-back approach is to base estimates on dynamic analysis. However, even if that is done, the details of the SFRS will not be known at first, and a seed estimate of T ₁ is required as the basis to begin iterative component sizing. For this, various empirical formulas have been suggested, with an example being:

T ₁ = 0.05 H ^0.75  (7.7)

For the tested seven-storey system, this yields the estimate of T ₁ = 0.05 × 23.5^0.75 = 0.53 seconds (fundamental frequency f ₁ = 1.9Hz). This compares with the actual measured T ₁ values of the system prior to the application of a series of earthquake records of 0.43 and 0.30 seconds for X-z and Y-z planes, respectively. Approximate formulas like Eq. (7.7) have limitations, and engineers need to satisfy themselves concerning what depth of analysis is appropriate for par-ticular projects.

IVALSA hold-down

        1         4         2

IVALSA hold-down

(d)

(a) (b) (c)

Fig. 7.17: Connection hardware in SFRS test: (a) IVALSA hold-down anchor used at bottom storey; (b) Simpson HTT22-hold-down used at elevated storeys; (c) shear connector used at upper storeys; (d) shear connector used at bottom storey

7.7 Additional comments

As implied by the types of applications discussed in this chapter, the construction of taller, multi-storey building superstructures using only timber plates, such as CLT, as primary load-bearing elements can handle certain niche situations. Examples are cases where the load trans-fers from above are widely dispersed such that neither floor platforms nor wall elements will be subjected to highly concentrated force flows on their horizontal surfaces. This means that the viable number of storeys is finite, as are viable room dimensions. Building superstructure geometries will largely determine the limits on the numbers of storeys and height, because the geometry along with building occupancy and building location controls the intensities of design stresses in wall and floor platform elements. Roughly speaking, the more squat a superstructure is the greater the number of storeys that will be feasible. Most probably timber plate buildings will be multi-occupant dwellings, hotels or non-mercantile business premises. The type of occu-pancy will bring with it certain restrictions (e.g. related to building functionality or occupant preferences). Therefore, storey limits will likely flow from the maximum heights that engineers can achieve within an architecturally defined footprint and not on unfettered limits of engineer-ing skill. What engineers will be able to achieve (with this particular type of construction sys-tem) will be governed by mechanical capabilities of timber panels and connection hardware and economics, once the architectural decisions are made. To make the engineering work beyond heights bounded in that way, structural systems will require fundamental alterations which mean solution are no longer classifiable as timber plate superstructures.

The rough rule of thumb is that for residential and similar building occupancies the minimum wall thickness for three-storey systems is about 90 mm. Each additional storey requires a 10 mm increase in wall thickness per storey (except for the top three floors). Therefore, a 14-storey all-timber plate building superstructure would require 200 mm thick walls in the lowest storeys.

Fig. 7.18: Fifteen-storey CLT building concept designed for Northern Italy (Dante O. Benini &

Partners, Milano)

Seizone Vista 3D

   5   0  m

 2 5 m

2  7  m 

Strutt ura in legno X-LAM Struttura in acciaio Struttura in cemento armato

7.7 ADDITIONAL COMMENTS 107 The practical limit for the number of storeys or height for timber plate buildings cannot be stated exactly. To date, buildings up to nine CLT plate superstructure storeys have been successfully constructed in diverse locations like Berlin (Germany), London (UK), Milano (Italy), and Melbourne (Australia). A 15-storey CLT plate superstructure design concept has been created for Northern Italy (Fig. 7.18). However, as the cutaway schematic for that building design shows, the building’s superstructure consists of a 13 storey CLT plate assembly that wraps around a structural framework at the building’s core and sits on top of two above ground RC plinth storeys. This reflects that, as has already been conjectured, using only timber plates becomes impractical at around that height. Economic and other non-technical considerations suggest that the maximum number of storeys that will ever be constructed from timber plates alone lies in the range 12 to 15 (40 m and 50 m).

Acknowledgements

This chapter incorporates ideas, information, diagrams, and photographs suppl ied by Professor Dr. Ario Ceccotti, who at the time of writing was the Director of the Italian National Research Council Trees and Timber Institute (CNR-IVALSA).

Thanks are also due to Dr. Andrea Polastri of CNR-IVALSA who assisted with finalization of this chapter.

109 Chapter

8

Example Project 1: Six-Storey Hybrid