13th International Brick and

Block Masonry Conference

Amsterdam, July 4-7, 2004

PREDICTING THE BEHAVIOUR OF LATERALLY LOADED

MASONRY PANELS
AN Fried1, JJ Roberts2, MC Limbachiya3 and O Kanyeto4

Abstract

The paper aims to compare predictions by different methods of analysis of the lateral load
capacity of unreinforced masonry walls, and to determine whether a relationship can be
drawn between theory and experiments. Several analysis methods were employed to predict
failure loads of a number of wall panels, and the results were compared with test values. It is
concluded that boundary conditions play a major role in the accuracy of analytical methods.
When the boundary conditions assumed in the analysis most nearly match the real
conditions of the test panels, predictions are most accurate. The authors conclude that
finding a rational analysis and design procedure for laterally loaded masonry walls is
possible.

Key Words

analytical expressions, flexural strength, lateral loads, testing.

1 Introduction
Intensive research has been directed towards lateral strength of masonry wall panels in
recent decades, with the main objective being the derivation of analytical equations for the
prediction of strength properties. The results of these researches widely vary, making it
difficult to draw any reliable conclusions about the behaviour of the material in flexure. The
various design approaches adopted by different codes of practice is testimony to the
designers’ limited understanding of the behaviour of the material. The UK adopted the yield
line theory as the basis for design of masonry panels subjected to lateral loading, while
some European countries employ equations that are based on elastic plate theory. Because
predictions using these theories have not produced consistent results, researchers have
also looked at the development of empirical relationships.

This paper reviews and compares the theories on which some of the codified methods of
design are based. Results of a series of tests performed by Lawrence (1983) are used to
compare the ability of these theories to match experimental data. It has been assumed in

1
Reader, School of Engineering, Kingston University, London, UK
2
Dean, Faculty of Technology, Kingston University, London, UK
3
Reader, School of Engineering, Kingston University, London, UK
4
Researcher, School of Engineering, Kingston University, London, UK
page
1 1
this review that material properties, such as mortar types, water absorption rate of units, etc.,
were kept constant during the tests, and do not influence the comparisons.

2 The need for design formulae

Traditionally, structural masonry was not treated as engineered material in the same sense
as steel, concrete and timber. This resulted in rule of thumb procedures being applied for
masonry construction, which in turn produced excessively thick walls with consequent cost
penalties and wastage of space within a building. Haseltine et al (1977) note that, with the
massive walls, the ability to resist such small loads as wind was never in question. As walls
became thinner, and lower strength materials such as aerated concrete blocks were
introduced, the need to carry out structural calculations could no-longer be avoided.
Moreover, Haseltine et al (1977) observe that the narrowing of walls and introduction of
lower strength materials were accompanied by an increase in the pressure used to
represent the effect of wind in design.

The need to find a simple strength-to-size relationship of panels subjected to lateral loads
has been a topic of discussion since the inception of masonry-design codes. In December
1978, a discussion group of The Institution of Structural Engineers, considered a paper by
Haseltine at el (1978), and declared that design data on this subject was sorely lacking. A
year earlier, Haseltine at el (1977) had reported that very little information was available to
enable engineers and checking authorities to design walls for lateral loads, and that, what
there was had to allow for a wide range of dissimilar materials and take into account
variations in workmanship from well controlled to totally neglected. The code of practice of
the time, CP121 (Haseltine et al, 1977), had given rule of thumb methods for the sizing of
panel walls.

According to Lovegrove (1988), the lack of theoretical basis for the design of masonry walls
subjected to lateral loads made it impossible to extend the design rules beyond what was
currently contained in the design codes. He cites, as an example, the design of walls
containing openings.

3 Existing theories
There are a number of theories on which prediction of lateral loading capacities of masonry
walls can be based, which include: (a) elastic plate methods, (b) yield line theory, (c) finite
elements, (d) strip method, and (e) fracture line theory. Analyses using these methods give
quite different results when compared with experimental data, thus prompting researchers to
focus on finding new techniques, or a rationale for existing methods. While there is some
indication that each of these theories does give reasonable results at certain times, the
results are random and dependant on the specific testing programme undertaken. That is to
say, there is no consistent evidence, so far, linking theory with all types of masonry.
Consistent results only occur over specific sectors of the subject.

3.1 Elastic Plate Theory

Elastic plate theory would appear to be the most promising analytical technique since, in
most of the tests recorded in literature, the load-deflection relationship for laterally loaded
panels is nearly linear. Allowance for the orthotropic properties of brickwork is also made
without any difficulty.

The European code (Eurocode 6, 1996) allows designers to choose between using the
moment coefficients derived by the yield-line method and those derived by the elastic plate
theory. The moment coefficients in Table 1 of this review were calculated using plate-
bending equations derived by Timoshenko (1959).

page
22
3.2 Yield Line Theory
Yield line theory was developed for use with reinforced concrete, and assumes that the
bending moment along a line or lines reaches a yield value, and stays constant until other
parts of the line reach that yield value. Thus, a pattern of yield lines develops with constant
moment along each line, when failure occurs. It has been argued by several researchers
(Sinha 1978, Lovegrove 1988, and Lawrence 1995) that, with masonry, this is theoretically
unsound as it assumes the existence of plastic hinges which cannot exist in a brittle
material.

The method of design of unreinforced masonry panels given in the British Code of practice,
BS 5628: Part 1, is based on the yield line theory. To the best of the author’s knowledge, the
UK is the only country which uses this method. The yield line analysis used in this review
applies the moment coefficients taken from Table 9 of BS 5628: Part 1.

3.3 Finite Elements

Finite Elements have been applied by many researchers to simulate the behaviour of
masonry structures and has often produced very good results when compared with
experiments. The method is suitable for the prediction of failure loads, as well as stress
distributions in the working stress range. Page (1978) used the method to investigate stress
distributions and found that it was able to reproduce these with good accuracy. It has even
been shown, by Bouzeghoub and Riddington (1994), that there is no need to use 3-D finite
elements since simpler 2-D elements are adequate to simulate the behaviour of masonry
structures. The biggest drawback of the method is the effort and time it takes to idealise the
structure, input the data and, interpret the results.

Although the finite element method is a good analytical tool, it is not suitable for design
purposes. Designers need a simple analytical approach which enables accurate predictions
of wall capacity. For this reason, this method is not reviewed here.

3.4 Fracture Line Theory

The Fracture line method was proposed, and applied, by Sinha (1978), for the design of
masonry wall panels against lateral loads. In this method, Sinha proposes that the variation
of Young’s modulus with direction should be taken into account, and that the all
deformations take place only along the fracture lines. Besides the variation of E with
direction, the method does not differ from the yield line theory, since the resulting equilibrium
equations are based on the cracked pattern, with assumed ultimate (constant) moments
along the crack lines. As with the yield line method, the fracture line theory assumes that the
individual parts of the failed panel rotate as rigid bodies, and the equations of equilibrium are
derived from energy principles.

To the best of the author’s knowledge, the fracture line method is not incorporated in any
code of practice, but it is reviewed here because it has been available for some time and is
widely known in the field of masonry. This method can be found in a number of published
texts. The reader is referred to references 3 and 19 for the derivations and discussion of the
crack line formulae from which the coefficients in Table 3 were determined.

3.5 Strip Method

In this method, the load is split into parts which are carried by individual systems of strips,
designed as beams. Since the static equations are satisfied, a lower bound solution is
obtained. The strip method basically assumes that the load is carried by bending only, that
is, twist-free moment fields are considered. Since the existence of twisting moments along
both bed and perpend joints in masonry panels cannot be avoided, a rather conservative
solution can be expected. In reinforced concrete slab design, this is compensated for
through savings in reinforcement by providing only the amount necessary to resist the

page
33
bending moments. The strip method equations used to calculate the values in Table 3 were
derived from crack patterns of the panels tested by Lawrence (1983).

4 Comparison of the theories

The theories discussed above have been employed to reproduce the results of a series of
tests that were carried out by Lawrence (1983). The tests used clay bricks laid in a 1:1:6
mortar (cement: lime: sand by volume). Table 1 shows failure loads of the wall panels as
calculated by the different theories, as well as experimental results as recorded by
Lawrence. Lawrence recorded three different failure stages: the load at initial cracking; the
load at full-crack pattern; and the ultimate failure load. The results have been plotted on
Figures 1 to 4.

In Figure 1 the failure loads of the panels as calculated by the different theories are
compared to the full-crack-pattern load as obtained by Lawrence. The wall panels are
grouped according to the boundary conditions, as can be seen from the figure. It is apparent
from the figure that, with exception of very few panels, the yield line method over-estimates
the failure load of the panels. The strip method, on the contrary, under-estimates the failure
loads of most panels. Elastic plate theory also predicts larger failure loads than obtained by
tests, with the exception of simply supported panels. The fracture line method was only
applied to panels with simple supports and to those with fixed edges and simply supported
along the top and bottom edges (boundary condition 3). For other support conditions of the
tested walls, the fracture line equations were not readily available. While it could have been
possible to derive such equations from first principles, the available equations were deemed
adequate for the purposes of this review. In the simply supported panels, the fracture line
method was found to under-estimate Lawrence’s test results, while in the other panels, the
fracture line method invariably over-estimated the failure loads. Closer observation of Figure
1 also shows that the test values for a few panels (panels 12, 14, 15, 16, 37 and 38) are
much higher than the corresponding theoretical predictions. All of these are square panels,
i.e., their aspect ratio is 1. This gives some indication that regardless of boundary conditions,
theoretical methods used here under-estimate the strength of square panels with the largest
error margin.

Figure 2 compares the theoretical results to the failure load at initial crack formation. With
the exception of the strip method when applied to panels with boundary conditions 4 and 5,
the test results indicate that cracks start forming in the panels at loads lower than those
predicted by theory. In the case of panels with all edges fully fixed (boundary condition 2),
test results are much lower than the theoretically predicted values. This means that panels
will start developing cracks under loads that are much lower than those predicted by theory.

In Figure 3, the theoretical methods are compared with experimental ultimate load capacities
of the panels. It is noticed here that results from the yield line method almost coincide with
the test results for the majority of panels. Comparing this figure to Figures 1 and 2, it is
evident that theoretical predictions are closest to the ultimate failure loads of the wall panels,
but still there is considerable variability.

Figures 4 to 8 are extracts from Figure 1. They show the full crack failure loads of the panels
grouped according to their support conditions. Figure 5, with 8 wall results, contains the
largest number of panels tested. This is a very small number of panels from which
meaningful conclusions regarding the behaviour of walls can be made, and more test data
is, therefore, required. A lot of other test data does exist, boundary conditions are
questionable, and the detail with which Lawrence tested his walls is not always evident.

In general, all of the methods of analysis reviewed here reveal some level of inconsistency
and large variability. An important observation from the Figures is that the results for simply
supported panels are less scattered than those for panels with any other support conditions.

page
44
This indicates either that all theoretical methods reviewed here predict the failure loads for
simply supported panels with better accuracy than for panels with other support conditions
or that it is easier to accurately produce simple supports than other supports when testing
walls. The methods also give better approximations for panels with mixed boundary
conditions as compared to panels with all supports fixed. It is observed that as the number of
built-in edges increases, the theoretical predictions drift further from test results. This trend
can be associated with the values of the moment coefficients as they are assigned with the
assumption of full continuity at the built-in edges. When the assumed built-in edge is not
capable of full moment resistance, the assigned coefficient becomes faulty. With the simple
supports or free edges better approximations of the coefficients can be made. This confirms
the influence of boundary conditions on the strength of panels, with accurate predictions
occurring when the boundary conditions assumed in the analysis nearly match the real
conditions of the test panels.

It is very important to note here that the major task of this review was to compare the ability
of the different theories to match test results, as opposed to determining the load-carrying
capacity of panels with different boundary conditions. In Table 2 the ratios of predicted
values to the test results are displayed. These ratios are then plotted in Figure 9. It is clear
from the figure that, for most panels, the strip method gives the best results of the walls
tested as its values are closest to unity. It is also apparent from the graph that, in general,
the yield line method over-estimates the load carrying capacity of these walls more than the
other methods. However, the yield line method gives the best results for panels with free top
and simple supports on the other edges (far right of graph, panels 28, 26, 10, 36, 17 and
15). Incidentally, almost all of the walls tested by Haseltine et al (1977) for the validation of
the yield line method in masonry panels were supported on three edges (free along the top
edge).

The foregoing discussion clearly demonstrates the intricacy of any attempt to draw
conclusions from the predictions of the existing theories, and illustrates the need to generate
more test data. With today’s improved testing equipment and methods, we should strive for
enhanced simulation of the boundary conditions.

5 Conclusions and recommendations

The following are the conclusions drawn from this review:
1. The methods used to predict lateral strength of wall panels are not consistent.
2. Boundary conditions play a major role in the strength of the walls, and, results of
theoretical methods can be greatly improved if a good representation of the boundary
conditions can be made.
3. Different methods yield good approximations for different types of panels
4. All methods reviewed here give reasonable results for panels simply supported on all
edges.
5. All methods give poor results for panels with all edges built-in.
6. The strip method yields very conservative values of the lateral strength of panels, while
the yield line method almost invariably over-estimates the failure load.

It is recommended that:
1. More test data be generated, paying particular attention to boundary conditions.
2. The possibility of applying more than one of the above-reviewed methods to derive a
method of design be investigated.
The possibility of combining two or more of the above-reviewed methods with empirical
methods be investigated.

page
55
6 References
Abboud, B.E., Xin, L. and Becica, I.J., 1995, Deflection of Reinforced Masonry Walls Under
Out-of-Plane Loads, Proceedings of the Seventh Canadian Masonry Symposium,
Hamilton, Ontario, 895 – 910.
Bouzeghoub, M.C. and Riddington, J.R., 1994, 2-D and 3-D Finite Element Analysis of
Masonry Walls: A Comparative Study, Proceedings of the 10th International Brick and
Block Masonry Conference, Calgary, Canada, Volume1, 31-40.
BS 5628: Part 1: 1992, Structural Use of Unreinforced Masonry, British Standards Institution,
UK.
Christiansen P., 1995, Theoretical Determination of Flexural Strength of Unreinforced
Masonry, Masonry International, The British Masonry Society, Vol.9 No.3, 84-90.
Eurocode 6: Design of Masonry Structures – Part 1-1: General Rules for Buildings, 1996
Hagsten, L.G. and Nielson, M.P., 2000, Laterally Loaded Masonry, Proceedings of the 12th
International Brick/Block Masonry Conference, Madrid, Spain, 841 – 858.
Hansen, K.F., 2001, Uniaxial Bending Strength of Masonry Walls, Masonry International,
The British Masonry Society, Vol.14 No.3, 96-100
Haseltine, B.A., West, H.W.H., and Tutt, J.N., 1977, Design of Walls to Resist Lateral Loads,
The Structural Engineer, Vol.55, 422 - 430.
Haseltine, B.A., West, H.W.H., Hodgkinson, H.R., and Tutt, J.N., 1978, The Resistance of
Brickwork to Lateral Loading: Discussion, The Structural Engineer, Vol.56A, 363 -
374.
Hendry, A.W., Sinha, B.P. and Davies, S.R., 1981, An Introduction to Load Bearing
Brickwork Design, Ellis Horwood Limited, Chichester, England.
Lawrence, S.J., 1983, Behaviour of Brick Masonry Walls Under Lateral Loading, Ph.D.
Thesis: Volume 2, School of Civil Engineering, University of New South Wales,
Australia.
Lawrence, S.J., 1995, The Behaviour of Masonry in Horizontl Flexure, Proceedings of The
Seventh Canadian Masonry Symposium, Hamilton, Canada, Volume 1, 525-536.
Li, S., Fried, A.N., and Roberts, J.J., 1995, Deformational Analysis for Reinforced Concrete
Blockwork Beams, Proceedings of the Seventh Canadian Masonry Symposium,
Hamilton, Ontario, Canada, 501 – 512.
Lourenco, P.B., 2001, A Review of Out-of-Plane Behaviour, Masonry International, The
British Masonry Society, Vol.14 No.3, 67-73.
Lourenco, P.B., Rots, J.D., and van der Pluijm R., 1998, Understanding the Tensile
Behaviour of Masonry Parallel to the Bed Joints: A Numerical Approach, Masonry
International, The British Masonry Society, Vol.12 No.3, 96-103.
Lovegrove, R, 1988, A Discussion of Yieldlines in Unreinforced Masonry, The Structural
Engineer, Volume 66, No.22, 371-375.
Page, A.W., 1978, Finite Element Model for Masonry, Journal of Structural Division,
American Society of Civil Engineers, Vol.104, No.8, 1267-1285.
Sinha, B.P., 1978, A Simplified Ultimate Load Analysis of Laterally Loaded Model
Orthotropic Brickwork Panels of Low Tensile Strength, The Structural Engineer,
Vol.56B, 81 - 84.
Timoshenko, S.P., and Woinowsky-Krieger, S., 1959, Theory of Plates and Shells, Second
Edition, McGraw-Hill International Editions, Singapore.

page
66
Table 1: Theoretical and Experimental Failure Loads
moment coefficient, α Wk (kN) Tests (Lawrence) (kN)
Test # B.C. fkx H t L µ Asp. ratio Ex Ey β YL elastic Strip Fracture Li YL Elastic Strip Frac. LinFull crack Initial crackUltimate
8 1 1.61 3 0.11 6 0.638 0.500 16.93 21.29 0.3003 0.024 0.0464 0.0270 0.0394 3.76 1.94 3.35 2.29 3 1.6 3.2
12 1 0.94 2.5 0.112 2.5 0.420 1.000 13.85 19.34 0.4318 0.061 0.0479 0.0881 0.0802 5.15 6.56 3.57 3.92 8.6 7.6 8.6
18 1 1.03 2.5 0.109 3.75 0.393 0.667 21.09 26.09 0.3088 0.043 0.0498 0.0473 0.0670 3.37 2.91 3.07 2.17 4.9 2.9 4.9
22 1 1.16 2.5 0.111 5 0.485 0.500 18.2 24.15 0.2761 0.029 0.0464 0.0279 0.0449 3.29 2.05 3.42 2.12 4.7 3.1 4.7
27 1 1.91 2.5 0.109 6 0.845 0.417 17.23 22.11 0.2932 0.018 0.04292 0.0189 0.0284 5.84 2.45 5.55 3.70 3.1 1.9 3.1
32 1 2.31 3 0.109 6 0.760 0.500 20.95 30.43 0.3374 0.0218 0.0464 0.0263 0.0325 5.83 2.74 4.84 3.92 3.5 1.7 3.5
6 2 1.71 3 0.11 6 0.810 0.500 19.05 20.81 0.01 0.0158 0.0173 9.58 6.06 5.53 4.4 1.9 8
7 2 1.46 3 0.11 6 0.699 0.500 18.46 20.67 0.011 0.0158 0.0177 7.44 5.18 4.61 4.4 2.3 8.1
13 2 1.31 2.5 0.112 2.5 0.413 1.000 14.07 17.35 0.03 0.0231 0.0590 14.61 18.97 7.43 9.1 9.1 12.1
20 2 0.93 2.5 0.109 3.75 0.416 0.667 16.18 22.29 0.021 0.0203 0.0313 6.24 6.45 4.19 5.2 3.6 11.6
23 2 1.54 2.5 0.111 5 0.545 0.500 18.72 23.67 0.015 0.0158 0.0183 8.43 8.01 6.90 5.5 2.9 9.9
31 2 1.98 2.5 0.109 6 0.768 0.417 18.6 23.93 0.008 0.0125 0.0128 13.61 8.71 8.53 4.2 1.8 6.9
33 2 2.04 3 0.109 6 0.957 0.500 20.63 28.03 0.0095 0.0158 0.0168 11.81 7.10 6.67 3.3 1.6 4.7
37 2 1.48 2.5 0.109 2.5 1.050 1.000 20.12 27.29 0.0205 0.0231 0.0406 22.87 20.30 11.54 10.7 9 24
9 3 1.51 3 0.112 6 0.457 0.500 15.92 21.33 0.6289 0.022 0.0142 0.0267 0.0165 3.99 6.18 3.29 5.32 2.5 1.6 5.5
14 3 1.31 2.5 0.112 2.5 0.519 1.000 15.77 19.93 0.4655 0.0355 0.0244 0.0703 0.0361 12.34 17.96 6.24 12.13 11.3 11.3 20
19 3 0.89 2.5 0.109 3.75 0.432 0.667 18.74 24.63 0.5802 0.028 0.0179 0.0431 0.0249 4.48 7.00 2.91 5.03 4.8 4.3 6.7
24 3 1.6 2.5 0.111 5 0.512 0.500 18.2 22.87 0.6246 0.021 0.0142 0.0262 0.0163 6.26 9.26 5.01 8.09 5 2.9 6.4
30 3 1.82 2.5 0.109 6 0.704 0.417 17.91 22.4 0.6291 0.015 0.0125 0.0183 0.0115 6.67 8.01 5.46 8.74 4.7 2.3 4.7
34 3 2.18 3 0.109 6 0.994 0.500 18.22 25.71 0.5431 0.014 0.0142 0.0228 0.0123 8.57 8.44 5.27 9.76 3 2.2 3.9
38 3 1.34 2.5 0.109 2.5 0.832 1.000 18.22 27.95 0.3805 0.0306 0.0244 0.0556 0.0241 13.87 17.40 7.63 17.60 9 9 18.8
16 4 0.84 2.5 0.109 2.5 0.336 1.000 19.16 24.49 0.0493 0.0972 0.1657 5.40 2.74 1.61 8 8 14
21 4 0.93 2.5 0.109 3.75 0.460 0.667 18.9 24.32 0.0415 0.0558 0.0998 3.16 2.35 1.31 3.9 3.6 4
25 4 1.48 2.5 0.111 5 0.550 0.500 18.28 23.96 0.035 0.0293 0.0685 3.47 4.15 1.78 2.6 2.6 3.9
29 4 1.85 2.5 0.109 6 0.726 0.417 17.08 21.57 0.03 0.01855 0.0494 3.39 5.49 2.06 2.4 2.4 3.5
35 4 1.53 3 0.109 6 0.760 0.500 18.41 24.82 0.0314 0.0293 0.0584 2.68 2.87 1.44 1.7 1.7 2.5
10 5 1.08 3 0.112 6 0.400 0.500 16.04 20.51 0.031 0.022 0.0893 2.02 2.85 0.70 1.7 1.7 1.7
15 5 1.31 2.5 0.112 2.5 0.493 1.000 14.65 20.49 0.083 0.039 0.1683 5.28 11.24 2.60 7.8 7.8 7.8
17 5 1.02 2.5 0.109 3.75 0.451 0.667 19 25.06 0.067 0.03 0.1234 2.14 4.79 1.16 3.4 3.4 3.4
26 5 1.46 2.5 0.111 5 0.503 0.500 19.08 24.25 0.056 0.022 0.0832 2.14 5.45 1.44 2.7 2.7 2.7
28 5 1.83 2.5 0.109 6 0.728 0.417 18.91 23.92 0.043 0.022 0.0576 2.34 4.58 1.75 2.3 2.3 2.3
36 5 1.6 3 0.109 6 0.904 0.500 19.52 25.86 0.047 0.022 0.0656 1.87 4.00 1.34 1.9 1.9 1.9

Boundary Condition reference number (B.C.): H - Wall height

1 = all sides simply supported t - wall thickness
2 = all sides built-in L - length of wall
3 = simply supported top and bottom, built-in sides µ − orthogonal strength ratio = fkx / fky
4 = simply supported bottom, free top, built-in sides Asp. Ratio = H / L
5 = simply supported bottomand sides, free top Ex - horizontal beam elastic modulus
Ey - vertical beam elastic modulus
β − a factor derived for fracture line method by Sinha, see reference 3

Table 2 : R atios of T heoretical to Experim ental Failure Loads

Test B.C . Full-C rack Pattern Load Initial-C rack Load U ltim ate Load
no. YL Elastic Strip Fractur YL Elastic Strip Fracture Li YL Elastic Strip Fracture Lin
27 1 1.25 0.65 1.12 0.76 2.35 1.21 2.09 1.43 1.17 0.61 1.05 0.72
8 1 0.60 0.76 0.42 0.46 0.68 0.86 0.47 0.52 0.60 0.76 0.42 0.46
22 1 0.69 0.59 0.63 0.44 1.16 1.00 1.06 0.75 0.69 0.59 0.63 0.44
32 1 0.70 0.44 0.73 0.45 1.06 0.66 1.10 0.68 0.70 0.44 0.73 0.45
18 1 1.88 0.79 1.79 1.19 3.07 1.29 2.92 1.94 1.88 0.79 1.79 1.19
12 1 1.67 0.78 1.38 1.12 3.43 1.61 2.85 2.30 1.67 0.78 1.38 1.12
31 2 2.18 1.38 1.26 5.04 3.19 2.91 1.20 0.76 0.69
6 2 1.69 1.18 1.05 3.23 2.25 2.01 0.92 0.64 0.57
7 2 1.61 2.08 0.82 1.61 2.08 0.82 1.21 1.57 0.61
23 2 1.20 1.24 0.81 1.73 1.79 1.16 0.54 0.56 0.36
33 2 1.53 1.46 1.25 2.91 2.76 2.38 0.85 0.81 0.70
20 2 3.24 2.07 2.03 7.56 4.84 4.74 1.97 1.26 1.24
13 2 3.58 2.15 2.02 7.38 4.44 4.17 2.51 1.51 1.42
37 2 2.14 1.90 1.08 2.54 2.26 1.28 0.95 0.85 0.48
30 3 1.59 2.47 1.31 2.13 2.49 3.86 2.05 3.33 0.72 1.12 0.60 0.97
9 3 1.09 1.59 0.55 1.07 1.09 1.59 0.55 1.07 0.62 0.90 0.31 0.61
24 3 0.93 1.46 0.61 1.05 1.04 1.63 0.68 1.17 0.67 1.04 0.43 0.75
34 3 1.25 1.85 1.00 1.62 2.16 3.19 1.73 2.79 0.98 1.45 0.78 1.26
19 3 1.42 1.70 1.16 1.86 2.90 3.48 2.37 3.80 1.42 1.70 1.16 1.86
14 3 2.86 2.81 1.76 3.25 3.89 3.84 2.39 4.43 2.20 2.17 1.35 2.50
38 3 1.54 1.93 0.85 1.96 1.54 1.93 0.85 1.96 0.74 0.93 0.41 0.94
29 4 0.67 0.34 0.20 0.67 0.34 0.20 0.39 0.20 0.11
25 4 0.81 0.60 0.34 0.88 0.65 0.36 0.79 0.59 0.33
35 4 1.34 1.60 0.68 1.34 1.60 0.68 0.89 1.06 0.46
21 4 1.41 2.29 0.86 1.41 2.29 0.86 0.97 1.57 0.59
16 4 1.58 1.69 0.85 1.58 1.69 0.85 1.07 1.15 0.58
28 5 1.19 1.68 0.41 1.19 1.68 0.41 1.19 1.68 0.41
10 5 0.68 1.44 0.33 0.68 1.44 0.33 0.68 1.44 0.33
26 5 0.63 1.41 0.34 0.63 1.41 0.34 0.63 1.41 0.34
36 5 0.79 2.02 0.53 0.79 2.02 0.53 0.79 2.02 0.53
17 5 1.02 1.99 0.76 1.02 1.99 0.76 1.02 1.99 0.76
15 5 0.99 2.11 0.71 0.99 2.11 0.71 0.99 2.11 0.71

page
77
 F ig u re 1 : C o m p a ris o n w ith F u ll-c ra c k lo a d
2 5 .0 0

B .C . 3

2 0 .0 0

Fa
ilu 1 5 .0 0
re
Lo
ad
(k
N)
1 0 .0 0

5 .0 0

0 .0 0
8 12 18 22 27 32 6 7 13 20 23 31 33 37 9 14 19 24 30 34 38 16 21 25 29 35 10 15 17 26 28 36
T e st N u m b e r
F ixity sim ply su p p orted fix e d all e d g e s B .C .4 B .C .
: 5
YL E la stic S trip F ra c . L in e F u ll cra c k (L a w re n c e )

F ig u re 2 : C o m p a r is o n w ith In itia l C r a c k L o a d
2 5 .0 0

B .C . 3

2 0 .0 0
Predicted Load (kN)

1 5 .0 0

1 0 .0 0

5 .0 0

0 .0 0
8 12 18 22 27 32 6 7 13 20 23 31 33 37 9 14 19 24 30 34 38 16 21 25 29 35 10 15 17 26 28 36
Test N um ber

F ix ity s im p ly s u p p o rte d fixe d a ll e d g e s B .C .4 B .C .5

YL E la s tic S trip F ra c . L in e In itia l c ra c k (L a w re n c e )

F ig u r e 3 : C o m p a r is o n w ith U ltim a te L o a d
3 0 .0 0

B .C . 3
2 5 .0 0

2 0 .0 0
Predicted Load (kN)

1 5 .0 0

1 0 .0 0

5 .0 0

0 .0 0
8 12 18 22 27 32 6 7 13 20 23 31 33 37 9 14 19 24 30 34 38 16 21 25 29 35 10 15 17 26 28 36
Test Num ber

F ix ity s im p ly s u p p o rte d fix e d a ll e d g e s B .C .4 B .C .5

YL E la s tic S trip F r a c . L in e U ltim a te L o a d ( L a w r e n c e )

page
88
 Figure 4: Simply Supported Panels

10.00

9.00

8.00

7.00

6.00
Failure Load

5.00

4.00

3.00

2.00

1.00

0.00
8 12 18 22 27 32
Panel Number

YL Elastic Strip Frac. Line Lawrence

Figure 5: Panels with All Edges Fixed

25.00

20.00

15.00
Failure Load

10.00

5.00

0.00
6 7 13 20 23 31 33 37
Panel Number

YL Elastic Strip Lawrence

Figure 6: Boundary condition 3

20.00

18.00

16.00

14.00

12.00
Failure Load

10.00

8.00

6.00

4.00

2.00

0.00
9 14 19 24 30 34 38
Panel Number

YL Elastic Strip Frac. Line Lawrence

page
99
 Figre 7: Boundary Condition 4

9.00

8.00

7.00

6.00
Failure load

5.00

4.00

3.00

2.00

1.00

0.00
16 21 25 29 35
Panel num ber

YL Elastic Strip Lawrence

Figure 8: Boundary Condition 5

12.00

10.00

8.00
Failure load

6.00

4.00

2.00

0.00
10 15 17 26 28 36
Panel num ber

YL Elastic Strip Lawrence

F ig u re 9 : R a tio o f th e o ry to T e s t R e s u lts (fu ll c ra c k p a tte rn )

4 .0 0

3 .5 0 B .C . 3

3 .0 0

2 .5 0
Ratio

2 .0 0

1 .5 0

1 .0 0

0 .5 0

0 .0 0
8 12 18 22 27 32 6 7 13 20 23 31 33 37 9 14 19 24 30 34 38 16 21 25 29 35 10 15 17 26 28 36
Test N um ber

s im p ly s u p p o rte d fixe d a ll e d g e s B .C .4 B .C .5
F ix ity
YL E la s tic S trip F ra c tu re

page1010