Professional Documents
Culture Documents
Scaffolds Under Eccentric Loads
Scaffolds Under Eccentric Loads
Scaffolds Under Eccentric Loads
article info a b s t r a c t
Article history: Steel scaffolds collapse quite often in many places with a considerable number of reported casualties,
Received 25 August 2007 but their behaviour has not been studied to the extent of many other permanent structures. This paper
Accepted 25 March 2008 investigates the effect of eccentric loads on steel scaffolding systems used in construction sites. The type
of scaffold considered here is the door-shaped steel scaffold with an inner reinforced gable sub-frame. The
Keywords: single-side cross-brace scaffolding systems with various eccentric loads are mainly focused on two issues,
Critical load
namely, the unrestrained boundary and the removal of cross-braces at the access location. This study
Eccentric load
Scaffold
shows that regardless of the lowest layer of cross-brace in a scaffold being removed or not, the critical
Scaffolding load of a scaffolding system under an eccentric load is the lowest, whereas that of scaffolding system
under a concentric load is the maximum. If the bottom jack base of a scaffolding system in construction
sites is strengthened to a fixed end, the critical load of this scaffolding system will be greatly increased.
If a scaffolding system is erected more than 8 stories high, the critical load of the scaffolding system
with the fixed end base can be increased to 2.4 times that with the hinged base. However, whether the
cross-braces at the lowest story of a scaffolding system are removed or not, the simulated scaffolding test
indicates that the critical load of a used scaffolding system under the eccentric load is the lowest and its
load reduction also appears significant.
© 2008 Published by Elsevier Ltd
Fig. 1. Assembly of steel scaffold for using as working scaffold in construction site. Fig. 3. Scaffold collapse occurred at a construction site in southern Taiwan.
This study mainly focuses on the effect of eccentric load on the This study is carried out by using a 3-dimensional second-order
critical load of the door-shaped steel scaffold. The steel scaffolding elastic semi-rigid joint analysis. Equivalent lateral notional forces
system considered in this study is based on the assembly obtained of 0.1%–0.5% of the total vertically static loads are used to simulate
from site investigations in Taiwan. Basically, the cross-braces of the the initial imperfection of the scaffolding system. The computer
scaffolding system discussed here are setup on the single-side only. program GMNAF (i.e. Geometric and Material Nonlinear Analysis of
Furthermore, two more common setups of a scaffolding system Frames) developed by Chan and his co-workers [6,7] is used in this
in construction sites are investigated. Namely, they are (A) the study. In GMNAF, the pointwise equilibrating polynomial element
conditions of different boundaries, and (B) removal of the lowest is adopted and a second-order elastic/plastic analysis can be used.
layer cross-braces. The effects of the two cases on the steel scaffold In plastic analysis, the sectional strength is kept unchanged during
are described respectively as follows: the iteration for equilibrium when the section capacity factor
A. Different boundary conditions (φ = F /py A + M/Mc in which φ is the section capacity factor, F
After the assembly of a scaffolding system is completed in is the applied axial force, py is the design strength, A is the area,
construction sites, the uppermost layer of the scaffolding structure M is the applied moment at the section and Mc is the sectional
is commonly erected in an unrestrained condition. However, due moment capacity) is equal or greater than 1. Compatibility at the
424 J.-L. Peng et al. / Journal of Constructional Steel Research 65 (2009) 422–435
end nodes and the equilibrium for moment and shear at mid-
span as well as end nodes are maintained, which overcomes the
error associated with a conventional displacement-based straight
cubic finite beam-column element satisfying equilibrium and
compatibility at the end nodes only. For details, references 6 and
Fig. 5. Load vs. Deformation Relations of Structure in Different Analyses.
7 should be referred.
Six boundary conditions are imposed to evaluate the shape
functions of the system. As shown in Fig. 4, four boundary
conditions satisfy the compatibility as follows.
4. Test setup
4.1. Basic arrangement of scaffold Fig. 8. Arrangement of scaffold with unstrained boundary and eccentric load.
4.2. Arrangement of lateral unrestrained boundary B at the bottom is similar to the displacement condition of point B
at the top in Fig. 9(I). If the own weight of the scaffolding specimen
Fig. 8 shows the assembly of the scaffolding system with an AB is not taken into consideration, the critical load of structure AB
unrestrained boundary condition provided with lateral movement in Fig. 9(I) resembles the condition shown in Fig. 9(III).
and under the eccentric load. To simulate an unrestrained The steel scaffold applied in the test is brand new; nevertheless,
boundary condition for the scaffolding system in Fig. 8, two pieces most scaffolds used in construction sites have been used several
of steel plate are provided at the bottom of vertical posts of the times. For simulating the worst condition of the steel scaffold
scaffold. Further, between the two steel plates, 9 steel balls are used in construction sites, the new steel scaffolds are loaded twice
installed in a cross-frame window shape, i.e. 3 balls provided at in the study. This means that the steel scaffold is loaded to the
each of the 3 rows equally spaced as shown in Fig. 8. This type ultimate load in its first loading test. After unloading, the tested
of setup enables the bottom steel plate to be movable after the scaffold is re-setup to the original configuration and re-loaded to
scaffolding structure is under the vertical axial load. the ultimate load for the second time. In this loading arrangement,
Fig. 9 indicates the different conditions of the boundary’s the second ultimate load of the tested scaffold is considered as
lateralmovement during the test. Fig. 9(I) shows the lateral ideally the lower bound of the critical load of the used scaffold in
movement at the top layer of the tested specimen AB. Nevertheless, real construction sites.
due to expansion in a mono-axial direction of the hydraulic thrust
head at a fixed position, the loading position cannot move along the 4.3. Arrangement of eccentric load
lateral displacement during the test as shown in Fig. 9(II). Fig. 9(II)
differs greatly from the expected phenomenon in Fig. 9(I). Fig. 10 shows the top view of the setup of scaffolding systems
Fig. 9(III) shows the loading condition of the tested specimen AB and different types of eccentric loads on the tested steel scaffold
being placed upside down. In Fig. 9(III), the displacement of point discussed in this paper. The letters L, R, T, B shown in Fig. 10(I)
426 J.-L. Peng et al. / Journal of Constructional Steel Research 65 (2009) 422–435
Table 1
Loading test result of basic setup scaffolds
Setup model Two-stories Three-stories
Double-side cross-brace Single-side cross-brace Double-side cross-brace Single-side cross-brace
Test No. A 109.1 (62.9) 103.1 (47.7) 112.9 (46.3) 68.3 (46.0)
B 126.3 (75.3) 102.8 (50.0) 95.3 (55.3) 72.2 (50.0)
Test Setup
represent the loading locations on the Left, Right, Top and Bottom,
respectively, of the steel scaffold. As shown in Fig. 10(I), two steel
scaffolds are placed at the top and the bottom and a pair of cross-
braces are setup on the right side. On the left side, no cross-
brace is installed. To compare the buckling strength of scaffolds,
a concentric loading test is applied. Fig. 10(I) is the concentric
loading arrangement.
In Fig. 10(I), “Center” indicates that the applied load is applied
to the center of the scaffold; i.e. with respect to the xy coordinate,
the loading coordinate lies at (d/2, L/2). The eccentric load in
Fig. 11. Deformation condition of concentric loading test for basic setup scaffold
Fig. 10(II) is T/4 with the eccentric load placed at the position (d/2, with single- side cross-brace and restrained boundary.
3L/4) of the xy coordinate. The setup for L/4 eccentric loading test is
the same as T/4 eccentric loading test. The difference between the
two tests is the loading location of the test L/4 placed at (d/4, L/2) The discrepancy between single and double cross-braces of steel
of xy coordinate shown in Fig. 10(III). Similar to the TL/4 eccentric scaffolds is first investigated in this study. Fig. 11 shows the test
loading test, the load was placed at asymmetric location along result of a 2-story basic setup scaffold with single-side cross-brace
the upper & lower and left & right directions, i.e. at the location and restrained boundaries at the top and bottom. A restrained
(d/4, 3L/4) of xy coordinate shown in Fig. 10(IV). These 4 types of horizontal boundary condition at the top and bottom and applying
eccentric loading tests are separately processed with two types of a concentric load are considered for these basic steel scaffolds
cross-braces on the single-side and without cross-brace which is setup in the tests. The critical load of the tested basic setup scaffold
to allow easy access by workers in construction sites. is considered as a basis of comparison with other conditions in this
paper.
5. Test, analysis and discussion The test results of the basic setup scaffolds are shown in Table 1
which reveals that the averaged critical load of the 2-story scaffold
5.1. Test result of basic setup scaffold with cross-braces at both sides is 117.7 kN, and the critical load
obtained from second-time load is 69.1 kN, which is approximately
The tested scaffold studied in this research is the door-shaped 41% lower. The averaged critical load for the 2-story scaffold with
steel scaffold with an inner reinforced gable-type sub-frame. only single-side cross-braces is 102.9 kN, and the failure load
J.-L. Peng et al. / Journal of Constructional Steel Research 65 (2009) 422–435 427
Table 2
Analyzed joint stiffness of scaffold under eccentric loads and unrestrained boundary
Load types Setup model Test No. Test results (kN) Analytic results (kN)
Test strength average 1569.6! Error (%) 1765.8! Error (%) 1962! Error (%)
Table 3
Tested critical loads of scaffolds with various single-side cross-braces under eccentric loads
Setup model Without eccentricity T/4 eccentricity
Single-side cross-brace Removal of cross-braces in lowest layer Single-side cross-brace Removal of cross-braces in lowest layer
Test No. A 62.2 (33.8) 37.5 (26.7) 40.8 (30.3) 34.8 (5.1)
B 61.8 (32.8) 33.8 (21.0) 43.6 (17.1) 28.2 (5.8)
B. T/4 Eccentricity
A T/4 eccentric loading arrangement is as shown in Fig. 10(II).
This shows the damage to scaffolding structures after being placed
withloads that are close to the topside where damage due to
deformation of the steel scaffold occurred. The failure mode of
T/4 loading case is similar to the concentric loading condition. The
averaged critical load of the first-time load is 42.2 kN; compared
to the concentric loading case, the critical load has been reduced
by 32% (= (62.0 − 42.2)/62.0). The averaged critical load of the
second-time load is 23.7 kN. When compared to the first-time load,
the critical load has been reduced by 44% (= (42.2 − 23.7)/42.2).
The critical load of the T/4 loading case is shown in Table 3. Fig. 14
shows the applied vertical load vs. the vertical displacement for the
first-time load, and Fig. 15 indicates the applied vertical load vs. the
vertical displacement for the second-time load. The curve tendency
of the two figures is similar when compared with the condition of
scaffolds with a concentric loading condition.
C. L/4 Eccentricity
Fig. 10(III) indicates the arrangement of a L/4 eccentric loading
case. The deformation of the L/4 eccentric loading test indicates
that the failure occurred mainly in the scaffold near the left side (L
side) without cross-braces. This failure type is similar to that of the
concentric loading test. The averaged critical load of a L/4 eccentric
load is slightly less than a T/4 eccentric load. When compared to
the concentric loading case of the same installation condition, the
critical load of a T/4 eccentric load is reduced by approximately 35%
(= (62.0 − 40.1)/62.0) as shown in Table 3. The averaged critical
Fig. 13. Deformation condition of concentric loading test for scaffold with single- load of the second-time load is 23.7 kN; compared to the first-time,
side cross-brace. it is reduced by approximately 41% (= (40.1 − 23.7)/40.1).
D. TL/4 Eccentricity
overall scaffolding structure. Furthermore, the critical load of the A TL/4 eccentric loading arrangement is shown in Fig. 10(IV).
second loading case compared to the first-time load has an average In this case, the test result shows that the small deformation in
difference of approximately 46% (= (62.0 − 33.3)/62.0). This the scaffolding system initially occurred close to the side near
implies that the variation between brand new and old sets of the loading jack. When the vertical load reached its ultimate
scaffolds is significant to the critical loads of the scaffolds as shown maximum load, a large deformation occurred at a steel plate of the
in Table 3. bottom boundary of the scaffolding system. The type of damage
J.-L. Peng et al. / Journal of Constructional Steel Research 65 (2009) 422–435 429
Fig. 14. T/4 eccentric load vs. axial deflection of scaffold with single-side cross-
Fig. 16. Analysis and test results of scaffolding system with single-side cross-braces
braces for 1st – time load.
and top guided-end boundary (bottom stiffness: 1765.8 kN cm/rad).
Fig. 15. T/4 eccentric load vs. axial deflection of scaffold with single-side cross-
braces for 2nd – time load. Fig. 17. Analysis results of scaffolding system with single-side cross-braces and top
free-end boundary (hinged bottom).
is similar to the case of a concentric loading condition, and the
obtained averaged ultimate critical load is 32.0 kN. Compared under concentric loads. The critical load is 43.1 kN for the 2-story
to the critical load of the concentric loading condition with a case and 41.6 kN for the 12-story case under the T/4 eccentric
restrained boundary, the critical load is reduced by approximately loading condition. For the L/4 eccentric loading case, the critical
69% (= (102.9 − 32.0)/102.9). Similarly, the deformation of loads are 40.5 kN and 39.7 kN for the 2-story and 12-story cases,
the second-time load was gradually increased from the originally respectively. For the TL/4 eccentric loading case, the critical load
deformed position until the scaffolding failed. The second-time is 31.5 kN for the 2-story and 30.9 kN for the 12-story. According
averaged critical load is 8.8 kN; compared to the first-time load, the to the tendency of the 4 curves shown in Fig. 16, the critical loads
critical load has been reduced for 73% (= (32.0 − 8.8)/32.0). All the
of the scaffolding systems slightly decrease with the increase in
failure loads of the scaffolding systems are tabulated in Table 3.
the number of stories of the scaffolding systems. The slope of
5.2.2.2. Comparison of analysis and test results. Fig. 16 shows the these curves is almost a constant and the tendency differs greatly
computed critical loads of 2-story to 12-story scaffolding systems. from that of the critical loads of steel scaffolds used as formwork
In the analysis, all assumptions are based on the conditions in the supports.
experimental tests; i.e. the scaffolding top is laterally unrestrained, Fig. 17 shows the critical loads of the 4-types with the
joint stiffness between scaffolds is 784.8 kN cm/rad, and semi- conditions like those found in real construction sites; i.e. the top
rigid stiffness of the base is 1765.8 kN cm/rad. In Fig. 16, the of the scaffolds is free, the joint stiffness between scaffolds is 784.8
circle symbol shows the test results of 2-story scaffolds for 4 kN cm/rad, and the bottom base is hinged. The biggest difference
types of eccentric loading conditions. From Fig. 16 and Table 2, between Figs. 17 and 16 lies on the assumption of the lowest
the computed critical loads of 2-story scaffold for the 4 types of bottom part of the scaffold. The base stiffness of scaffold in the
eccentric loads are close to the experimental tested failure loads. construction site is simulated as a lower bounded “hinge”. There
In Fig. 16, the computed critical load is 58.3 kN for the 2-story is no processed test for the condition shown in Fig. 17; the analysis
scaffolding system and 57.0 kN for the 12-story scaffolding system is processed purely numerically. The analysis is conducted with
430 J.-L. Peng et al. / Journal of Constructional Steel Research 65 (2009) 422–435
Table 4
Tested critical load comparisons to brand new and simulated worst scaffolds
Single-side cross-brace Removal of cross-braces in lowest layer
1st load 2nd load Reduction (%) 1st load 2nd load Reduction (%)
TL/4 Eccentricity(kN)
32.0 8.8 73 18.1 4.4 76
C. L/4 Eccentricity
Fig. 10(III) shows the L/4 eccentric loading arrangement. After
the scaffolding system is loaded, a large deformation occurs on the
left side (L side) of the scaffold without cross-braces. Moreover, the
scaffolding posts close to the loading location deformed first. When
the applied load was close to the ultimate load, a large deformation
occurs simultaneously in the two scaffolds and in their out-of-
plane directions.
As shown in Tables 3 and 4, the averaged critical load of the
structure is 23.2 kN. Compared to the scaffold with single-side
cross-braces and under the same eccentric loading condition, the
averaged critical load is reduced by approximately 42% (= (40.1 −
23.2)/40.1). If compared to the same assembly with removal of
cross-braces at the access location and under a concentric loading
condition, the averaged critical load is reduced by 35% (= (35.7 −
23.2)/35.7). Comparing the critical load of the second-time load
with that of the first-time load, the averaged critical load is reduced
by 47% (= (23.2 − 12.3)/23.2) as shown in Tables 3 and 4.
D. TL/4 Eccentricity
The TL/4 eccentric loading arrangement is shown in Fig. 10(IV).
Fig. 19. Analysis and test results of scaffolding system with removal of cross-
The test result reveals the damaged model of structure; i.e. the braces at the access location and top guided-end boundary (bottom stiffness: 1765.8
lateral deformation of the system also occurs in the out-of- kN cm/rad).
plane direction of the scaffold. The averaged critical load of the
scaffolding system is 18.1 kN. In Table 3, tests A and B have a big
variation due to the hard setup in tests as well as a variation in
imperfection which is somehow random in nature. Compared to
the assembly for single-side cross-braces under the same eccentric
condition, the averaged critical load is reduced by 43% (= (32.0 −
18.1)/32.0). This indicates that under TL/4 eccentric load, the
removal of cross-braces at the access location has a big impact to
the critical load of the overall scaffolding structure.
Comparing to the scaffolding system with the removal of cross-
braces at the access location under the concentric load, the critical
load of the scaffolding system under the TL/4 eccentric load is
reduced by 49% (= (35.7 − 18.1)/35.7). The deformation of the
second-time load similarly occurs as the deformed position of the
first-time load and the averaged critical load of the scaffolding
system is 4.4 kN. Comparing the critical load of the second-time
load with that of the first-time load, the averaged critical load is
reduced by 76% (= (18.1 − 4.4)/18.1) as shown in Tables 3 and 4.
Fig. 21. Deformation shapes of various stories of scaffolding systems with bottom hinged base and top free-end boundary.
Fig. 23. Deformation shapes of various stories of scaffolding systems with bottom fixed-end base and top free-end boundary.
Table 5
Analyzed critical loads of scaffolding systems with top unrestrained boundary and
bottom hinged/fixed-end bases
No. of stories Hinged base (kN) Fixed-end base (kN)
2 22.8 92.4
3 22.8 75.7
4 22.8 65.8
5 22.7 60.6
6 22.6 58.0
7 22.6 56.3
8 22.6 55.4
9 22.6 54.7
10 22.6 54.3
11 22.6 54.0
12 22.4 53.9
are very different. The deformation of fixed base case in Fig. 23,
except 2-story scaffold, was found to be outward at odd numbers
of scaffolding stories inward at even numbers of scaffolding stories.
Further, Fig. 23 also reveals that the deformation of top scaffold
is larger than at the bottom. This deformation mode is totally
different from the case of bases hinged shown in Fig. 21, which Fig. 24. Comparison of analyzed critical loads between single-side cross-braces and
has almost the same deformation in the scaffolding units above the removal of cross-braces at access location with base stiffness 1765.8 kN cm/rad.
second scaffolding level.
the scaffolding base is placed on the concrete slab or wooden
5.4. Effect of removing cross-braces at access floor whose stiffness is zero like a hinged base, the joint stiffness
between scaffolds is 784.5 kN cm/rad, and the top boundary is
As previously discussed, steel scaffolds are frequently setup movable and rotational as a free end.
with a single row close to the façade of building during
construction. Moreover, the cross-braces near the façade are 5.4.1. Semi-rigid joint base
always removed for easy access so that the scaffolding system Fig. 24 shows the analysis results based on the laboratory
is fabricated with single-side cross-braces [11]. Furthermore, for boundary conditions. The symbols, N and in Fig. 22, indicate the
the convenient entrance of workers and materials in construction analyses of scaffolding systems with removal of cross-braces at
sites, the cross-braces on the lowest story at access location are access location under “concentric load” and “TL/4 eccentric load”
frequently removed. For comparing the variations of the critical respectively. Further, the symbol 4 and indicate the two types
loads of scaffolding systems between before and after removing of analysis results under the condition of single-side cross-braces.
the cross-braces at the lowest story scaffold access location, this In Fig. 24, for the concentric load over the scaffolding system
study investigates the critical load of scaffold under concentric with single-side cross-braces, the critical loads are 58.3 kN for 2-
load and under the TL/4 eccentric load. The concentric load can be story and 57.0 kN for 12-story scaffolds. The average critical load
considered as the upper bound critical load, and the TL/4 eccentric of the all scaffolds of 2-story to 12-story is 57.7 kN. Furthermore,
load is taken as the strength lower bound. The actual condition in for the concentric load over the scaffolding system with removal
construction sites should lie within these two bounds. of cross-braces at access location, the critical loads are 35.7 kN for
The boundary conditions of the scaffolding systems are also 2-story and 35.2 kN for 12-story scaffolds. The average critical load
assumed in two types in the comparison. One is the laboratory of the all scaffolds of 2-story to 12-story scaffolds is 35.5 kN. For the
boundary condition; i.e. the scaffolding base is placed on a steel convenience of comparison, the averaged critical load is adopted.
plate whose base stiffness is 1765.8 kN cm/rad, the joint stiffness After comparing the averaged critical loads of the scaffolds under
between scaffolds is 784.5 kN cm/rad, and the top boundary is the conditions of removing cross-braces at the access location and
laterally movable but not allowed to rotate. The other case is single-side cross-braces, it is found that the difference values of
the construction site boundary condition; i.e. in construction site scaffolds under removing and without removing cross-braces is
434 J.-L. Peng et al. / Journal of Constructional Steel Research 65 (2009) 422–435
Table 6
Ratios of critical loads of scaffolding systems with single-side cross-braces and removal of cross-braces at access location
:
1. ! the analytic stiffness of base is 1765.8 kN cm/rad (with its top boundary movable & un-rotational)
2. !! with its top boundary movable & rotational.
Fig. 26. Failure shapes of Scaffolding systems with single-side cross-braces and
Fig. 25. Comparison of analyzed critical loads between single-side cross-braces and
different Bases.
removal of cross-braces at access location with hinged base.
system with single-side cross-braces is 12.7 kN. The load for 12-
0.62 (= 35.5/57.7). The critical load is reduced by 38% as shown story scaffold is 12.3 kN and the averaged critical load is 12.5 kN.
in Table 6. Moreover, for the condition of removing cross-braces at access
In Fig. 24, for the TL/4 eccentric load over the scaffolding system location, the critical load of 2-story scaffold is 7.9 kN and the load
with single-side cross-braces, the critical load of 2-story scaffold of 12-story scaffold is 7.7 kN. The averaged critical load of total
with single-side cross-braces was obtained as 31.5 kN and as 30.9 scaffolding systems is 7.8 kN. The ratio of the averaged critical loads
kN for 12-story. The averaged critical load of all scaffolds from 2- of the two conditions of removing and without removing cross-
to 12-story scaffolds is 31.2 kN. However, for the TL/4 eccentric braces is 0.62 (= 7.8/12.5). The critical load is reduced by 38% as
load over the scaffolding system with removal of cross-braces at shown in Table 6.
access location, the 2-story is 26.4 kN and the 12-story is 25.9 Summarizing the above findings, for a scaffolding system under
kN. The average critical load of all scaffolds from 2-story to 12- the TL/4 eccentric load and after removal of the cross-braces at the
story scaffolds is 26.2 kN. The ratio of averaged critical loads of the lowest layer at the access location, a great reduction in the critical
scaffolds with removal cross-braces at access location to that with load of scaffolding system will result. The practice of removing
single-side cross-braces is 0.84 (= 26.2/31.2). The critical load is the scaffoldingsystem’s cross-braces at lowest layer for easy access
reduced by 16% as shown in Table 6. should be avoided.