Content uploaded by Hercend Mpidi Bita
Author content
All content in this area was uploaded by Hercend Mpidi Bita on Aug 12, 2018
Content may be subject to copyright.
LONG-TERM PERFORMANCE OF TIMBER CONCRETE COMPOSITE
FLOORS
Thomas Tannert1, Hercend Mpidi Bita2, Md Shahnewaz3, Md Mehdi Ebadi4, Adam Gerber5
ABSTRACT: Timber-concrete-composite (TCC) floors, composed of timber and concrete layers connected by a shear
connector are a successful example of hybrid structural components and are commonly used in practical applications.
The connection of the two components is usually achieved with mechanical fasteners where relative slip cannot be
prevented and the connection cannot be considered rigid. The growing availability of panel-type engineered wood
products (EWPs) offers versatility in terms of architectural expression and structural and building physics performance.
Preceding research determined the properties for a range of TCC connector systems in several EWPs using full-scale
short-term bending tests. In the research presented herein, nine TCC floor segments (one specimens of each previously
investigated configuration) were exposed to serviceability loads for approximately 2.5 years. During this time, the
environmental conditions and the deflections of each floor were monitored. After having been long-term loaded, the
floor segments were tested to failure. The results show an increase of deflection over time but neither bending stiffness,
load-carrying capacity nor vibration performance were impacted by the long-term loading. This research provides input
data to develop design guidance for TCC floors.
KEYWORDS:
Composite floors, engineered wood products, e
ffecti
ve bending stiffness
,
destructive testing
1 INTRODUCTION 1
Timber-concrete composite (TCC) systems are generally
comprised of a timber beam or plate connected to a
concrete topping. TCC offer significant improvements
over conventional reinforced concrete and light wood-
frame flooring systems with regard to strength, section
depth, stiffness, and vibration performance [1,2].The
growing availability of panel-type engineered wood
products (EWPs) offers designers greater versatility in
terms of architectural expression and structural and
building physics performance. EWPs are created by
bonding together graded strands, veneers or lumber into
larger structural elements resulting in stable products
with uniform properties. Such products include
Laminated Strand Lumber (LSL), Laminated Veneer
Lumber (LVL) and Cross-Laminated Timber (CLT). The
efficiency of TCCs depends largely on the properties of
the shear connection between them. Connectors range
from low-stiffness, where partial composite behaviour is
achieved, to rigid connectors.
1 Thomas Tannert, Wood Engineering, University of Northern
British Columbia, thomas.tannert@unbc.ca
2 Hercend Mpidi Bita, Wood Science, The University of British
Columbia, hercend_mb@alumni.ubc.ca
3 Md Shahnewaz, Civil Engineering, The University of British
Columbia, md.shahnewaz@alumni.ubc.ca
4 Md Mehdi Ebadi, Wood Engineering, University of Northern
British Columbia, MohammadMehdi.Ebadi@unbc.ca
5 Adam Gerber, Aspect Engineering, Vancouver, Canada,
adam@gerber.ca
The serviceability limit state is often the governing
design factor for TCC floors, [3,4]. Specific complexities
arise since deflections also increase over time due to
thermo-hygrometric variations of the environment and
time dependent behavior of the system: concrete
undergoes creep and drying shrinkage; timber exhibits
creep, mechano-sorptive creep and shrinking/swelling as
a function of moisture content and the connector
performance over time may change as well.
Preceding research at the University of British Columbia
focused on performance of different TCC configurations
in small-scale shear and full-scale bending tests [5,6].
Nine TCC configurations were selected and designed
according to the γ-method to achieve similar composite
efficiencies in the range of 90%. The results showed that
calculations according to the γ-method predict capacity,
stiffness and dynamic properties within a reasonable
degree of accuracy.
2 OBJECTIVE
The primary objective of the research presented herein
was to experimentally investigate the long-term
performance of multiple TCC systems. For this purpose,
nine full-size specimens were subjected to typical
climate yearly variations for approximately 2.5 years.
The secondary objective, which was contingent on the
first, was to determine the impact of long-term loading
on stiffness, ultimate load-bearing capacity and vibration
performance of the TCC systems.
3 EXPERIMENTAL INVESTIGATION
3.1 MATERIALS
Three EWP, namely LVL, LSL and CLT, and
commercial ready-mix concrete were used in the
research. Table 1 lists the relevant material mechanical
properties taken from the manufacturer’s specifications.
Table 1: Material properties
Fc
(MPa)
E
(MPa)
Fb
(MPa)
Ft
(MPa)
Concrete 30(1) -- -- --
LSL -- 10,685 33.3 20.4
LVL -- 13,790 37.6 18.6
CLT -- 9,500 11.8 5.5
(1) Average cylinder strength at time of testing = 45.5 MPa
Fc = specified compressive strength, E = modulus of elasticity,
F
b = specified bending strength,
F
t = specified tensile strength
Different connectors were combined with the EWP: self-
tapping screws (STS), (10mm diameter ASSY STS with
Canadian approval [7]), the proprietary glued-in steel
connector system Holz-Beton-Verbund (HBV) [8], and a
combination of STS and adhesive bond, see Table 2.
Insulation is sometimes desired for enhanced acoustic
performance of TCC floors; in this project, Foamular®
C-200 extruded polystyrene rigid insulation with a
compressive strength of 140kPa [9] was utilized. For the
systems using adhesive, Sikadur®32 Hi-Mod with a
shear strength after 14 days of 41MPa and a pot life of
approx. 30 minutes was used [10]. A plastic separation
layer, cut from rolls of clear 6mil polyethylene sheeting,
was placed between the timber and concrete elements.
Table 2: Test series overview
Series EWP Description
S1 LSL ASSY VG 10x240 installed at 30°
angle to grain (leff =140mm)
S3 CLT
S4 LSL
ASSY VG 10x240 pairs installed at
45° to grain through 25mm rigid
insulation (leff=90mm)
S5 LVL
ASSY VG 10x240 installed at 30°
angle to grain (leff =140mm) plus 2
continuous 50mm wide rows epoxy
applied wet just prior to casting
S6 LSL HBV mesh (90x1000), installed in
3mm wide saw kerf with 2C
polyurethane adhesive
S7 LVL
S8 CLT
S9 LVL
HBV mesh (120x1000), additional
25mm rigid insulation interlayer
between timber and concrete
3.2 SPECIMEN DESCRIPTION
Full-scale specimens were designed to exhibit composite
efficiencies in the range of 85-95% to allow comparing
the connector requirements for similar performance.
Calculations based on the γ-method were performed
[11]. Table 3 summarizes the panel configurations,
including the connector spacing. The parameters tc, tt,
and ti refer to concrete, timber and interlayer thickness
respectively, while rows refers to the number of rows of
fasteners across the width of the panel and s1 and s2 refer
to fastener spacing in the high and low shear zones of the
panel respectively, see Figure 2.
Table 3: Test series parameters
Series
EWP ti
(mm)
s1
(mm)
s2
(mm)
rows Figure
S1 LSL -- 150 300 3 2 (a)
S3 CLT -- 150 300 3 2 (a)
S4 LSL 25 300 -- 3 2 (b)
S5 LVL -- 300 -- 3 2 (c)
S6 LSL -- -- -- 2 2 (d)
S7 LVL -- -- -- 2 2 (d)
S8 CLT -- -- -- 2 2 (d)
S9 LVL 25 -- -- 2 2 (e)
3.3 BENDING TESTS
Two panels of each series were tested for strength and
stiffness under four-point bending. The panels spanned
5.8m between roller supports to simulate a true simply
supported condition. Load was applied to the panel at the
third points spread evenly across the width of the panel.
Roller bearings were located between the spreader beam
and the HSS tubes ensuring that the specimens were
entirely unrestrained by the test apparatus. A photo of
one mounted test specimen is shown in Figure 1. Loads
were recorded using a calibrated load cell while vertical
displacements were measured by a mid-span LVDT for
the stiffness tests and lasers at the mid and third points
for the entire test to failure. Four additional LVDTs were
installed to measure the relative slip between the
concrete and timber at each of the four corners directly
over the supports. Each of the full-scale specimens was
subjected to service level loadings from which the
effective bending stiffness (EIeff) was calculated.
Following the serviceability test, the panels were loaded
to failure at a constant displacement rate of 6mm/min.
Figure 1: Specimen subjected to bending testing
(a)
(b)
(c)
(d)
(e)
Figure 2: Panel layout (a) ASSY VG at 30◦ [S1-S3], (b) ASSY VG pairs at 45◦ through insulation [S4], (c) ASSY VG at 30◦ plus
SikaDur 32 adhesive [S5], (d) HBV mesh [S6-S8], (e) HBV mesh through insulation [S9]
3.4 VIBRATION TESTS
The dynamic performance of each TCC floor was
predicted based on established methods of mechanics
using the effective bending stiffness determined by the γ-
method [11]. The panels were subjected to dynamic
excitation from a heel strike impact and accelerations
were recorded using a digital accelerometer. The
required accelerometer has the resolution of 50 µg with
the sampling rate of 2000 Hz. From these acceleration
time histories, the fundamental frequency was obtained
using a Fast Fourier Transform (FFT). Before signal
processing, the recorded acceleration was pre-processed
including removing the offset and down-sampling.
Figure 4: Vibration tests of TCC test specimens
3.5 LONG-TERM TESTING
Nine TCC floor segments (one replicate from Series S1,
S3, S4, S6-S9 and two replicates from series S5) were
exposed to serviceability loads for approximately 2.5
years. The panels were located outside in a vertical
position, applying a uniform load on all nine floors by
means of four pre-stressed rods, see Figure 4. The
environmental conditions, the applied load (as the sum
of the four individual loads) and the resulting deflections
(as the average of two LVDT at mid-span of each floor
segment) were monitored with one recording every 30
minutes. Over the course of the long-term loading, the
rods were re-tightened six times, allowing for
comparisons of the change of the effective bending
stiffness over time. After completion of the long-term
loading, the panels were subjected to vibration and
quasi-static monotonic destructive bending tests,
identical to those described in section 3.3.
Figure 4: TCC test specimens under long-term loading
4 RESULTS
4.1 SHORT-TERM LOADING
Each of the full-scale specimens was subjected to
multiple loadings up to service level from which the
effective bending stiffness (EIeff) was calculated. The
results are summarized in Table 4 and compared with
-
method predictions.
Subsequently, the panels were tested to failure and the
maximum moment obtained are also listed in Table 4.
The typical load vs. displacement curves from one
specimens from each tests series are shown in Figure 5.
Figure 5: Typical load-displacement curves bending tests
The experimental work showed that TCC can fail in
either ductile or brittle modes, see Figure 6, based on
interpretation of the panel and connector load slip curves
along with visual observations during testing. Brittle
failure (concrete crushing and timber tensile fracture)
occurred at higher loads than more ductile failure modes
such as connector yielding or screw withdrawal.
Figure 6: Tested TCC panels: Brittle failure (top); connector
yielding (bottom)
4.2 LONG-TERM LOADING
The results from monitoring the ambient conditions
(temperature and relative humidity) are shown in Figure
7. The applied loads for the full period of long-term
loading are shown in Figure 8, while Figure 9 illustrates
in more detail the loss of load during the initial month of
loading.
Figure 7: Ambient conditions as averages over four hours for
duration of long-term loading
Figure 8: Applied load as average over four hours for duration
of long-term loading
Figure 9: Applied load each 30 minutes for the initial phase of
long-term loading
The measure average deflections for the full period of
long-term loading are shown in Figure 10, while Figure
11 illustrates in more detail the increase of deflection
during the initial month of loading. Finally, the load-
deformation curves obtained from the destructive testing
after completion of long-term loading are shown in
Figure 12.
Figure 10: Panel deflections as average over four hours for
duration of long-term loading
Figure 11: Panel deflections each 30 minutes for the initial
phase of long-term loading
Figure 12: Load-displacement curves from bending tests after
long-term loading
Table 4: Results summary
Series EIeff (1012 N*mm2) 1st Natural Freq. (Hz) Mult (kN*m/m) Failure
γ-Meth. Short Long1) γ-Meth.
Short Long Short Long mode
S1 (LVL + STS) 3.2 3.4 2.7 7.1 7.3 7.2 202 178 A
3.2 7.0 183
S3 (CLT + STS) 2.8 2.9 2.4 7.0 6.8 7.6 118 129 C
3.2 7.0 147
S4 (LSL + STS
+ insulation) 4.6 4.7 4.0 8.4 8.2 9.2 197 200 D
4.3 8.1 167
S5 (LVL + STS
+ adhesive) 3.9 4.2 3.8 7.9 7.7 8.5 165 156 E
3.9 4.0 7.9 8.8 140 157
S6 (LSL + HBV) 3.6 3.3 3.1 7.2 7.2 8.2 139 148 F
3.1 7.0 137
S7 (LVL + HBV) 3.9 3.9 3.5 7.9 7.9 8.4 139 152 F
4.2 8.1 141
S8 (CLT + HBV) 2.9 3.1 2.3 7.0 7.3 7.6 134 126 C
2.9 7.1 119
S9 (LVL + HBV
+ insulation) 5.9 5.7 5.5 9.7 8.9 9.9 143 145 F
6.1 9.6 141
Failure modes: A) brittle concrete crushing; C) brittle timber fracture; D) ductile screw withdrawal;
E) bond, screw withdrawal, & timber fracture; F) connector yielding
1) EIeff determined during destructive testing
5 DISCUSSION
5.1 SHORT-TERM PERFORMANCE
Good agreement was obtained between predicted overall
bending stiffness of panels and experimental results.
This indicates that stiffness values obtained in small-
scale shear tests as calculation input are an appropriate
characterization of the connector stiffness under realistic
bending conditions.
Where connector strength is the governing failure mode,
it is possible to introduce ductility to the system where it
would otherwise experience either concrete compression
or timber tensile failure. While both brittle failure modes
were observed during testing, it was observed that brittle
failures occurred at load levels in excess of four times
service load, demonstrating that these brittle failures are
unlikely to be the governing mode within the structure.
As a result, the failure mode is no longer of practical
importance since the TCC floor stands a very small
chance of being a weak link in the overall structural load
path. It should also be noted that even while some
connectors exhibited ductile behaviour at the connector
level, the global floor ductility was generally very low as
the slender panels have a large elastic deformation
capacity which was not greatly enhanced by the non-
linear contribution of the connectors.
5.2 LONG-TERM PERFORMANCE
Figure 7 clearly illustrates the yearly fluctuations in
ambient climate conditions with higher temperature and
lower relative humidity in summer compared to winter.
The applied load shown in Figure 8 mirrors this
behaviour: while there is a significant loss in load (cause
by relaxation) in winter when the EWPs absorbed
moisture, there is a slight gain of load during the summer
months when the EWPs were drying out.
Each of the full-scale specimens was subjected to
multiple re-loadings to keep to load level approximately
at service load. Figures 10 and 11 show the initial
increase in deflection caused by primary creep as well as
the increase of deflection after each re-loading. A
preliminary analysis of the results showed that the
increase in deflection is a function of the connector and
the material. Series S4 and S9, where pairs of crossed
STS were used show the smallest increase in deflection.
Series S3 and S8, where CLT was used, exhibited the
largest deflections and the largest increases in deflection.
The floor segments were tested to failure after long-term
loading. As shown in both Figure 12 and Table 4, the
ultimate load-carrying capacity of the floor segments
was not significantly impacted; the variation in ultimate
moment at failure (Mult) is on average 5% and not more
than 8% for any test series.
The effective bending stiffness (EIeff) determined during
the destructive testing, was on average 16% lower than
that determined on specimens that were not exposed to
long-term testing, varying between 3% for series S5
(adhesive bond in addition to STS) and around 28% for
series S3 and S8 (specimens using CLT).
The governing failure modes were unchanged comparing
the specimens that were exposed to long-term testing to
those that were tested shortly after fabrication.
5.3 VIBRATION PERFORMANCE
All fundamental frequencies for the TCC slabs range
between 6 and 10 Hz. Floors S4 and S9 have the highest
fundamental frequency (i.e. 8.4, 9.2, 9.7 and 9.9 Hz for
short-term and long-term loading), this attribute to the
presence of isolation layer which increases the moment
arm and consequently EIeff of the panel without changing
the mass considerably. The fundamental frequency of the
floors S5, S6, S7 was higher compared to floors S1 and
S3, also because of higher EIeff. It should be noticed that
the tested panels behave like shallow floors. Therefore,
lower stiffness and lower fundamental frequencies are
observed compared to panel with stiffening ribs.
The fundamental frequency of some panels is below 8
Hz. Eurocode 5 [12] mentions that special investigation
needs to be considered for the vibration design of floors
with the fundamental frequency below this threshold as
such floors might be prone to resonance with one of the
walking harmonics. Also, this frequency is the minimum
satisfactory natural frequency for lightweight floors [13].
One advantage of TCC floors is their higher mass
compared to lightweight floors which decreases the
amplitude of the response. Although using the thicker
concrete layers might reduce the fundamental frequency.
It should be noticed that damping plays an important role
in the vibration performance of timber floors [13] and
that further investigations are required to find the
damping ratio of TCC floor systems.
6 CONCLUSIONS
The work presented is an extract of the analytical and
experimental work performed of multi-phase program
with focus on: 1) the performance of TCC configurations
in small-scale shear tests; 2) the performance of eight
selected TCC configurations in full-scale bending and
vibration tests; 3) the performance of one specimen from
each configuration subjected to long-term servicability
loading; and 4) the performance of these specimens then
subjected to vibration tests and destructive bending tests.
The full-scale test specimens were designed according to
the γ-method to achieve similar composite efficiencies in
the range of 90%. Specimens were tested for elastic
stiffness under quasi-static bending and dynamic
properties were obtained using an accelerometer. The
results showed that calculations according to the γ-
method predict the stiffness and dynamic properties of
the panels within a reasonable degree of accuracy.
The vibration analyses showed that some TCC systems
exhibit a fundamental frequency below 8 Hz warranting
special attention for the vibration design. Adding a layer
of insulation increase the fundamental frequency of the
panel. Assuming constant mass, the panels with higher
effective bending stiffness provide higher fundamental
frequency. Further investigation needs to estimate the
damping ratio of such floors in buildings.
Results from the long-term tests showed an increase in
deformation over time. The long-term loading, however,
did not cause significant degradation, effective bending
stiffness decreased by on average 16% while load-
carrying capacity decreased by onloy 5% on average.
ACKNOWLEDGEMENTS
The research was supported by Forest Innovations
Investment through Wood First Program; NSERC
through Engage program. Weyerhaeuser, Louisiana
Pacific, Brisco Manufacturing; Structurlam Products,
MyTiCon Timber Connectors Inc, TiComTec GmbH,
Sika Canada, and Lafarge donated material. The help of
the technicians at FPInnovations and UBC is
appreciated.
REFERENCES
[1] Yeoh D, Fragiacomo M, De Franceschi M, Heng
Boon K. State of the art on timber- concrete
composite structures: Literature review. J Struct Eng
2011;137:1085-1095.
[2] Dias A, Skinner J, Crews K, Tannert T. Timber-
concrete-composites increasing the use of timber in
construction, European J of Wood and Wood
Products. 1016;74(3):443–451
[3] Ceccotti A, Fragiacomo M, Giordano S. Long-term
and collapse tests on a timber-concrete composite
beam with glued-in connection, Mater and Struct
2006;40:15-25.
[4] Tannert T, Endacott B, Brunner M, Vallée T. Long-
term performance of adhesively bonded timber-
concrete-composites. Int. Journal of Adhesion and
Adhesives. 2016;72: 51-61
[5] Gerber A, Tannert T. Timber-concrete composites
using flat-plate engineered wood products. ASCE
Structures Congress, April 2015, Portland, USA.
[6] Gerber G, Popovski M, Tannert T. Performance of
Timber-Concrete Composite Floors Using Flat-Plate
Engineered Wood Products. World Conference on
Timber Engineering, August 2016, Vienna, Austria
[7] CCMC 13677-R (2013) CCMC Evaluation Report
SWG ASSY® VG Plus and SWG ASSY® 3.0 Self-
Tapping Wood Screws. Canadian Construction
Materials Centre, Ottawa, Canada.
[8] Clouston P, Schreyer A (2008) Design and use of
wood–concrete composites. ASCE - practice
periodical on Struct design 13:167-174.
[9] Ciarlo S (2009) Allowable Loads on Owens Corning
Foam Products Technical Bulletin No.SC33, Owens
Corning
[10] Sika Canada Inc (2012) Sikadur 32 Hi-Mod Product
Data Sheet Version 1.
[11] EN 408 (2012) Timber structures – Structural timber
and glued laminated timber – Determination of
some physical and mechanical properties. CEN
European Committee for Standardization, Brussels.
[12] EN 1995, Eurocode 5 (2004) Design of timber
structures. Part 1-1: General- Common rules and
rules for buildings. CEN European Committee for
Standardization, Brussels, 2004.
[13] Weckendorf, J., Toratti, T., Smith, I. and Tannert T.
Vibration serviceability performance of timber
floors, Eur. J. Wood Wood Prod., 74(3), 353-367,
2015.
