Ultimate Load Capacity Analysis of Steel Scaffoldings Using

Direct-Analysis Method
Wiryanto Dewobroto1 and Wawan Chendrawan2

Abstract: For ensuring the safe load capacity of steel scaffoldings, the ultimate load capacity analysis is performed including an experi-
mental test and numerical modeling. The test is relatively costly; therefore, it is used specifically for small configurations. Then, the para-
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

metric methods based on the calibrated numerical modeling are used. The ultimate strength analysis of steel structure requires a
sophisticated nonlinear inelastic computer analysis and cumbersome procedure. As an alternative, elastic second-order computer analysis
and the direct analysis method (DAM) are tried. The loading is given incrementally, and at each stage the available strengths and required
strengths are evaluated. The ultimate limit load is considered achieved when the available capacity in the critical member equals that
required. It has been proven that the calibrated DAM can be used to trace the ultimate load well. It can even be used to evaluate the effec-
tiveness of bracing or lateral support placements. It is important especially for multistory steel scaffoldings that belong to sway frames
because their strengths are significantly influenced by the stability issues. DOI: 10.1061/(ASCE)SC.1943-5576.0000392. © 2018
American Society of Civil Engineers.
Author keywords: Direct analysis method (DAM); Ultimate load capacity analysis; Steel scaffoldings.

Introduction result the validation process can obtain an accurate outcome (Toma
and Chen 1992; Toma et al. 1995).
Building construction projects in Indonesia are generally realized The numerical modeling of structural failure requires nonlin-
as reinforced concrete structures; hence, a fairly large number of ear inelastic computer analysis. Currently, the most popular type
steel scaffoldings are needed in the construction stages. This is a is the FEM-based computer program (Weesner and Jones 2001;
prospective market in which to sell local steel scaffolding prod- Kim et al. 2003; Adam 2013). In practice, this method is rela-
ucts because the prices for similar products imported from abroad tively complicated. The method requires detailed geometric mod-
are more expensive. One of the locally produced steel scaffold- eling; thus, it is suitable for research purposes only and not for
ings that has been marketed is the ring-lock modular scaffolding practical daily design efforts. A simpler alternative is the direct
system (Fig. 1). analysis method (DAM) from AISC (2005, 2010, 2016), which
The successful use of the ring-lock scaffolding system in various requires only an elastic second-order computer analysis. DAM is
construction projects cannot be denied. This system is selected the latest method for designing a steel structure. It is thought to be
because it has been proven in the field (Valerii 2011). To ensure the capable of accurately predicting the ultimate strengths of a struc-
serviceability load capacity for the steel scaffoldings, an ultimate ture under an experimental load test compared with the older
limit state analysis must be performed; thus, the relatively safe max- design method, i.e., the effective length method (ELM) (AISC
imum workload of the steel scaffoldings can be determined. The 2010). As a result, the strategy of steel structural collapse analysis
strengths and reliability of the steel scaffolding system used are with DAM is starting to be the preferred solution (East and Rutz
extremely important because missing the safety margin can become 2016).
the cause of construction failure (Hadipriono and Wang 1987; This paper presents an alternative numerical analysis strategy
Andresen 2012). that is relatively simple using DAM based on the structural model,
The ultimate load capacity analysis is performed using an exper- which has been calibrated first with the experimental test results.
imental loading test in the laboratory with numerical simulation or Furthermore, a parametric concept is used to predict the ultimate
computer modeling. The empirical test is relatively costly and very load capacity of the ring-lock modular steel scaffolding system for
risky; hence, it is used only in one-story steel scaffoldings, which is configurations with many stories.
the smallest module that can be built. The test result is also needed
for validating the numerical modeling of two-story steel scaffold-
Principle of Ultimate Strength Analysis Using the DAM
ings. By comparing the numerical modeling to the empirical test
A numerical analysis using DAM requires only an elastic second-
1
Senior Lecturer, Dept. of Civil Engineering, Univ. Pelita Harapan, order computer analysis (AISC 2016). This is different from the
Tangerang, Banten 15811, Indonesia (corresponding author). ORCID: https:// FEM-based numerical analysis, which requires a nonlinear
orcid.org/0000-0002-9773-0581. Email: [email protected] inelastic computer analysis. This enables the automatic search of
2
Director, PT. Gistama Intisemesta Structural & Civil Engineering
inelastic and nonlinear parameters for tracing the structural
Consultant, Jakarta 11610, Indonesia. Email: [email protected]
Note. This manuscript was submitted on April 12, 2018; approved on
behavior systematically up to its ultimate condition; thus, the col-
May 11, 2018; published online on August 28, 2018. Discussion period lapse load can be estimated. However, DAM cannot automati-
open until January 28, 2019; separate discussions must be submitted for cally trace the structural behavior under the inelastic condition,
individual papers. This paper is part of the Practice Periodical on triggering the collapse as a result of exceeding the material
Structural Design and Construction, © ASCE, ISSN 1084-0680. strength and resistance, which happens because DAM is made for

© ASCE 04018028-1 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

 DAM

General Information
DAM is a new computer-based method for structural steel design
(AISC 2005, 2010, 2016). This method was developed to accurately
calculate the influence of structural loading at the analysis stage and
to remove the requirement to calculate the effective length factor,
K, of the column. This is possible because the stability issue has
been accommodated at the structural analytical stage. This method
can be used for all types of structure in which their elements will be
designed as a beam-column. As for the older design method, ELM,
its use is limited to where it is valid only for a structural system,
which meets the criterion requiring the ratio of maximum second-
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

order drift to maximum first-order drift to be less than 1.5. If the ra-
tio exceeds this criterion, then the steel design will be determined
by DAM (AISC 2005).
Fig. 1. Ring-lock modular scaffolding system.
Stability Parameters in Steel Structures
The parameters influencing the behavior of element stability are
analyzing the stability in structural steel design. The influence is (1) geometric nonlinearities, (2) spread of plasticity, and (3)
mostly determined by the structural geometric condition as a member limit states. These three issues influence the deformation
whole (global), which is why when determining the nominal com- of the structures being loaded (AISC 2005).
pressive strengths using DAM, the effect of K value should be The first parameter is geometric nonlinearities. In slender
omitted with K = 1 (AISC 2016). structures, deformation resulting from loading cannot be ignored.
The basic principle of structural steel design using the LRFD is Therefore, second-order analysis is needed in which the struc-
to ensure that the required strengths (Ru) of the elements due to the tural equilibrium is calculated against the geometric condition on
loading on the structural system cannot be greater than the available deformation. The factors evaluated include the influence of
strengths ( f Rn) of the cross section of the structural element under second-order effect, known as P-D and P-d . When the influence
review. The general design requirement is Ru ≤ f Rn. of geometric nonlinearities is significant, the parameters of initial
Using the LRFD and DAM principles, the structural computer geometric imperfections, in the form of member out-of-straight-
analysis will deliver the information on the required strengths
ness, frame out-of-plumbness, and resulting from material and
(Ru), and the cross-sectional design results in information on the
fabrication tolerances, become crucial.
available strengths ( f Rn). To find the ultimate strengths or maxi- The second parameter is spreading of plasticity. From the fabri-
mum Ru, the incremental loading procedure is required. At each
cation of steel profiles, residual stresses are left in the steel profile as
loading stage, the result is compared with condition Ru % f Rn to
a result of the cooling process and restraint. This reduces the ele-
see if the failure has taken place. When Ru < f Rn, the load can be
ment resistance in case of stability loss.
increased, and when Ru > f Rn, the load given at this stage is too The third parameter is member limit states. The available
great and should be reduced. The ultimate load occurs when the
strengths of a structural element are determined by one or more
Ru % f Rn condition is met. The analysis strategy and design at
of its limit state conditions, such as material yielding, local buck-
each loading stage is basically identical to the structural steel
ling, global buckling in the forms of flexural buckling, torsional
design strategy.
buckling, or torsional-flexural buckling, depending on the cross
The structural modeling for numerical analysis using DAM is
section conditions. The procedure to calculate the available
relatively simple and just uses the Line element [one-dimensional
strengths of members and connections will be done in accordance
(1D) finite-element analysis formulation] or the commonly called
with the provisions of AISC (2010), as applicable, with no further
Frame element (CSI 2011). This is different from the FEM based
consideration of overall structure stability.
on commercial computer programs, such as ABAQUS, ANSYS,
and other similar programs, which can even model the structural
geometry realistically. Such programs, in addition to providing the Computer Software Specification for the DAM
Line element, also provide a Solid element for three-dimensional DAM requires a program capable of second-order elastic analyses.
(3D) modeling. A more detailed modeling geometry influences In general, commercial structural analysis programs have this facil-
the accuracy of the numerical analysis results. Therefore, it is rea- ity. Nevertheless, there is no guarantee that the outcome will be the
sonable to question whether the simple modeling using DAM has same from all of the programs. The accuracy programmed in the
represented the real conditions of the structure being evaluated. software will have some influence on the design output. To avoid
Thus, the concern of numerical analysis using DAM is to obtain a any doubt, the computer program used should be tested by launch-
structural model configuration, which can represent the real ing the analysis benchmark problems prepared by AISC (2010).
structure. When DAM is not representative, it means that the The SAP2000 computer program (CSI 2011) has been tested by
analysis result is inaccurate or not correlated with the real Dewobroto (2016) and has met the intended criteria.
condition.
A simple strategy for getting an accurate numerical modeling to
Consideration of Initial Imperfections
represent the real structure is to calibrate it against the experimental
test results, i.e., the test is required for accurate results from the nu- Initial imperfection results in destabilizing effects. DAM deals with
merical analysis. it by explicitly modeling the geometry, which contains the initial

© ASCE 04018028-2 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

 imperfection, or giving a notional load or equivalent lateral load of cross section element. Furthermore, the outcome will be even more
the working load percentage. different if the second-order effect on the structure under review
Direct modeling is performed by placing additional nodal points becomes more and more significant. The difference from the old
on the Line element. The node coordinates can be taken from the method is caused by not explicitly accounting for initial out-of-
material and fabrication tolerance. The pattern is chosen to give the plumbness and member-stiffness reduction due to the spread of
greatest destabilizing effect; either the occurring deflection pattern plasticity in the analysis.
due to loading or buckling pattern can be chosen. ELM is calibrated to give a resultant axial strength, Pu, consist-
Notional load constitutes the lateral load at nodal points based ent with the actual response, but it underestimates the actual internal
on the percentage of vertical load working at that nodal point moments under the factored loads. The “actual response” curve has
level. The notional load is added together with other lateral loads, larger moments due to the combined effects of distributed yielding
for all combinations, except for certain cases explained in Section and geometric imperfections, which are not included in the second-
2.2b(4) of AISC (2010). The amount of notional load is Ni = order elastic analysis.
0.002Yi. The factor 0.002 % 1/500 represents the nominal out-of- A major advantage of DAM is that it captures the internal forces in
plumbness value for the story and is in line with the AISC Code of the structure more accurately, particularly for the cases in which there
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Standard Practice. When the structure has a greater value, a rear- are high gravity loads and low lateral loads. This advantage comes at
rangement is then needed. the expense of applying notional lateral loads to the structure and
reducing the frame stiffness as part of the input for the analysis.
Adjustments to Stiffness
Partial yielding due to residual stress in the steel cross section (hot DAM and Experimental Loading Test
rolled or welded) produces stiffness reductions when given a load
close to the limit state conditions, and it gives a destabilizing effect, DAM is considered simpler than FEM; hence, it is chosen as an al-
such as geometric imperfection. DAM deals with partial yielding ternative numerical method to predict the ultimate strengths of a
by reducing the structural stiffness. The value is found by calibrat- steel scaffolding system. The advantage of DAM is that it can do
ing it against the experimental test result, namely EI* = 0.8t bEI and the stability analysis of a structure by calculating the P-D effects.
EA* = 0.8EA. This stiffness reduction factor applies only to the cal- This effect has a significant influence on sway portal buildings
culation of the ultimate strength condition and the stability of steel (sway frames). For truss structure (nonsway frames), the influence
structure; it is not used to calculate the drift, deflection, vibration, or of P-D is relatively insignificant; hence, it can be ignored. Such an
vibration periods. assumption happens because the design usually uses the old
method (ELM), which has not taken imperfection into considera-
Notional Load to Represent the Inelastic Condition tion. Meanwhile, DAM has taken imperfection into account as a
standard procedure because the performance of steel scaffolding
Giving a notional load can also be used to represent the reduction of
is significantly influenced by imperfection and other matters
flexural rigidity, t b, resulting from the inelastic condition due to
related to its stability. For this reason, DAM is deemed appropri-
cross-sectional residual stress. This strategy is suitable to simplify the
ate. Consequently, steel scaffolding should be modeled as a por-
calculation of DAM in any member with a large compression force of
tal structure, and the amount of each of its connection system
aPr > 0.5Py, where the value of t b < 1.0. If this strategy is used, then
rigidities needs to be predicted carefully.
t b = 1.0, and an additional notional load is given by Ni = 0.001 Yi.
The structural modeling process, which considers the influence
The load is given simultaneously with the previous notional
of rigidity between secondary elements, such as bracing and the
load, which represents the previous initial geometric imperfection,
connection system, produces varied outcomes. Therefore, the nu-
and because it has the nature of enlarging, the final notional load
merical simulation with DAM can produce something accountable,
would be Ni = 0.003Yi and t b = 1.0 for all loading combinations.
but calibration needs to be performed first. In this way, the modeling
strategy closest to the real condition can be chosen. The experimen-
Available Strengths of Structures tal loading test of steel scaffoldings serves as the calibrator. Hence,
By using DAM for structural stability analysis to calculate the nom- the structural modeling in the numerical simulation gives an out-
inal cross-sectional strengths of the structure, a regular procedure come that correlates with the experimental test result, which can
can be used, such as the one used in ELM, i.e., in the AISC code then be used to evaluate similar steel scaffolding parametrically.
(AISC 2005, 2010, 2016), applying Chapters E to I for the nominal
strength of the cross section and Chapters J to K for the nominal
strength of the connection, but by taking K = 1 in the member slen- Experimental Loading Test: One-Story Scaffoldings
derness KL/r.
General Information
Accuracy of DAM (New) against ELM (Old) An experimental loading test is a simple and fast method to find the
maximum strengths of a structure, yet it is a costly and risky choice;
DAM and ELM use the same column strength equations of Section therefore, it is only used for the smallest configuration, i.e., one-
E3 of the AISC code (AISC 2005). Therefore, if DAM is used to an- story steel scaffoldings. Despite its limited amount, its existence is
alyze the structural system of a truss, which only receives an axial extremely important for calibrating numerical modeling analysis
force, the result would not be different or is equally accurate. For using DAM. It is understandable given that each (1) amount of rigid-
slender columns, the calculation of the effective length KL is critical ity of structural elements and (2) notional load location (representing
to achieve an accurate solution when using ELM. imperfection) will influence stability. As a result, the final outcome
The different design outcomes begin when the loaded steel struc- can vary, making it hard to determine which model can be deemed
ture creates an axial force and a simultaneous flexural moment to its as the most suitable (accurate) with real structural conditions.

© ASCE 04018028-3 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

 Steel Scaffolding Loading Test structural strengths (rigidity) is deemed to have taken place. The
final condition of structural collapse can be seen in Fig. 3.
An experimental loading test is conducted at the Laboratory
One reason steel scaffolding collapses is because of the buckling
of Structure and Construction Research Institute for Human
deformations in the columns. The horizontal and diagonal pipe
Settlements, Ministry of Public Works, Cileunyi, Bandung,
Indonesia. The loading test with tubular columns with a 58-mm
diameter yielded a capacity of Pmax = 47.29 t. Because the setup
consists of four columns, the ultimate capacity of each column is
116 kN (11.82 t) corresponding to the maximum load recorded
during the test. The load on the structure in the test is recorded
with the load cell on the scaffolding column support, and the re-
spective horizontal displacement is observed with a transducer
set in between the ends of the column in which the displacement
is not prevented by bracing. This is based on the assumption that
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

its collapse takes the form of column flexural buckling.

The configuration of the steel scaffolding loading test and testing
atmosphere can be seen in Fig. 2.

Structural Collapse Behavior

In addition to information on the maximum load, which causes the
steel scaffolding to collapse, the behavior is observed when the
scaffolding collapses. The loading is produced using a hydraulic
jack (Fig. 2), which can automatically stop when a significant dis- Fig. 3. Damage of tested scaffolding.
placement change occurs. This is a condition in which a loss of the

Fig. 2. Frame and loading arrangements.

© ASCE 04018028-4 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

 elements, which serve as bracing, remain intact and show no dam-
age. The connections between elements are still set well, with no
sign of damage. Hence, the damage is of the local type, appearing
only in the columns.
The part of the columns connected to the bracing element tends
to remain intact. This shows that its connection system is rigid
enough to “hold” the columns, even though its rigidity is not the
same when compared with the continuous connection system.
Although the weakest parts in the scaffolding are the pipe columns
loaded vertically, the structural modeling for the analyses using
DAM becomes easier, especially when using the notional loads for
simulating the imperfection effects. Thus, it is best to have the
notional loads in the parts of columns that are free from lateral
restraint.
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Calibrated Numerical Simulation: One-Story Scaffoldings

General Information
The numerical analysis of structural collapse can generally be per-
formed using a FEM-based nonlinear inelastic computer program,
such as ABAQUS, ANSYS, ADINA, and so forth. Using these pro-
grams, the structural geometry can then be modeled realistically
using the 3D solid element. Nevertheless, the implementation pro- Fig. 4. Node and element numbering for a one-story model.
cess is relatively laborious for practical daily design efforts. As an
alternative, the elastic second-order analysis with SAP2000 (CSI
2011) and DAM are used. Table 1. Specification components of the structural model
To obtain an accurate analytical result, the computer output of
numerical analysis using DAM will be compared with the experi- Element number Component Dimension Note
mental test result. At this stage, the numerical modeling made can 1–4 and 13–16 Column f 58 mm t = 3.25 mm Steel pipe
be rearranged in such a way that the results can be considered as 5, 17 Top column f 58 mm t = 3.25 mm Reduced
representing the real conditions; this is called the calibration stage. stiffness
To do so, the facility of the SAP2000 program, namely the Display- 8–10 Bracing f 48 mm t = 3.25 mm Steel pipe
Show Plot Function (F12), can be used to record the output of inter- 6, 7, 11, 12 Connection 4  100 mm Steel plate
nal forces and deformation (P-D) for each stage of the loads given.
Moreover, the data can be shown in the form of a curve, making it
easier to do proper comparisons. From the P-D curve, the load just
by modifying the flexural stiffness in the upper part of the col-
before the collapse ensues can be detected due to large deformation
at the observation point. The internal forces then become the umn. Whether or not this modeling assumption is appropriate
required strengths (Ru). The next stage is the evaluation of the avail- depends on the calibration results against the experimental test
able cross-sectional strengths ( f Rn). The maximum load is obtained data.
when the criterion Ru % f Rn, is met. To find the determining imperfection conditions, the notional
load (Ni) location is varied, as shown in Fig. 5. The variations are
denoted as Case 1 through Case 4.
Modeling the One-Story Scaffoldings
The modeling of the steel scaffolding connection (Fig. 1) is consid- Stability Analysis Using the DAM
ered as semirigid. For convenience, it is deemed to be a plate whose
size is found by trial and error to suit the experimental test results. To calculate the influence of the residual stress, the elastic modulus
From a two-dimensional (2D) perspective, a structural modeling is is reduced to 80%. A notional load of Ni = 0.003·Yi is placed
a plane frame. The perpendicular direction is considered as confined according to locations shown in Fig. 5. The notional load varia-
from horizontal deformations. The direction is chosen because the tions will be evaluated based on the experimental test result data.
geometry of the scaffolding is symmetrical; thus, only one side Afterward, a second-order elastic analysis using the SAP2000
needs to be reviewed. program is made, and using the Display-Show Plot Function the
In reference to the details of the experimental test plan and amount of incremental load and displacement at Nodal Point 7
results, a structural model as a plane frame is made with a notation can be recorded. The results are presented in a P-D curve, as seen
of the nodal point and member element numbering as shown in in Fig. 6.
Fig. 4, in which the data for the structural elements are tabulated as The notional load variations influence the analytical results. The
shown in Table 1. experimental loading test indicates that scaffolding columns experi-
The top end of the column is connected to the hydraulic jack ence some buckling, and the numerical simulation with notional
as a point of loading; thus, it is modeled as being fixed in the rota- loads in Case 3 shows no buckling. Therefore, the valid model
tion but free in the translation of any direction. However, the includes Cases 1, 2, and 4 only, showing how the buckling occurs
flexural stiffness of the column at its upper part is somewhat after the load exceeds 100 kN. The buckling in these cases is consid-
questionable; hence, it is considered as a special element member ered to take place when a small incremental load causes a large

© ASCE 04018028-5 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

 Fig. 5. Different notional load (Ni) positions in the one-story model: (a) Case 1; (b) Case 2; (c) Case 3; and (d) Case 4.
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. Load-displacement curve due to Ni placement for the one-story model.

lateral deformation, changing the P-D curve to be practically This is where the importance of the experimental test result data
horizontal. lies. The calibration is performed by comparing the P-D curve of
After the buckling, the P-D curve line tends to be horizontal; the experimental loading test results versus the P-D curve of the nu-
thus, it is concluded that the elastic second-order analysis computer merical simulation. Thereafter, a trial-and-error procedure is con-
program still works well. The load can be increased further, yet the ducted until a model structure is found in which the results are
result seems irrational or odd; hence, when it happens, the incre- deemed to represent the experimental ones accurately.
mental load is ceased manually. The additional loading after buck- The trial-and-error procedure shows that the P-D curve of the
ling is deemed as meaningless. numerical simulation results is influenced by the flexural stiffness
When learning the behavior of buckling resulting from Cases 1, of the end part of the column ( f 58-mm pipe) of Elements 5 and 17
2, and 4, Cases 2 and 4 are similar; thus, only Cases 1 and 2 will be (Fig. 4). To adjust this flexural stiffness, the second moment of area
compared. The behavior of buckling caused by Case 1 is preceded of f 58-mm pipe is multiplied by factors 1.0, 1.25, and 0.75. The
by a large deformation. Meanwhile, Case 2 has a collapse from a results are, respectively, notated as Cases 2, 2A, and 2B in Fig. 7.
sudden failure, which is similar to the collapse of the column during Correspondingly, TR-3 through TR-13 are the lateral displacements
the experimental loading test. Therefore, the notional load configu- measured in the columns of the scaffolding.
ration, which is deemed as right, is Case 2, and it will be used Three numerical models were made based on Case 2. Out of
further. these three, the Case 2A model (end column of 125% of the intact
cross section) gives a P-D curve that is closest to the experimental
one and implies that the Case 2A model can be taken as the cali-
Comparisons of Simulation Results and brated one. The model will be developed using parametric princi-
Real Conditions ples for the next ultimate load capacity analysis of two-story steel
scaffoldings.
For the behavior of collapse resulting from Case 2 numerical simu- It is proven that when a good calibration is performed, the struc-
lation (Fig. 6) to be correlated with the real conditions (experimen- tural analysis to find the maximum load for the scaffolding can also
tal results), the modeling configuration should then be calibrated. be performed using DAM (AISC 2010). Most important is that the

© ASCE 04018028-6 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Load-displacement curves, experimental versus numerical.

Table 2. Results of the stability analysis of one-story scaffoldings (Case 2A)

Step UX 7 (mm) RZ 1 (kN) P 14 (kN) M33 14 (kN·m) Note

37 1.10080 110.74471 −111.07195 −0.18151 Check 3
38 1.94281 113.71173 −114.06932 −0.26949 Check 2
39 3.52047 116.65575 −117.06472 −0.43269 Check 1
40 12.93401 119.35227 −120.03893 −1.39921 Buckle
Note: Several steps are deliberately deleted.

Fig. 8. Deformation changes at each step approaching instability (Case 2A).

numerical modeling correlates with the real conditions using cali- Nevertheless, no review has been made in relation to the material
bration in which the P-D curve from the model is adjusted against strengths (material boundary conditions). Hence, the available
that of the experimental test for the same structure. strengths ( f Rn) will be evaluated using the provisions of the LRFD
(AISC 2010). The smallest loading quantity, which fulfills both
conditions, is considered as the maximum load of structure.
Evaluation of Column Strengths Furthermore, some loading stages that cause buckling of Case
2A output in Element 14 (as a column; Fig. 4), which are considered
The P-D curve shows the behavior of the structure, which has as maximum, will be shown as seen in Table 2.
accommodated the stability element (geometric boundary condi- To understand the collapse behavior, it would be helpful to
tion); thus, it can be used to predict the load, which causes buckling. see the deformation for each of these steps (Fig. 8). The stable

© ASCE 04018028-7 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

 Table 3. Column interaction check for a one-story model
Pu 8 Mu
R¼ þ
Check Step Pu (kN) Mu (kN·m) Pu = f Pn f Pn 9 f Mn Note
1 39 −117.06472 −0.43269 1.004 Not check Not OK
2 38 −114.06932 −0.26949 0.979 1.053 Not OK
3 37 −111.07195 −0.18151 0.953 1.002 OK
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 10. Different notional load (Ni) positions in the two-story model:
(a) Case 5; (b) Case 6; and (c) Case 7.

If the loading condition at Step 39 is deemed as the final condi-

tion before instability (buckle), then the amount of the load based
on the stability or geometry boundary condition is Pu = 117 kN
(Table 2). The material boundary condition will be evaluated as fol-
lows. The smallest load is the determining one. The material bound-
ary condition is evaluated based on the nominal strengths of the
structural cross section of the ELEMENT #14 (see Figure 4) as
follows.
Check the available strength of Element 14 using pipe f 58 mm
(L = 1.5 m) (Fig. 4), Fy = 371 MPa (yield strength from mill certifi-
cate), E = 200,000 MPa (modulus elasticity of steel), D = 58 mm
(outside diameter of pipe), t = 3.25 mm (wall thickness), Ag =
559 mm2 (cross-sectional area), rmin = 19.4 mm (radius gyration),
and Zx = 9753.5 mm3 (plastic-section modulus).
To calculate the compressive strength, first check the local buck-
ling criteria according to Table B4.1a (AISC 2010). The section is a
nonslender element because D/t = 17.85 less than 0.11 E/Fy = 59.3.
The nominal strength, Pn, will be determined based on the limit
state of flexural buckling following the provisions of E3 (AISC
2010)
 qffiffiffiffiffiffiffiffiffiffiffi 
Fig. 9. Node and element numbering for a two-story model. ðKL=rmin ¼ 77:3Þ  4:71 E=Fy ¼ 109:4 then Fe

¼ ðp 2 EÞ=ðKL=rmin Þ2 ¼ 330 MPa

change in the geometric form of the scaffolding is seen in Steps
Fy
37–39 and unstable from Step 40 forward. Under such condi- Fcr ¼ 0:658Fe
Fy ¼ 231:7 MPa therefore Pn ¼ Fcr Ag
tions, the structural instability or buckle is assumed to occur.
The structure experiencing the buckle cannot be used; thus, the ¼ 129:5 kN
loading condition in Step 40 does not need to be examined fur-
ther because it cannot be used. Therefore, it is necessary to To calculate the flexural strength, check the local buckling crite-
review the maximum load before buckling, i.e., Steps 37–39. ria according to Table B41b (AISC 2010). The steel pipe is a com-
The check is started from Step 39 and if it fails to fulfill the load pact section because D/t = 17.85 less than 0.07 E/Fy = 38. The nomi-
bearing condition, then Step 38 with a smaller load is examined. nal flexural strength, Mn, will be determined based on the limit state
If the condition Pu % f Pn is met, then the ultimate limit load is of yielding according to F8–1 (AISC 2010), therefore, Mn =
obtained. Fy·Zx = 3.618 kN·m.

© ASCE 04018028-8 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 11. Load-displacement curve due to Ni placement in the two-story model.

forward the flexural moments. Further, following the calibration

Table 4. Results of the stability analysis of two-story scaffoldings result, the flexural stiffness at the upper ends of the column modules
(Case 7)
should be increased to be 125% of full cross sections, i.e., Elements
Step UX 7 (mm) RZ 1 (kN) P 2 (kN) M33 2 (kN·m) Note 5, 10, 29, and 34 (Fig. 9). The model structure and its nodal and ele-
ment numbering are now as shown in Fig. 9.
30 3.960 90.18136 −90.14139 −0.458 Check 3
The assumption of notional load placement, to represent imper-
31 7.465 91.99879 −91.96898 −0.815 Check-2
fection, is given according to the deflection prediction when experi-
32 15.416 93.47612 −93.46949 −1.623 Check-1
encing buckling or according to mode shape. There are three
33 28.467 95.04253 −95.07358 −2.956 Buckle
notional load placement configurations classified as load Cases 5–7.
Note: Several steps are deliberately deleted. From these three, the smallest incremental load that causes the
instability condition is selected. The configuration is shown in
The interaction of flexure and compression in doubly symmetric Fig. 10.
members (pipe) will be limited by Eqs. H1–1 (AISC 2010) as seen
in Table 3. Stability Analysis Using the DAM
The numerical analysis result indicates that the capacity of the
Similarly with the strategy presented previously with the numerical
steel scaffolding is determined by the nominal strength of cross sec-
analysis based on the form of the loaded structure, the analytical
tion that fulfills the criterion Pu % f Pn. Thus, the ultimate load
results will be treated as the P-D diagram of a representative point
capacity of the column of one-story steel scaffoldings is Pu = 111
in the column in which buckling is predicted to occur. For this rea-
kN (from Step 37).
son, Nodal Point 7 is chosen for lateral displacement and Nodal
Note that the maximum load of the experimental load test is 116
Point 1 for reaction force.
kN (11.82 t).
As shown in Fig. 11, Case 7 gives a shape for the curve that best
matches the sudden buckling behavior as well as the smallest load
Numerical Simulations of the Two-Story Scaffoldings without the preceding initial deformation.

General Information Evaluation of Column Strengths

Based on the calibrated model (Case 2A), a parametric concept for Representative loading stages of Case 7 that precede buckling are
constructing a two-story scaffolding model is developed and ana- presented in Table 4 for Element 2 (column) (Fig. 9), which is con-
lyzed using DAM as explained before. sidered as the most deformable.
As mentioned earlier, to understand the collapse behavior, it
would be helpful to learn the deformation for each of the loading
Modeling the Two-Story Scaffoldings steps before the collapse (Fig. 12). The stable changes in the geome-
Two-story steel scaffoldings are made by stacking up two of the try form of the scaffolding include Steps 30–32, and the stability
one-story steel scaffolding frame modules discussed previously. loss appears in Step 33. Under such conditions, the structural insta-
For its structural modeling, the column connection between the two bility or buckle is assumed to occur. The structure experiencing
modules is considered as continuous. This condition is taken with buckling cannot be used; thus, the loading condition in Step 33 does
the assumption that although it takes the form of a connection not need to be examined further because it is obvious that it cannot
because it has been designed to be easy to set and dislodge, the con- be used. Therefore, it is necessary to review the maximum load
dition is considered precise enough and strong enough to distribute before buckling, i.e., Steps 30–32. The check begins in Step 32, and

© ASCE 04018028-9 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 12. Deformation changes at each step approaching instability (Case 7).

Table 5. Column interaction check for the two-story model

Pu 8 Mu
R¼ þ
Check Step Pu (kN) Mu (kN·m) Pu = f Pn f Pn 9 f Mn Note
1 32 −93.46949 −1.623 0.802 1.245 Not Ok
2 31 −91.96898 −0.815 0.789 1.011 Not Ok
3 30 −90.14139 −0.458 0.773 0.898 OK

Fig. 13. Ultimate load response of one- and two-story steel scaffolding: (a) Case 2A; and (b) Case 7.

if it fails to fulfill the limit load condition then the lower loads in from the mean value of both, i.e., Pu = (91.97 þ 90.14)/2 % 91 kN.
Step 31 and smaller are checked. The ultimate load is obtained Compared with the one-story scaffolding, making a combination of
when the limit condition Pu % f Pn is fulfilled. stacked modules of two single-story scaffoldings will result in a
capacity that is 82% of the previous configuration.
Check the Available Strength of Element 2 (Column Left)
Because the configuration of f 58 mm (L = 1.5 m) steel pipe Discussions
remains the same as earlier, the previously calculated data, i.e., Pn =
129.5 kN and Mn = 3.618 kN·m, are used. An ultimate load analysis using DAM has been made using two
The interaction of flexure and compression in doubly symmetric configurations of steel scaffolding, each consisting of the same
members (pipe) will be limited by Eqs. H1–1 (AISC 2010) as seen frame modules. One-story steel scaffolding (Figs. 4 and 5) is
in Table 5. a single module. Two-story steel scaffolding consists of two mod-
Judging from the results of Checks 2 and 3, it can be estimated ules, which are arranged stacked up (Figs. 9 and 10). Despite the
that the maximum load occurs between the Step 30 and Step 31 same constituent frame modules, the resulting carrying capacities
incremental loading stages. Briefly, the maximum load is taken are different. The carrying capacity of one- and two-story steel

© ASCE 04018028-10 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 14. Ultimate load response of standard and modified scaffolding: (a) Case 7; (b) Case 8; and (c) Case 9.

Fig. 15. Load-displacement curves of standard and modified scaffolding.

scaffolding is 111 kN (100%) and 91 kN (82%), respectively, and lateral restraint. To see the extent of the sway effect, a numeri-
which means that increasing the number of stories will decrease cal simulation using the previous model as modified is performed
the load capacity of the steel scaffolding. by adding new bracing in parts that had no previous bracing (Case
DAM evaluates elements as beam-column; their capacities are 8) and by using an initial condition with additional lateral support at
determined by the interactions of flexure and compression accord- the point of loading (Case 9). The ultimate load is taken at Pu = 91
ing to Eqs. H1–1 (AISC 2010). Hence, the compression capacity is kN to enable its comparisons with the previous analysis (Fig. 14).
correlated with the occurring moment. Therefore, when the load To see the effect of structural geometry changes (addition of
capacity is reduced it is due to the existence of a greater moment, new bracing in the parts of the frame with no previous bracing and
and vice versa. To prove this, DAM results will be shown regarding adding lateral support to the point of load at the column end) a rean-
the moments occurring for the axial load when it reaches the ulti- alysis is performed on the two-story scaffolding, which yields the
mate condition. P-D curves shown in Fig. 15.
The addition of stories results in the greater influence of P-D. Case 7 is a configuration of a steel scaffolding model, which is
This can be seen from the occurring moments when the load deemed as a standard, and Cases 8 and 9 are the modifications of
capacity of the steel scaffolding decreases (Fig. 13). The condition that standard model. Case 8 is made by adding new diagonal brac-
occurs because the structural system experiences sway deforma- ing to a part of the frame that had no previous bracing. The resulting
tions that are significantly influenced by the bracing configuration impact is highly significant; the bending moment, which results

© ASCE 04018028-11 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028

 from the second-order effect, decreases and seems even to disap- thank Mr. Sutadji Yuwasdiki, the chief of the experimental
pear. As a result, as can be seen in Fig. 15, the occurring P-D curve loading test executing team at the Research Institute for Human
takes the form of a vertical straight line. This means that during the Settlements, Ministry of Public Works, Bandung. Without them,
loading stages no buckling or instability occurs. This is possible this research on numerical analysis of steel scaffolding systems
because the behavior of the structural system in the frame shifts would have never been completed.
from the sway mode to the nonsway mode. This is the best solution
for a multistory steel scaffolding system. A similar influence
appears for Case 9, which gives some lateral restraint to the point of References
loading, although it does have some influence in increasing its load-
ing capacity, but the impact is not as great as when using Case 8. Adam, J. M. 2013. “Special issue on analysis of structural failures using nu-
merical modeling.” J. Perform. Constr. Facil. 27 (1): 2–3. https://doi
.org/10.1061/(ASCE)CF.1943-5509.0000420.
Conclusion AISC. 2005. Specification for structural steel buildings. ANSI/AISC 360-
05. Chicago: AISC.
AISC. 2010. Specification for structural steel buildings. ANSI/AISC 360-
Downloaded from ascelibrary.org by Wiryanto Dewobroto on 09/15/23. Copyright ASCE. For personal use only; all rights reserved.

The local product of the ring-lock scaffolding system is made

10. Chicago: AISC.
based on similar products abroad, and it has been proven reliable AISC. 2016. Specification for structural steel buildings. ANSI/AISC 360-
as an aid for construction execution at a more economical price. 16. Chicago: AISC.
To ensure its strengths when used on-site, an ultimate load Andresen, J. 2012. “Investigation of a collapsed scaffold structure.” Proc.
capacity analysis has been conducted to find the maximum load Inst. Civ. Eng. Forensic Eng. 165 (2): 95–104. https://doi.org/10.1680
that can be supported by the structure. Generally speaking, for /feng.11.00021.
such an analysis, a relatively sophisticated FEM-based computer CSI (Computers & Structure, Inc.). 2011. SAP2000 linear and nonlinear
program and cumbersome procedure will be used. However, this static and dynamic analysis and design of three-dimensional structure.
research has successfully completed the analysis by using a sim- Berkeley, CA: CSI.
pler method, i.e., DAM, which is the latest structural steel design Dewobroto, W. 2016. Steel structure–Behavior, analysis & design–
method (AISC 2005, 2010, 2016). AISC 2010. 2nd ed. Tangerang, Indonesia: Jurusan Teknik Sipil
UPH.
Despite its relatively simple structural modeling, the accuracy of
East, J., and F. R. Rutz, 2016. “Stability of trusses: Direct analysis method
its results can be improved by calibrating it against the experimental compared to experimental results.” In Proc., Geotechnical and
loading test result data, which are relatively costly. Therefore, it is Structural Engineering Congress 2016, edited by C. Y. Chandran and
used only for relatively small structural modules, which in this case M. I. Hoit, 201–211. Red Hook, NY: Curran Associates.
is one-story steel scaffolding. Furthermore, using a parametric Hadipriono, F. C., and H. K. Wang. 1987. “Causes of falsework collapses
study technique, the numerical analysis is used for analyzing the during construction.” Struct. Saf. 4 (3): 179–195. https://doi.org/10
ultimate load of more complex structural modules, i.e., two-story .1016/0167-4730(87)90012-9.
steel scaffoldings. Kim, S. E., K. W. Kang, and D. H. Lee. 2003. “Full-scale testing of space
This research found that the steel scaffolding modules arranged in steel frame subjected to proportional loads.” Eng. Struct. 25 (1): 69–79.
a stack up are highly susceptible to stability issues. As a result, this https://doi.org/10.1016/S0141-0296(02)00119-0.
module’s load capacity is less than the one-story scaffolding module. Toma, S., and W. F. Chen. 1992. “European calibration frames for second-
order inelastic analysis.” Eng. Struct. 14 (1): 7–14. https://doi.org/10
In addition, adding new bracing can effectively improve its resistance
.1016/0141-0296(92)90003-9.
against stability issues, hence, improving the load capacity. Toma, S., W. F. Chen, and D. W. White. 1995. “A selection of calibra-
tion frames in North America for second-order inelastic analysis.”
Eng. Struct. 17 (2): 104–112. https://doi.org/10.1016/0141-0296
Acknowledgments (95)92641-K.
Valerii, V. 2011. “Comparison of scaffolding systems in Finland and in
The author would like to extend his gratitude to Mr. Susiato and Russia.” Bachelor’s thesis, Saimaa Univ. of Applied Sciences.
Mr. Yosep Tan, the owner and head of PT. Putracipta Weesner, L. B., and H. L. Jones. 2001. “Experimental and analytical
Jayasentosa, manufacturers of steel scaffoldings and funding of capacity of frame scaffolding.” Eng. Struct. 23 (6): 592–599. https://doi
the experimental loading tests. The author would also like to .org/10.1016/S0141-0296(00)00087-0.

© ASCE 04018028-12 Pract. Period. Struct. Des. Constr.

Pract. Period. Struct. Des. Constr., 2018, 23(4): 04018028