Behavior of AAC Reinforced Thin-Panel Floor

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

https://www.researchgate.

net/profile/Jennifer_Tanner4/
publication/330235785_Behavior_of_Autoclaved_Aerated_Concrete_Reinforced_Thin-
Panel_Floor_Sheathing_Elements/links/5cffc3e2299bf13a384d63c5/Behavior-of-Autoclaved-Aerated-
Concrete-Reinforced-Thin-Panel-Floor-Sheathing-Elements.pdf
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/330235785

Behavior of Autoclaved Aerated Concrete Reinforced Thin-Panel Floor


Sheathing Elements

Article  in  Aci Structural Journal · January 2019


DOI: 10.14359/51710863

CITATION READS
1 104

3 authors:

Caleb Jennings Nicholas A. Owen


University of Wyoming
3 PUBLICATIONS   1 CITATION   
1 PUBLICATION   1 CITATION   
SEE PROFILE
SEE PROFILE

Jennifer Tanner
University of Wyoming
34 PUBLICATIONS   116 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Jennifer Tanner on 11 June 2019.

The user has requested enhancement of the downloaded file.


ACI STRUCTURAL JOURNAL TECHNICAL PAPER
Title No. 116-S12

Behavior of Autoclaved Aerated Concrete Reinforced


Thin‑Panel Floor Sheathing Elements
by Caleb Jennings, Nicholas A. Owen, and Jennifer Eisenhauer Tanner

This paper presents findings from laboratory tests on reinforced Very few studies have been conducted on thin precast
autoclaved aerated concrete (AAC) thin-panel floor diaphragms. panel systems on either steel or wood joist systems. Yeoh
Data were generated through targeted AAC material property tests, et al. (2009) and Judd and Fonseca (2002) evaluated the
out-of-plane testing of reinforced AAC panels, and in-plane testing strength of panels resting on timber joists. Another experi-
of light-frame AAC floor diaphragm assemblages. A group of six
mental study was conducted by Easterling and Porter (1994)
reinforced panels resting on joists was tested to determine a design
considering moderately thick concrete panels supported by
load for one-way floor systems subjected to gravity loads. Results
of 10 floor diaphragm specimens were used to develop a strut- steel joists.
and-tie model that provides a reasonable lower-bound approxima- Research pertaining to basic material properties of plain
tion of the ultimate in-plane capacity of one-way floor systems with AAC exists. For example, Valore (1954) compiled and
AAC sheathing. The testing provided reliable design values and the compared AAC material test data from European-, Russian-,
strut-and-tie model fits well with the experimental test data. and American-based studies. Data from these studies provided
density, compressive strength, flexural strength, modulus
Keywords: autoclaved aerated concrete (AAC); diaphragms; experimental; of elasticity, and drying shrinkage data. Similarly, Short
strut-and-tie; thin panel.
and Kinniburgh (1961) compared and analyzed combined
test results for AAC compressive strength and modulus
INTRODUCTION AND BACKGROUND
of rupture from several European laboratories. Soon after
Developed in 1923 by Swedish architect Johan Axel
AAC was introduced into the United States, Matthys and
Eriksson, autoclaved aerated concrete (AAC) is a precast,
Nelson (1999) conducted targeted material testing. This was
lightweight, cellular, cementitious building material that
further developed by Fouad et al. (2005) and Dembowski
simultaneously provides structure, thermal insulation, sound
(2001), who evaluated three different grades of plain AAC
insulation, dimensional stability, and resistance to fire,
units supplied by different manufacturers—in particular,
mold, and decay. Since its development over 90 years ago,
compressive strength, flexural strength, shear strength, and
AAC has progressed into a reliable building material among
modulus of elasticity; those results influenced the develop-
many European countries (Coduto and McDonough 2012).
ment of several ASTM testing standards for AAC (Fouad
Consequently, many of the laboratory studies performed on
et al. 2005). The work was synthesized by Argudo (2003) and
AAC have been European-based. After AAC was introduced
incorporated into a design guide for AAC (ACI Committee
into the United States in the early 1990s, relevant testing
523 2009). Chen et al. (2013) conducted an investigative
standards and design provisions were developed and AAC
study in China on the material properties of plain AAC and
gained recognition as a viable construction material. In the
confirmed that ACI 523.5 predictive equations are accurate.
United States, AAC has promise in low-rise commercial and
residential structures. As new AAC products come to the
RESEARCH SIGNIFICANCE
U.S. market, it is necessary to quantify their behavior char-
In this study, material properties and capacities of floor
acteristics and strength attributes through laboratory testing.
systems are determined experimentally. To the best of the
Although past researchers have performed destructive
authors’ knowledge, data does not exist to evaluate thin
tests on AAC structural floor diaphragm assemblages, they
AAC panels over a one-way light-gauge steel or wood joist
used AAC panels surrounded by a grout ring. Griebenow
system. As a result, this paper presents entirely new find-
et al. (1989) performed monotonic, in-plane load testing
ings regarding the behavior of AAC thin-panel construction
of floor diaphragms constructed of reinforced AAC panels
subjected to out-of-plane and in-plane loading. After testing,
connected at the joints and encased along the diaphragm
mechanical models to explain the behavior were developed.
perimeter by a reinforced bond beam. This research was
Structural engineers will benefit from using the models when
complimented by Parker (2005), who tested a series of floor
laboratory testing has not yet been performed.
diaphragms constructed in a similar manner and subjected
to in-plane monotonic loads to failure. Parker showed that
a strut-and-tie model provided reasonable lower-bound
approximations for the ultimate capacity of AAC floor ACI Structural Journal, V. 116, No. 1, January 2019.
MS No. S-2017-452.R2, doi: 10.14359/51710863, was received December 19,
diaphragms. Storlie (2009) extended Parker’s research by 2017, and reviewed under Institute publication policies. Copyright © 2019, American
investigating the effects of cyclic loads on similar AAC Concrete Institute. All rights reserved, including the making of copies unless
permission is obtained from the copyright proprietors. Pertinent discussion including
floor diaphragms. author’s closure, if any, will be published ten months from this journal’s date if the
discussion is received within four months of the paper’s print publication.

ACI Structural Journal/January 2019 129


Fig. 1—Fastener shear test: (a) using wood backup before testing; (b) after screw deformation and AAC failure; and (c) steel
specimens after testing.
Fastener shear testing
Five fastener shear specimens, each consisting of a 3.5 x
3 x 6 in. (90 x 75 x 150 mm) AAC block fastened to a 3.5 x
1.5 x 8 in. (90 x 40 x 200 mm) wood block by means of a
single No. 12, 4.5 in. (110 mm) long, 4 in. (100 mm) thread
length screw were tested to failure. This fastener size and
panel thickness would be typically used in AAC thin-panel
construction. Because there is no governing ASTM standard
for fastener shear tests using AAC and wood, an adaptation
of ASTM D1761-12 was designed to apply loads through the
longitudinal centerline axis of the AAC block and backup
block, respectively. Monotonic compressive loads were
applied until crushing of the AAC block or shear of the
fastener occurred, as illustrated in Fig. 1 and 2. The average
maximum load was 480 lb (2.1 kN) with a coefficient of
variation (COV) of 0.071. Failures initiated with localized
crushing of the AAC, leading to splitting of the AAC block.
A 5.5 in. (90 mm) wide AAC block was connected to a
5.5 x 1.5 x 8 in. (90 x 40 x 200 mm) standard steel joist
section (550S162-54) by means of a single No. 12, 3.75 in.
(95 mm) long, 3 in. (75 mm) thread length screw. After five
tests were completed, the average maximum load was 530 lb
(2.4 kN) with a COV of 0.096. The primary failure mecha-
nism was vertical splitting of the AAC as shown in Fig. 1(c)
Fig. 2—Elevation view of fastener shear test.
(Jennings 2016).
MATERIALS AND METHODS
The testing performed in this project began with material Fastener pull-through testing
testing and concluded with system-level testing. It resulted A total of eight fastener pull-through specimens, each
in an Engineering Service Report for AAC floor diaphragms consisting of a 12 x 12 x 1-7/8 in. (300 x 300 x 50 mm) AAC
supported by one-way steel or timber joists (UES 350 2017). panel sections attached to a 1.5 x 5.5 x 7.5 in. (40 x 140 x
190 mm) wood block by means of a single fastener were
AAC material property testing tested to failure. Specimen assemblies were constructed by
To quantify the basic material properties of the AAC under aligning the centers of the 12 x 12 in. (300 x 300 mm) face of
investigation, material tests on plain AAC were performed the AAC panel section with the 1.5 x 7.5 in. (40 x 190 mm)
in accordance with ASTM C1692 and C1693. Plain AAC face of the wood block and securing the two pieces by means
specimens had an average compressive strength of 610 psi of a single No. 12, 3-1/4 in. (80 mm) long screw fastener
(4.2 MPa), modulus of elasticity of 340 ksi (2.3 GPa), inserted transverse to the 12 x 12 in. (300 x 300 mm) face of
and modulus of rupture of 200 psi (1.4 MPa). The drying the AAC block.
shrinkage value of 0.0071% is less than the maximum limit The specimen was supported by a test fixture consisting
of 0.02% allowed by ASTM C1693. Data from this study of two 12 in. (300 mm) channel sections bolted together at
agreed with predictive equations of ACI 523.4 (2009) and the corners (Fig. 3 and 4). The bottom channel served as
a complete record of the test data, procedures, and analyses the support platform for the AAC block and featured an
are included in Jennings (2016). 8 in. (200 mm) diameter hole at its center to allow the wood
block to pass through; this ensured that the volume of AAC
surrounding the screw was not confined against a cone-like
pull-through failure. Two steel plates bolted to the wood

130 ACI Structural Journal/January 2019


Fig. 3—Fastener pull-through test: (a) general setup before testing; and (b) during testing.
block provided a connection point for a U-bolt to eye-bolt
attachment, which ensured concentric load application along
the axis of the fastener. The average value of fastener pull-
though for a single No. 12 fastener was 340 lbf (1.51 kN).
The 95% confidence interval for fastener pull-through spec-
imens is 310 to 360 lbf (1.4 to 1.6 kN).

Individual panel transverse gravity load testing


Panels have nominal dimensions of 3 x 24 x 72 in. (0.076 x
0.61 x 1.83 m). All panels were reinforced with one layer of
steel mesh, as illustrated in Fig. 5.
To aid in easy specimen identification, each specimen
name contains the following information: test type, joist
material, and replicate number. For example, the specimen
identifier GP-W-1 indicates a gravity panel test, wood joist
subsystem, and the first replicate.
To quantify the behavior of AAC panels subjected to
gravity loads, structural testing was conducted on a three-
span segment. Each panel rested atop four subsystem
elements consisting of either dimensional lumber or light-
gauge steel sections spaced at 24 in. (0.61 m) on center. Wood
subsystem elements consisted of 2 x 6 nominal dimensional
lumber or lightweight steel joist sections (550S150-54), with Fig. 4—Elevation of fastener pull-through test.
a specified yield strength of 33 ksi (228 MPa). Lateral floor diaphragm testing
A steel beam and reaction frame provided support for All diaphragm specimens had an aspect ratio of 1.5 and
three hydraulic actuators positioned above the centers of specimens featured 3 in. (76 mm) thick panels as sheathing
panel clear spans. Aluminum blocks served as intermediate elements. Panels were laid out in a one-third running
supports and distributed actuator loads across the panel bond configuration and attached using mechanical screw
width (Fig. 6). Hydraulic load actuators supplied monotoni- fasteners to an underlying framing system. Thin-bed mortar
cally increasing loads that were applied until target span bonded panels at both longitudinal and end faces. Four
deflections of L/1080, L/540, and L/360 were reached. At specimens were constructed using wood joists while the
these deflections, loads were held for 5 minutes to monitor remaining six were constructed using light-gauge steel joists
short-term deflection changes. Upon reaching a deflection of (Fig. 7). Mechanical fasteners (No. 12) secured AAC panels
L/360, the corresponding load was held and monitored for to the wood or steel frame. Screws were oriented at a diag-
24 hours to determine the deflection characteristics of panels onal where panel edges met over a joist member. Wood-
subjected to sustained loads. Finally, panels were loaded to frame substructures were constructed with 2 x 6 nominal
failure to determine ultimate load capacity. Linear displace- dimensional lumber. Interior joists were spaced at 24 in.
ment transducers measured displacements at the centers of (0.61 m) on center. Framing members were doubled along
panel spans and at interior panel supports. A pressure trans- the diaphragm perimeter.
ducer recorded pressure supplied to the hydraulic actuators. Steel framing substructures consisted of standard steel
stud sections (550S162-54). Joist members were attached to
chord members by angle brackets and self-drilling screws.

ACI Structural Journal/January 2019 131


As with gravity panel specimens, diaphragm specimens tests differed by more than 10%, a third test was performed
were given a unique specimen identifier, where the first vari- and the average of the lower of the two test results was
able indicates subsystem material (Wood or Steel), followed reported. To estimate the beneficial effect of mortared joints,
by the load orientation (Perpendicular [⊥] or Parallel [II] to the test program included one diaphragm specimen (S-II-
joists), specimen depth, length, and replicate number. 8-12-4) constructed without the use of mortar. Although
Floor diaphragm specimens were subjected to mono- mortar is typically used, the research team wanted to deter-
tonic loads to the point at which the specimen load capacity mine its contribution.
either fell below 85% of the maximum load, or a minimum
midspan deflection of 2.5 in. (63 mm) was reached. Loads RESULTS AND DISCUSSION
were induced by two hydraulic load actuators positioned at Individual panel transverse gravity test results
1/3 points along the diaphragm edge, as illustrated in Fig. 6. Panels were able to withstand span loads ranging from
A strip of plywood was used to raise each side of the spec- 760 to 1060 lbf (3.38 to 4.71 kN) prior to initial damage.
imen parallel to the loading direction to eliminate friction Flexural cracking occurred first at the midspan of exterior
between the laboratory floor and the specimen. Deflec- panel spans and shortly thereafter at the interior supports.
tions were measured at several locations along the side of Following cracking, ultimate capacity was governed by flex-
diaphragm nearest the supports using linear potentiometers. ural failure in one of the exterior panel spans. Most tests were
A minimum of two diaphragms of each type (wood or terminated after significant panel span deflections between
steel) and load orientation (loading parallel or perpendicular 2 and 3 in. (50 and 75 mm) but before any appreciable loss
to joists) was conducted. If the ultimate loads of the first two in panel load-carrying capacity, as illustrated in the force-
displacement plots of Fig. 8. Table 1 summarizes critical
test results; reported values reflect load per actuator or panel
span. As described in the methods section, extended 24-hour
loading periods give an indication of panel deflection char-
acteristics under sustained load. After this duration, midspan
deflections ranged from negligible to 0.029 in. (0.75 mm),
suggesting that sustained gravity loads are not a problem.

Discussion of individual panel transverse


gravity test
For design purposes, it is typical to express sheathing
load capacity in terms of a distributed area load. Because
the gravity panel tests used line loads rather than area loads,
an equivalent area load was calculated that would induce the
same maximum stresses as those induced by experimental
Fig. 5—AAC panels used in construction of floor diaphragm
line loads. By matching the maximum moments for line and
specimens: (a) plan view; (b) elevation of panel end; and (c)
distributed area loads and ensuring that the resulting panel
elevation of panel length with scross = 13.8 in. (0.35 m) and
shear stresses remained within shear design limits based on
all bar diameters of 0.16 in. = 4.1 mm.

Fig. 6—(a) Panel gravity test apparatus; and (b) schematic of test setup. (Note: LP is linear potentiometer.)
Table 1—Panel test results (strength attributes)
Cracking load, lbf Service load, Ultimate load, Cracking load, Service load, Ultimate load,
Specimen ID (kN) lbf (kN) lbf (kN) Specimen ID lbf (kN) lbf (kN) lbf (kN)
GP-W-1 1060 (4.71) 1140 (5.07) 1660 (7.38) GP-S-1 760 (3.38) 965 (4.29) 1930 (8.58)
GP-W-2 1020 (4.54) 1180 (5.25) 1475 (6.56) GP-S-2 760 (3.38) 920 (4.09) 1950 (8.67)
GP-W-3 950 (4.23) 920 (4.09) 1890 (8.37) GP-S-3 905 (4.03) 865 (3.85) 1830 (8.14)
Average 1010 (4.50) 1080 (4.80) 1672 (7.44) Average 808 (3.60) 917 (4.08) 1903 (4.47)
COV 0.055 0.130 0.121 COV 0.104 0.055 0.034

132 ACI Structural Journal/January 2019


Fig. 7—As-built floor diaphragm test apparatus: (a) steel framing configured to load specimens parallel to joists, 8 x 12 ft
(2.4 x 3.6 m); and (b) wood framing configured to load specimens perpendicular to joists, 12 x 8 ft (3.6 x 2.4 m).

Fig. 8—Load-deflection paths for failure-controlling spans.


ACI 523-R.4 (2009), a maximum distributed area load of
425 lbf/ft2 (20.4 kN/m2) was achieved.
Fig. 9—Crack sequence of floor diaphragm specimen
Lateral floor diaphragm test results W-II-8-12-1.
Due to the interlocking nature of the running bond configu- Table 2—Reported ultimate strength values of
ration of AAC panels in diaphragms loaded parallel to joists, diaphragm specimen types
a direct load path existed from the points of load applica-
tion to the specimen reaction supports. Figure 9 illustrates Number of replicates Average ultimate load,
Specimen type tested kip (kN)
a typical sequence of damage events for diaphragms loaded
parallel to joists. A full description of the damages labeled W-II-8-12 2 16.3 (72.5)
in Fig. 9 and the approximate loads and deflections is avail- S-II-8-12 3*
18.6 (82.7)
able in Jennings (2016). The resulting failure in diaphragms W-⊥-8-12 2 15.5 (68.9)
loaded parallel to joists was governed by the compressive
S-⊥-8-12 2 20.9 (93.0)
strength of the AAC at reaction supports.
Unlike diaphragms loaded parallel to joists, instead of *
Non-mortared Specimen S-II-8-12-4 was not considered in reporting average
ultimate load.
relying on direct edge bearing between panels, diaphragms
loaded perpendicular to joists initially relied on mortar-joint Load-deflection plots of diaphragms loaded parallel to
shear strength to transfer forces from interior panels to exte- joists and perpendicular to joists are presented in Fig. 10 and
rior panels and ultimately to specimen supports. Diaphragms Fig. 11, respectively.
loaded perpendicular to joists underwent midspan deflec- Table 2 gives the reported average ultimate loads for the
tions exceeding 2.5 in. (65 mm), with some loaded in excess two lowest-strength diaphragms of each specimen type and
of 3 or 5 in. (75 or 130 mm) with no pronounced degradation load orientation. Diaphragms loaded parallel to joists expe-
in load capacity. After longitudinal mortar joint cracking, rienced a general increase in load until testing was stopped
loads transitioned to underlying joist members and failure at displacements exceeding 3 in. (75 mm). Specimen S-II-
was governed by the composite action between AAC panels, 8-12-4 was tested without mortar bonding the AAC panels.
panel fasteners, and the flexural capacity of joist members. As a result, the load capacity was only 75% of the remaining
Figure 9 illustrates the sequence of damage events for a floor three specimens. Likewise, the initial stiffness is signifi-
diaphragm specimen loaded perpendicular to joists; a full cantly reduced because the AAC panels are segmented and
description is available in Jennings (2016). do not function as one specimen.

ACI Structural Journal/January 2019 133


forces on perimeter cross wires, which in turn prevent the
inclined compression struts from translating in the global
x-direction. Cross wires are attached to longitudinal wires
which transfer strut forces to subsequent cross wires within
the panel. These subsequent cross wires are confined by and
bear directly on the surrounding AAC. The panel itself is
anchored to the substructure frame by mechanical screw
fasteners, which prevent the panel from moving outward
under compression strut forces. Ultimately, the global
x-component of the inclined strut forces is transferred to
the tension chord of the underlying frame, as shown by the
heavy dashed green line. This load path is further described
Fig. 10—Plot of loads and midspan deflections for through a series of free body diagrams presented in Fig. 13.
diaphragms loaded parallel to joists. To develop strength predictions for floor diaphragms
using the strut-and-tie model described herein, several
failure modes must be considered. The failure modes consid-
ered in the analysis of the proposed strut-and-tie model are
as follows:
1. Crushing of the AAC under load bearing plates or
within compression struts.
2. Crushing of the AAC at perimeter cross-wire bearing
interface.
3. Screw anchorage failure.
4. Crushing of the AAC at reaction supports.
The observed failure mode for experimental floor
diaphragm test specimens loaded parallel to joists was
always controlled by crushing of the AAC at reaction
Fig. 11—Plot of loads and midspan deflections for
supports (failure mode 4). To verify the model, however, the
diaphragms loaded perpendicular to joists. Specimen iden-
capacity as governed by each failure mode was calculated;
tifier “L = ⊥”.
the limiting capacity would be the predicted failure mode.
Strut-and-tie model for diaphragms loaded parallel The model assumes all load is transferred directly to the
to joists AAC based on its relative stiffness with respect to the frame.
Conventional strut-and-tie models feature concrete Failure Mode 1—Theoretically, Failure Mode 1 (crushing
compression struts and one or more continuous tension ties of the AAC under load bearing plates) occurs when the exter-
(typically steel reinforcement) spanning the width of the nally applied load, P/2, divided by the bearing area of load
structural element or assembly. Although the AAC light- distribution plates exceeds the tested compressive strength
frame floor diaphragms discussed herein feature longitu- of the AAC, fAAC. Using this method, the externally applied
dinal panel reinforcement in the form of welded wire mesh load required to cause Failure Mode 1 was determined to be
(Fig. 5), the wires cannot be solely relied upon to transfer P1 = 18.3 kip (81.3 kN).
tensile forces between reaction nodes because of discontinu- Failure Mode 2—Upon the occurrence of transverse panel
ities at panel boundaries. cracking, longitudinal tension forces within an individual
Despite reinforcement discontinuities, a load path still AAC panel are carried by wire mesh reinforcement. Unlike
exists to transfer tension forces between diaphragm reaction typical steel reinforcing bars which feature raised lugs to
nodes. This load path relies on contributions from several provide mechanical bond between the steel and concrete
components of the floor diaphragm, including wire mesh along the length of the bar, welded mesh wires are smooth
reinforcement; AAC; screws; joists; and, finally, the frame (ACI Committee 523 2009). Because the wires are smooth,
tension chord. Before providing a detailed account of the adhesive bond along their length was assumed to be negli-
load path, a global schematic of the proposed strut-and-tie gible, and therefore ignored. As illustrated in Fig. 14, wire
model for light-frame floor diaphragms is presented in mesh reinforcement relies on AAC bearing resistance at
Fig. 12. Compression struts and nodal zones are shown as cross-wire interfaces to develop tensile forces within longi-
light gray and dark gray shaded regions, respectively. The tudinal wires. AAC bearing capacity at cross-wire interfaces
y-component of inclined strut forces is resisted by reaction is a function of the diameter of the cross wires, dcross, and
supports at the diaphragm corners, while the x-component of the compressive strength of the AAC, fAAC. Although AAC
the inclined strut forces is resisted by a complex load path. crushing initially occurs at cross-wire interfaces nearest the
The inclined compression struts flare out along the most highly stressed sections of longitudinal wires, with
diaphragm sides to use anchorage provided by perimeter increasing damage, stresses redistribute to subsequent cross
cross wires within the AAC panels. The AAC that makes wires until, eventually, all cross wires are equally engaged,
up the inclined compression struts exerts transverse bearing resulting in a uniform stress distribution. At this point, the
wire mesh reinforcement is assumed to have reached its ulti-

134 ACI Structural Journal/January 2019


Table 3—External applied loads required to cause
various failure modes in strut-and-tie model for
light-frame AAC floor diaphragms
Failure Required external
mode Description load, P, kip (kN)
Crushing of the AAC under load bearing
1 18.3 (81.3)
plates or in smallest section of strut.
Crushing of the AAC at perimeter cross
2 13.9 (62.0)
wire bearing interfaces.
3 Screw anchorage failure 16.9 (75.3)
4 Crushing of the AAC at reaction supports 12.8 (56.9)

required to cause Failure Mode 3 and a value of P3 = 16.9 kip


(75.3 kN) was determined.
Failure Mode 4—This failure is characterized by AAC
crushing at the diaphragm reaction support interfaces. By
statics, the forces resisted by the reaction supports is P/2.
Fig. 12—Proposed strut-and-tie model for light-frame AAC The bearing area of each reaction support is simply the
floor diaphragms. product of the width of the reaction support and the thick-
ness of the AAC panels. By multiplying the compressive
mate state, regardless of whether the longitudinal wires are strength of AAC, fAAC, by the sum of the two bearing areas,
at their full tensile capacity. a value of P4 = 12.8 kip (56.9 kN) was determined. As with
As previously stated, perimeter cross wires are responsible experimental floor diaphragm tests, Failure Mode 4 requires
for resisting the global x-component of inclined compres- the smallest external applied load, and thus controls over
sion strut forces. The capacity of which is governed by the all the other investigated failure modes. A summary of the
compressive strength of AAC bearing on the cross wires as external applied loads required to cause the investigated
shown in Fig. 14. The magnitude of force that can be resisted failure modes described in the previous sections is presented
by perimeter cross wires is determined by multiplying fAAC in Table 3.
by the cross-wire length and diameter. The value obtained The magnitudes of loads in Table 3 indicate that crushing
will be equal to the x-component of the inclined compres- at the support will govern the behavior. In most cases,
sion strut force, which can easily be correlated back to an ultimate capacity was reached at the same time AAC was
externally applied diaphragm force. Through this method, crushed at the supports. Thus, the observed damage was
the externally applied load required to cause Failure Mode 2 consistent with the predicted failure modes. Steel-framed
was determined to be P2 = 13.9 kip (61.8 kN). diaphragms sometimes withstood higher loads accompanied
Failure Mode 3—Screws anchor panels to substructure by increased deflections after the initiation of AAC support
framing elements are responsible for distributing panel crushing; this is attributed to the ductile nature of the steel
forces to the joists and tension chord of the substructure frames in comparison to the wood. If larger supports had
frame (Fig. 13(c)). Although the floor diaphragms discussed been used, then crushing of AAC under the cross wires
herein featured thin-bed mortar at panel joints, for analysis (Failure Mode 2) would occur with a capacity increase
purposes, the tensile strength contributions of those joints to of roughly 10%. Although higher load capacities were
resist the x-component of inclined compression strut forces observed with steel substructure diaphragms as compared to
have been conservatively ignored. The x-component of strut wood substructure diaphragms, the capacity of the proposed
forces is assumed to be resisted by engaging the screws that strut-and-tie model discussed herein is independent of spec-
anchor perimeter panels. imen framing material. Finally, Table 4 provides a compar-
Because of the running bond panel orientation, each spec- ison between strut-and-tie predicted load capacity and actual
imen has one small panel measuring 2 ft (0.61 m) square that experimental capacities for floor diaphragms loaded parallel
is located along the path of the compression strut. This small to joists.
panel is anchored to the underlying frame by the fewest As expected, the strut-and-tie model provides a conser-
number of screws and therefore represents the weakest link. vative prediction (ratios of observed to predicted capacity
Theoretically, the six screws anchoring the panel in place consistently greater than unity), even in cases where no
are responsible for transferring half of the x-component mortar was used at panel joints. Therefore, the proposed
force imposed by the inclined compression strut. Using the strut-and-tie model provides an appropriate lower-bound
assumption that forces are evenly distributed to the screws, solution for the capacity of floor diaphragms.
the shear force resisted by each screw is established by For design purposes, the value predicted using the strut-
dividing the number of screws by half of the x-component and-tie model should be reduced by the appropriate strength
force. By statics, the total shear resistance capacity of the reduction and beta-factors prescribed in ACI 318 (ACI
engaged screws was related to the externally applied load Committee 318 2011). This document requires that the
effective compressive strength of concrete, 0.85fc′, within

ACI Structural Journal/January 2019 135


struts or nodal zones be adjusted by a factor of βs or βn, be applied to the nominal strengths of struts, ties, and nodal
respectively. In addition, if the tested material strength fAAC zones. Application of ACI provisions to the strength of struts
was replaced with the design strength fAAC′ and multiplied and nodal zones gives values of factored allowable external
by the beta factors, the predicted capacities would be further applied loads, P, listed in Table 5.
reduced based on Eq. (1) or Eq. (2), respectively Once again, Failure Mode 4 controls. As with the exper-
imental testing, if the bearing area would have been suffi-
f ce = 0.85β s f AAC
′ (1) ciently large, the next failure would be in the compression
strut. This value of P is a very safe capacity for floor
f ce = 0.85β n f AAC
′ (2) diaphragms of the configuration discussed herein. Applica-
tion of ACI strut-and-tie factors results in an average factor
where 0.85fAAC′ represents the effective AAC strength under of safety of 3.5. The data suggest that it is not necessary to
sustained compression. In addition to the βs and βn values, use the lightweight factor, λ, for AAC.
ACI 318 prescribes a strength reduction factor  = 0.75 to
Table 4—Comparison of strut-and-tie model Evaluation of specimens loaded perpendicular
predicted capacity to experimental test results to joists
The strut-and-tie model does not apply to floor diaphragms
Observed diaphragm Ratio of observed/ loaded parallel to joists because the longitudinal cracks that
Specimen
capacity, kip (kN) predicted* capacity
formed in the AAC panel joints could interfere with forma-
W-II-8-12-1 17.0 (75.6) 1.33 tion of a compression strut. As a result, the unmortared shear
W-II-8-12-2 15.7 (69.8) 1.23 wall equation of ACI 523.4R was applied to the AAC portion
S-II-8-12-1 21.2 (94.3) 1.66
of the diaphragm between the load and support (Eq. (3)). The
wall height was the projected distance between the load and
S-II-8-12-2 17.5 (77.8) 1.37
support 4 ft (1.2 m) and the length was 12 ft (3.7 m). Based
S-II-8-12-3 17.2 (76.5) 1.34 on these values, the predicted diaphragm capacity would be
S-II-8-12-4 †
13.9 (61.8) 1.09 P = 2VnAAC = 12.8 kip (56.9 kN). The average experimental
*
Predicted diaphragm capacity is 12.8 kip (56.9 kN). load of the four diaphragms is 18.2 kip (81 kN), resulting in

Non-mortared specimen. a ratio of observed (tested) to predicted capacity of 1.4

Table 5—ACI factored capacities as determined by VnAAC = 0.6lw t f AAC (3)


strut-and-tie model analysis
ACI factors
Failure ACI factored design
mode Component type βs βn  capacity, P*, kip (kN) CONCLUSIONS
A total of six individual panels of the type used in floor
1 Nodal zone NA 1.0 0.75 11.6 (51.8)
diaphragm constructions were tested under transverse, mono-
1
Compression
0.45 NA 0.75 5.24 (23.3) tonically increasing loads. Panels typically failed in flexure
struts and were capable of withstanding large span deflections with
2 Nodal zone NA 0.6 0.75 5.33 (23.7) negligible loss in load-carrying capacity. The ultimate load
3 Tension tie NA based on 3 in. (75 mm) thick panels is 425 lb/ft2 (20.3 kPa).
Specimens supported by steel joists carried roughly 10%
4 Nodal zone NA 0.6 0.75 4.89 (21.8)
more load than specimens supported by wood joists.
*
Values of P are reduced by relevant ACI factors and coefficient 0.85, which
represents effective concrete strength under sustained compression.

Fig. 13—Free body diagrams describing: (a) cross wire and support reactions at compression strut interface; (b) reaction
forces on wire mesh reinforcement; and (c) external forces exerted on panels.

136 ACI Structural Journal/January 2019


L = length between center lines of supports in individual gravity
panel tests
Px = load in floor diaphragm based on failure mechanism x
scross = spacing of cross wires (transverse steel) in reinforced AAC
panel
sh = spacing of longitudinal wires in reinforced AAC panel
t = specified thickness of AAC panel, 3 in. (76 mm)
VnAAC = nominal shear strength of AAC shear wall
βn = factor used to account for effect of anchorage of ties on effective
compressive strength of nodal zone
βs = factor used to account for effect of cracking and confining rein-
forcement on effective compressive strength of concrete in strut
 = strength reduction factor

Fig. 14—Free-body diagram of wire mesh reinforcement at REFERENCES


ultimate state. ACI Committee 318, 2011, “Building Code Requirements for Structural
Concrete and Commentary (ACI 318-11),” American Concrete Institute,
A total of 10 AAC-sheathed floor diaphragm specimens Farmington Hills, MI, 503 pp.
ACI Committee 523, 2009, “Guide for Design and Construction with
were built and tested under in-plane, lateral, monotonic Autoclaved Aerated Concrete Panels (ACI 523.4R-09),” American
loading. Based on five of the specimens with loads oriented Concrete Institute, Farmington Hills, MI, 81 pp.
parallel to the joists, an applied strut-and-tie model was Argudo, J. F., 2003, “Evaluation and Synthesis of Experimental Data for
Autoclaved Aerated Concrete,” master’s thesis, The University of Texas at
developed as an analysis tool for AAC floor diaphragms of Austin, Austin, TX, 297 pp.
similar constructions. Predicted strength values obtained ASTM C1692, 2011, “Standard Specification for Autoclaved Aerated
from the model agreed closely with experimental results, Concrete,” ASTM International, West Conshohocken, PA, 4 pp.
ASTM C1693, 2011, “Standard Practice for Construction and Testing of
indicating that this model can be used to predict reason- Autoclaved Aerated Concrete (AAC) Masonry,” ASTM International, West
able lower-bound strength values for AAC-sheathed floor Conshohocken, PA, 7 pp.
diaphragms. Furthermore, when ACI factors are added, ASTM D1761, 2012, “Standard Test Methods for Mechanical Fasteners
in Wood,” ASTM International, West Conshohocken, PA, 10 pp.
the factor of safety grows to 3.5. Floor diaphragms tested Chen, Y.; Peng, M.; Zhang, Y.; and Liu, Y., 2013, “Mechanical Proper-
with loads perpendicular to the joists experienced longitu- ties of Autoclaved Aerated Concrete with Different Densities,” Advances
dinal cracking that could limit the strut-and-tie model. As in Civil Engineering Materials, V. 2, No. 1, pp. 441-456. doi: 10.1520/
ACEM20130063
a result, equations from ACI 523.4R were applied and a Coduto, A., and McDonough, M., 2012, “Autoclaved Aerated Concrete
reasonable value of experimental to predicted strength of 1.4 as a Holistic Building System,” Structure, May, pp. 34-35.
was achieved. Dembowski, J., 2001, “A Study of the Material Properties and Struc-
tural Behavior of Plain and Reinforced AAC components,” master’s thesis,
A robust finite element solution could be used to model University of Alabama at Birmingham, Birmingham, AL, 489 pp.
diaphragm behavior in both directions as part of future work. Easterling, W. S., and Porter, M. L., 1994, “Steel-Deck Reinforced
Interaction between panels and fasteners would be helpful Concrete Diaphragms I,” Journal of Structural Engineering, ASCE, V. 120,
No. 2, pp. 560-576.
to further explain the progressive crushing observed in the Fouad, F. H.; Vichare, A. C.; and Barnett, R. E., 2005, “Development of
testing. In addition, experimental work should include modi- ASTM Standards for AAC,” Autoclaved Aerated Concrete Innovation and
fying the aspect ratio of floor diaphragms. Other variables Development – Proceedings of the 4th International Conference on Auto-
claved Aerated Concrete, Kingston University, London, UK, pp. 281-286.
that could be adjusted include AAC strength, thickness, or Griebenow, G.; Koch, R.; Sitka, R.; and Vokel, G., 1989, “Loadbearing
fastener type or layout. Finally, an assemblage specimen Floor Plates of Autoclaved Aerated Concrete Components,” Concrete
that combines the diaphragm with AAC-clad walls could Precasting Plant and Technology, V. 55, No. 6, pp. 56-66.
Jennings, C. J., 2016, “Behavior of Autoclaved Aerated Concrete Rein-
be tested. forced Thin Panel Sheathing Elements for Walls and Floor Diaphragms,”
master’s thesis, University of Wyoming, Laramie, WY, 208 pp.
AUTHOR BIOS Judd, J., and Fonseca, F., 2002, “Strength and Behavior of Hybrid
Caleb Jennings is a Structural Engineer at Martin-Martin in Cheyenne, Diaphragms,” Journal of Composites for Construction, ASCE, V. 6, No. 6,
WY. He received his BS in architectural engineering in 2012 and his MS pp. 215-233. doi: 10.1061/(ASCE)1090-0268(2002)6:4(215)
in civil engineering in 2017 from the University of Wyoming, Laramie, WY. Matthys, J. H., and Nelson, R. L., 1999, “Structural Properties of Auto-
claved Aerated Concrete Masonry,” ASTM Special Technical Publication,
Nicholas A. Owen is a Structural Engineer at Martin-Martin. He received V. 1356, pp. 13-22.
his BS in architectural engineering in 2012 and his MS in civil engineering Parker, C. K., 2005, “Evaluation of a Strut-and-Tie Model to Design
in 2016 from the University of Wyoming. Autoclaved Aerated Concrete Floor Diaphragms to Resist Lateral Loads,”
master’s thesis, University of Wyoming, Laramie, WY, 200 pp.
Jennifer Eisenhauer Tanner, FACI, is an Associate Professor in the Short, A., and Kinniburgh, W., 1961, “The Structural Use of Aerated
Department of Civil and Architectural Engineering at the University of Concrete,” The Structural Engineer, V. 39, No. 1, pp. 1-16.
Wyoming. She received her PhD in 2003 from the University of Texas at Storlie, V. L., 2009, “Behavior of Autoclaved Aerated Concrete (AAC)
Austin, Austin, TX. She is Chair of ACI Committee 526, Autoclaved Aerated Floor Diaphragms Subject to In-Plane Reverse Cyclic Loading,” master’s
Concrete, and a past Co-Chair of ACI Subcommittee 440-L, FRP-Dura- thesis, University of Wyoming, Laramie, WY, 223 pp.
bility. From 2005 to 2015, she was a member of the ACI 530/TMS 402/ UES 350, 2017, “Hebel Power Floor Panels and Hebel Thin-Bed
ASCE 6, Masonry Building Code. Mortar,” Uniform Engineering Services, Evaluation Report 350, Ontario,
CA, 5 pp.
Valore, R. C., 1954, “Cellular Concretes Part 2: Physical Properties,”
NOTATION ACI Journal Proceedings, V. 50, No. 6, June, pp. 817-836.
dcross = diameter of transverse (cross) wire in reinforced AAC panel Yeoh, D.; Fragiacomo, M.; Buchanan, A.; and Gerber, C., 2009,
dlong = diameter of longitudinal wire in reinforced AAC panel “Preliminary Research Towards a Semi-Prefabricated LVL– Concrete
fAAC = tested compressive strength of AAC Composite Floor System for the Australasian Market,” Australian
fAAC′ = specified compressive strength of AAC Journal of Structural Engineering, V. 9, No. 3, pp. 225-240. doi:
fc′ = specified compressive strength of concrete 10.1080/13287982.2009.11465025
frAAC = tested modulus of rupture of AAC

ACI Structural Journal/January 2019 137


ACI Honors and Awards
Program Nominations

The American Concrete Institute is seeking ACI selects the winners of its
annual awards through an
your nominations for our Honors and
open nomination process.
Awards Program. Nominations are now
You can participate in the
being accepted for ACI’s Honorary Member Honors and Awards Program
as well as for ACI’s personal awards for by nominating worthy
candidates for award
individual or organizational merit, the ACI
consideration.
Young Member Award for Professional
Nomination forms can be
Achievement, the ACI Education Award, the found on the ACI website,
ACI Certification Award, the ACI Concrete www.concrete.org.
Sustainability Award, and more.

  
+1.248.848.3700 • www.concrete.org

View publication stats

You might also like