A Review On Flat-Jack Testing: January 2000
A Review On Flat-Jack Testing: January 2000
A Review On Flat-Jack Testing: January 2000
net/publication/237334295
CITATIONS READS
66 6,433
2 authors, including:
Paulo B. Lourenco
University of Minho
1,046 PUBLICATIONS 17,794 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Seismic Retrofitting Project (SRP), Getty Conservation Institute (GCI) View project
All content following this page was uploaded by Paulo B. Lourenco on 24 July 2019.
ABSTRACT
1. INTRODUCTION
1
Associate Professor ([email protected])
2
Socrates exchange student
2.1. Background
Flat-jack testing originates from the field of rock mechanics. Italian researcher Paolo
Rossi adapted the method for use with masonry in the early 1980s and, since then, different
researchers worldwide focus on this technique, e.g. Abdunur [1] carried out tests with very
small semi-circular flat-jacks, and conducted idealized photoelastic stress analyses on plastic
models. Atkinson-Noland & Associates [5] has been engaged in the evaluation of flat-jack
testing for use in the evaluation of existing old brick masonry buildings in the United States.
Qinglin and Xiuyi [10] developed a thick flat-jack with large displacement capabilities for use
on very soft masonry materials typically found in China.
Two separate standards for masonry evaluation with flat-jacks were developed in the
United States by the ASTM and approved in 1991. ASTM Standard Test Methods C 1196-91,
In-Situ Compressive Stress Within Solid Unit Masonry Estimated Using Flat-jack
Measurements [2], and C 1197-91, In Situ Measurement of Masonry Deformability Properties
Using the F1atjack Method [3]. European practice follows RILEM standards LUM.D.2 [6]
and LUM.D.3 [7] , which were first introduced in 1990.
Flat-jack is a “thin envelope-like bladder with inlet and outlet ports which may be
pressurized with hydraulic oil” [2] (some typical configurations are shown in Figure 1). A
flat-jack may be manufactured in many shapes and sizes - the actual dimensions are
determined by its function, slot preparation technique and the properties of the masonry being
tested. Flat-jacks with curved edges (types c, d in Figure 1) are designed to fit in a slot cut by
a circular saw. Rectangular jacks (types a, b) are used where mortar must be removed by hand
or with stitch drilling. Regardless of the shape, a flat-jack must fit the slot well. The thickness
of the flat-jack is determined by its specific function: An ideal flat-jack will completely fill
the slot in the mortar joint. However, if such flat-jack is not available, then shims are used
together with the flat-jack to completely fill the slot thickness.
a) b)
10.2
10.2
20.3 40.6
0.38 0.38
c) d)
8.9
8.9
r=
r=
15
15
.2
.2
27.7 48.0
This test is based on the principle of partial stress release and involves the local
elimination of stresses, followed by controlled stress compensation (see Figure 2).
p p p
di d<di pf d=di
a) b) c) d) pressure pf
before the cut after the cut
The principle of the test is similar to a standard compressive test. The difference is that
it is performed in-situ and two flat-jacks are used to apply the load. A typical setup of the in-
situ deformability test is shown in Figure 3.
By cutting two parallel slots, part of the wall is isolated from the surrounding masonry
forming a “specimen”. Masonry between the flat-jacks is assumed to be unstressed. Flat-jacks
are then introduced into both slots, and the initial distances between gauge points are
measured. By pressurizing flat-jacks, the load is applied to the “specimen” creating an
approximately uniaxial state of compressive stress. With a pressure increase in the flat-jacks,
the distances between gauge point pairs decrease. By gradually increasing the pressure, the
stress-strain relationship can be determined. Loading-unloading cycles can also be performed.
Based on an experimental stress-strain curve, the value of Young’s compressive
modulus can be calculated. If extended damage in the specimen is acceptable, the
compressive strength of masonry can be obtained. Obviously, this can only be done if the
2.5. Equipment
The following equipment is required: one or more flat-jacks; a hydraulic system with a
pump, a gauge and hoses; a displacement measurement equipment with an appropriate
number of gauge points, brackets and plugs; tools for mortar removal such as a masonry saw
or a drilling machine, a hammer and a chisel; safety equipment. Additional optional
equipment includes shims (single or multi-piece), a data acquisition unit and a power source.
2.6. Flat-jacks
B
CIRCULAR
some flat-jacks are fitted only with one port. The
shape of the flat-jack depends of the equipment used
B
R
to create the slot, see Figure 4. The size of the flat-
jack depends on the application, ranging from a few A
mechanics).
Figure 4 - Flat-jack dimensions.
Flat-jacks are designed to have an output pressure (one that is applied to masonry) that
is linearly dependent on the internal hydraulic pressure. The coefficient that provides
conversion (Km) is determined during the calibration process. An example of a curve obtained
during calibration, illustrated in Figure 5, shows the relationship between the internal
hydraulic pressure in the flat-jack and external pressure.
1.0
0.9
ideal flatjack
Ex 0.8
ter
10 mm steel flatjack
nal 0.7
Pr
ess 0.6
ure 0.5
M
Pa 0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Internal Pressure MPa
Usually, for new flat-jacks, the calibration factor Km is supplied by the flat-jack
manufacturer. However, since flat-jacks softens with repeated use, they must be re-calibrated
after 4 or 5 tests, or sooner, if during test, flat-jacks develop excessive deformations. Flat-
jacks should be restored to the original thickness following each test to ensure that the
calibration factor remains unchanged. The flat-jack calibration procedure is described in [2].
Stitch drilling is only appropriate for weak mortars (as usually found in old masonry
structures) and not for modern, strong, cement-based mortars. The use of high-power hammer
drills is not recommended due to the possible disturbance caused to surrounding masonry. In
the case of strong mortars, irregular stone masonry, thick joints, or need of cutting through
units, a masonry saw should be used. A water-cooled saw with a carbide or diamond tipped
disk is suitable.
It is not allowed to grout the flat-jack in the slot because grout would flow into void
and cracks resulting in local change of masonry behavior. To ensure a uniform transfer of
pressure over the complete area, the flat-jack must fit tightly into the slot. ASTM standards
allow a difference in plane dimension up to ½ in. (1.25 cm). Shims are used to fill completely
the thickness of the slot. Shims should have the same size and shape as the flat-jack being
used and can be of three types: single piece, multiple pieces and fluid cushion, see Figure 7.
For the stress test, both the RILEM and ASTM standards recommend placing
reference points symmetrically on second courses (counting from the slot) above and below
the slot. The ASTM recommends placement of at least four pair of equally spaced points and
RILEM recommends that at least three pairs, placed in the middle part of flat-jack length. For
the deformability test, both standards require that reference points are placed symmetrically in
the masonry courses immediately above and below slots.
Figure 8 - Position of strain reference points for stress test ASTM and RILEM proposals.
The ASTM standard recommends the value of pressure increment to be equal to 25%
of estimated maximum flat-jack pressure, while the RILEM standard recommends small
increment without specifying its value. Authors of reference [8] recommend increments of 70
to 140 kPa. The pressure at which original distances are restored (the canceling pressure) is
the base for compressive stress calculation An acceptable difference (deviation) between
original and restored distances is [2]: – for each deviation less than ±0.0025 cm or 0.1 of
maximum initial deviation – for average deviation less than ±0.0013 cm or 0.05 of maximum
initial deviation. In order to reduce the creep effect, the time taken for the load application
should be approximately the same, as the time required for making the cut and preparing the
test (after strain measurements are stable). In this case creep deformation will be symmetrical
and balance itself out, as stated by ASTM and RILEM standards. Figure 9 presents typical
examples results obtained in a stress and a deformability test.
180
600
160
120
Deformation
Stress [psi]
[in. x 10 -4 ]
400
100
40 200
20
280 psi 100
0
-20 0
0 4 8 12 16 20 -3 -3 -3 -3
0.0 0.5 · 10 1.0· 10 1.5 · 10 2.0· 10
location [in.] strain [in./in.]
(a) (b)
Figure 9 - Example test results from: (a) in-situ stress test (deformations at different stress
levels) and (b) cyclic in-situ deformability test.
For both the stress and deformability tests it is necessary to convert the flat-jack
pressure to the actual compressive stress. The stress can be calculated by
σm = Km Ka p (1)
where Km is the calibration factor (<1), Ka is the ratio of measured area of the flat-jack to the
average measured area of the slot (<1) and p is the flat-jack pressure.
For the stress test [2], states that this test shows a 20% coefficient of variation and no
inherent bias in predicting the state of compressive stress present in the masonry. That is, this
method can as well over and underestimate actual stress. Authors of reference [8] state that:
“laboratory testing has shown that the in-place stress test has a margin of error of up to 20%”.
For the deformability test, [3] gives a coefficient of variation of 24% but the test typically
overestimates the average Young’s modulus of masonry up to 15%. This is due to the actual
boundary conditions. Influence of surrounding masonry (including effect of collar joints when
only part of wall thickness is tested) decreases toward the middle of the loaded area. For this
reason, location of strain reference points only in middle part of loaded area (as found in the
RILEM standard) would lead to more accurate results than equal distribution (recommended
in the ASTM standard).
Although flat-jack stress test was originally proposed to measure only compressive
stresses, it can be also used to estimate tensile stresses [1].
Flat-jacks can also be used to perform the in-situ shear test (“shove test” or “push
test”). This is achieved by horizontally displacing a single masonry unit with a hydraulic jack.
Load applied to masonry by a flat-jack is assumed to be constant over its area. Actual
stress distribution for typical stainless steel flat-jack is however different. Example of pressure
contours on surface of the flat-jack, obtained experimentally, is shown in Figure 10. In this
test, internal flat-jack pressure was 0.396 MPa. A more uniform stress distribution can be
achieved, with the use of a rubber jack. The relationship between internal and external
pressures of the rubber flat-jack provides a calibration factor Km close to 1.
(a) (b)
Figure 10 - Pressure Contours for: (a) Steel Flat-jack and (b) for Rubber Flat-jack2 [4].
It must be noted that the objective of the tests carried out at University of Minho was
not to confirm the suitability or accuracy of the flat-jack method as this has already been done
in many laboratory experiments worldwide. Since there is some disagreement in
recommendations for flat-jack tests between standards and authors, these different
recommendations were followed in different experiments. Additionally, it is known that good
slot quality is essential for obtaining correct test results. Since knowledge of the true contact
area during the flat-jack testing is valuable for decisions about test validity, a very simple and
inexpensive method was introduced which allowed that information to be obtained. To
investigate the existing possibilities, also different techniques and tools for mortar removal
were used during experiments.
Rectangular flat-jacks, with dimensions 40.6 × 10.2 × 0.42 cm3, supplied by Atkinson-
Noland & Associates were adopted in the tests. The flat-jack was pressurized using a
manually operated hydraulic jack. Displacements were measured using a removable
Whittemore Gauge with a resolution of 0.0001 cm. Three different gauge lengths were used
during experiments: 11, 25, 31 cm. Square metal plates with dimensions of 2 × 2 × 0.2 cm3
served as gauge points. Each plate had a conical depression compatible with pointing
elements of the removable extensometer. Before testing, all reference points were attached to
surface of masonry with an epoxy glue. Both, multiple and single piece shims were used.
Two test walls (further referenced as TW1 and TW2) were constructed with the same
type of masonry units (YTONG blocks) and two different mortars: a cement-lime mortar for
the specimen TW1 and lime-mortar provided by Weber & Broutin for the specimen TW2.
YTONG blocks were chosen due to a small weight of this material as well as because this
material was readily available, see Figure 11.
2
The significant area of high pressure, shown as 0.4-0.5 MPa, is somewhat misleading as the pressure only just
creeps into that range.
The load was applied on top of the wall by an external hydraulic jack. The load was
distributed along the length of the wall with a rigid steel element. The load cell was placed
between the jack and the steel element. Three different stress tests have been carried out in the
two walls. The slot for the first test was made partly by stitch drilling and partly by saw
cutting. Stitch drilling proved not adequate at all for the combination of masonry units (soft
YTONG blocks) and modern hard mortars. The drilling was quite irregular and difficult. For
this reason, the slot was further opened by saw cutting. The second and third tests were fully
carried out in a saw cut slot, see Figure 12. Four pairs of reference were adopted, following
ASTM (test no. 2) and RILEM (test no. 3) proposals.
(a) (b)
Figure 12 - Mortar removal: a) by stitch drilling b) saw cutting.
Since knowledge of the contact area during flat-jack testing is important when
deciding about test validity, to gain better understanding of this area, a very simple and
inexpensive method was introduced in experiments following experiment no. 1: a sheet of
carbon paper, sandwiched between two sheets of ordinary paper, was placed between the flat-
jack and a surface of the shim. The paper was marked in places of contact, conversely to
places without contact, where the paper remained white. Examples of the contact area
contours obtained with this method are shown in Figure 13.
8.6
10.2
39.2
(a)
40.6
10.2
7.0
38.0
(b)
Figure 13 - Contact contours (max. pressure 0.5 MPa) for two different tests.
As reported before, the first test must be ignored due to the inadequacy of the slot.
Only the results of the second and third tests will be given here.
Figure 14a shows that the distances between reference points at 0 and 100 kPa were
practically the same for reference point pairs 1, 2, 4. Probably, there was no contact between
the flat-jack and the masonry at this pressure level. Please note that an error in reading at
pressures below the canceling pressure (in this case at 0, 100, 200, 300 kPa) does not have
any influence on the calculation of test results (i.e. estimated value of compressive stress).
The result of the flat-jack test is calculated based on the value of the canceling
pressure. Note that, because during the stress test, displacement measurements are not taken
continuously, but at pressure intervals that there are two possible ways of determining the
canceling pressure. One possible approach is to recognize certain pressure level as the
canceling pressure when displacements at this level and initial displacements (measured
before slot creation) does not differ more than certain tolerance value. Such an approach is
recommended by ASTM standards, which contain requirements for allowable tolerance.
Another possibility is to determine pressure corresponding to the zero displacement by
interpolation between two pressure levels (corresponding to displacements greater and lower
than initial displacements. In other words to find point of intersection between the
pressure/displacement graph (as one shown in Figure 14b) and the pressure axis.
Here, the latter approach was used and the canceling pressure was determined as
pf = 420 kPa. Note that in the case when inelastic deformations are present, the “zero
displacement” for each pair of reference points will be reached at different pressure level. In
this case, a criterion of "equal residual displacements" of points affected by similar load-
displacements history should be used [8].
The results for test no. 3 are similar to the results above and will not be shown here,
see [9] for a complete report. The obtained canceling pressure was 400 kPa.
Displacement [mm]
0.070
100 kPa
0.050 0.000
200 kPa
0.030 0 100 200 300 400 500 600 700 800
300 kPa
0.010 -0.050
400 kPa
-0.010
1 2 3 4 500 kPa -0.100
-0.030
-0.050 -0.150
-0.070
Reference point pair -0.200
Pressure [kPa]
(a) (b)
Figure 14 – Results for test no. 2: (a) deformations at different pressure levels;
(b) Pressure/displacement diagrams for each pair of reference points
The obtained stress values are close and represent an error of +22% (test no. 2) and
+18% (test no. 3). These values are expected according to the standards, even if it is peculiar
that the results of both tests are similar and biased to the overestimation side. For practical
applications in historical constructions such error in the stress estimation is reasonable and
represents valuable information for the analyst / designer.
4. CONCLUSIONS
5. REFERENCES
[1] C. Abdunur, Stress and deformability in concrete and masonry, IABSE Symposium on
Strengthening of Building Structures- Diagnostic and Therapy, Venice, Italy, 1983
[2] ASTM, In-situ compressive stress within solid unit masonry estimated using flat-jack
measurements, ASTM Standard C 1196-91, 1991
[3] ASTM, In-situ measurement of masonry deformability properties using flat-jack
method, ASTM Standard C 1197 -91, 1991
[4] T. G. Hughes, R. Pritchard, In-situ flat-jack tests matching new mechanical
interpretations, 10th Int. Brick/Block Masonry Conf., Calgary, Canada, 1994
[5] J.L. Noland, R.H. Atkinson, M.P. Schuller, A review of the flat-jack method for
nondestructive evaluation, Proc. Nondestructive evaluation of civil structures and
materials, Boulder, USA, 1990
[6] RILEM, LUM.D.2, In-situ stress tests on masonry based on the flat jack, 1990
[7] RILEM, LUM.D.3, In-situ strength/elasticity tests on masonry based on the flat-jack,
1990
[8] P. Ronca, C. Tiraboschi, L. Binda, In-situ flat-jack tests matching new mechanical
interpretations, 11th Int. Brick/Block Mas. Conf., Shanghai, China, 1997
[9] P. Gregorczyk, Analysis of historical structures: Two aspects of advanced experimental
and numerical possibilities, University of Minho, Guimarães, Portugal, 1999
[10] W. Qinglin, W. Xiuyi, The evaluation of compressive strength of brick masonry in-situ.
8th Int. Brick/Block Mas. Conf., Dublin, Ireland, 1988