See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/343935156

Design of Reinforced Concrete Beams with Web Openings

Conference Paper · August 2020

CITATION READS
1 919

2 authors:

Gerry Sweeney Suhaib Salawdeh

1 PUBLICATION   1 CITATION   
Galway-Mayo Institute of Technology
45 PUBLICATIONS   347 CITATIONS   
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Seismic design of steel braced frames View project

Design, model, and simulation of steel hollow sections using the physical testing and novel extended finite element models for application in earthquake engineering View
project

All content following this page was uploaded by Suhaib Salawdeh on 28 August 2020.

The user has requested enhancement of the downloaded file.

 &LYLO (QJLQHHULQJ 5HVHDUFK LQ ,UHODQG 

Design of Reinforced Concrete Beams with Web Openings

Gerry Sweeney1, Suhaib Salawdeh2
1
Chartered Engineer, Main St, Loughrea, Co Galway, Ireland
2
Department of Civil Engineering, Galway-Mayo Institute of Technology, Dublin Road, Galway, H91 T8NW, Ireland
email: sweeneygerry@hotmail.com, suhaib.salawdeh@gmit.ie

ABSTRACT: Web openings in Reinforced Concrete (RC) beam construction are common place. An extensive review of the
current research on RC beams with web openings was conducted. It was identified that a numerical model could be developed for
RC beams with openings and validated against existing experimental research data. The Finite Element (FE) application, Ansys,
was used to develop five different FE models. The numerical results compared well with experimental and analytical data. Overall
model results are provided and discussed, with key findings and recommendations for future work presented.

KEY WORDS: Web Openings; Reinforced Concrete; Finite Element; Ansys; Numerical Model.

1 INTRODUCTION web opening sizes under point load. Their work was presented
in the high impact journal of Composite Structures and is well
General Background cited within this area of research. Five beams were tested
Web openings within Reinforced Concrete (RC) beams are where FRP sheets were applied to the opening areas; four were
often required to be created so as to pass through mechanical un-strengthened and the final test beam was a solid RC beam
and electrical building services. For example, electrical cables with no openings which was used as a control specimen. The
are often passed through the downstands (or webs) of RC beam dimensions were the same in all specimen cases having a
beams. Better floor to ceiling heights may be achieved by cross-section of 250mm (H) x 100mm (W) and of length
passing services through beams as opposed to bringing them 2000mm (L). A schematic of the test rig is given in Figure 1
under beams; passing services through beams may also reduce below.
the overall height of a multi-storey building which may often
be a critical consideration when obtaining planning permission.
Structural engineers need to be able to safely design for the size
and locations of these openings within the web of RC beams
without adversely affecting the structural integrity of the
building – the structural design question that often arises is how
is the ultimate load capacity of an RC beam affected by varying
web opening size and location?
Previous Research
Some notable experimental research work into the effect of web
openings in RC beams exists in the research literature: dating
back to 1963, Bresler and Scordelis [1] tested 12 RC beam,
some with openings; Somes and Corley [2] experimentally
tested 19 specimens – 12 with openings, and, Mansur et al. [3]
experimentally tested eight RC beams with large openings. Figure 1. Four-point bending test rig from Abdalla et al. [6]
Other experimental RC beam researchers include Allam [4] and (source: Abdalla et al. [6]).
Aykac et al. [5]. While their findings are wide ranging, the
experimental research indicates that deep openings relative to Two equal monotonic loads were symmetrically applied to a
beam height (>35%) may significantly affect the load capacity beam specimen from a central hydraulic load cell. Central
of an RC beam. deflections were measured and plotted against the associated
applied loads. Figure 2 below shows the output deflection and
In more recent times, the research relating to web openings has ultimate strength results from the un-strengthened beam
focused on the use of Fibre Reinforced Polymer (FRP) sheets specimens having varying web opening sizes. As is evident
to improve the shear and axial load resistance around web from Figure 2 below, the results show that the ultimate strength
openings. A specific piece of research relating to the use of of each beam diminishes as the opening size increases.
FRP sheets in strengthening web openings prominently appears
in the research literature – Abdalla et al. [6] have presented
experimental work on ten different RC beams with different


&LYLO (QJLQHHULQJ 5HVHDUFK LQ ,UHODQG 
Figure 2. Mid-span deflections and ultimate strength results
for the solid and un-strengthened RC beams with web
openings from Abdalla et al. [6] (source: Abdalla et al. [6]).
Two specific research papers have incorporated the work from
Abdalla et al. [6] and have used Finite Element (FE) analysis to
endeavour to replicate part of the findings in relation to the un-
strengthened beam types. Amiri and Masoudnia [7] used the
results from Abdalla et al [6] in relation to the solid beam as a
verification study for its FE analysis. Similarly, Hafiz et al. [8]
almost replicated the work conducted by Amiri and Masoudnia
[7] and again used results from Abdalla et al. [6] for the solid
beam as a verification study for its FE analysis in Ansys. It is
notable in both numerical studies (i.e. Amiri and Masoudnia [7]
and Hafiz et al. [8]), the researchers did not endeavour to fully
validate the experimental results from Abdalla et al [6]; Amiri
and Masoudnia [7] endeavoured to validate the beam specimen
with the 300x100mm opening; Hafiz et al. [8] did not
endeavour to validate any of the experimental results presented
by Abdalla et al. [6] for RC beams with web openings. The
primary focus in this study was to extend work done by Amiri
and Masoudnia [7] and Hafiz et al. [8] to endeavour to validate
the experimental results obtained in Abdalla et al. [6].
Motivation and Justification of Research
First and foremost, the motivation for this work is that there
appears to be clear merit and practical value in developing a
validated FE model relating to the analysis of web openings in
RC beams. Web openings in RC beams is a common
occurrence in RC construction and having a practical
understanding of the structural implications of the effects of
openings within RC beams is of clear benefit.
Secondly, while there is a lot of research conducted in the
area, full validated FE models against real experimental data on
web openings in RC beams do not appear to be that common.
Having closely reviewed the research literature in the area of
RC beams with web openings, the research in the area is mainly
one of the following (a) purely experimental - presenting
findings, (b) experimental with analytical analysis or (c) FE
numerical analysis using part of an experimental study as a
verification study. There appears to be very few studies that
use full experimental data as validation of an FE numerical
model. This research used the full RC beam un-strengthened
experimental data from Abdalla et al. [6] to endeavour to
validate an FE model developed as part of this work.

Aim of Research
Primary Aim:
• To build a FE 3-D model using a FE analysis tool, Ansys,
for an RC beam with web openings.
• To validate the model against experimental & analytical
results.
Specific Aims:
• To build five different FE 3-D models using Ansys –
from an initial basic solid beam model to a final fully
reinforced beam with web openings.
• To model the non-linear behaviour of concrete and
reinforcement within an RC beam.
• To validate Models 1-3 against Analytical Results and
Models 4 and 5 against Experimental Data.
2 METHOD
Method Overview
Figure 3 below shows a schematic of the four point bending
problem that was ultimately modelled using the FE analysis
tool, Ansys, showing (a) the simply supported beam with
horizontal and vertical reinforcement (b) two monotonic loads
positioned symmetrically about the mid-span of the beam and
(c) a web opening located between the left hand beam support
and point load. The geometry in Figure 3 was specifically
designed to replicate the experimental test rig, shown in Figure
1 above, used by Abdalla et al. [6] so that the numerical results
obtained from Ansys could be directly validated against
experimental test results obtained from Abdalla et al. [6].
Figure 3. Geometry of the four-point bending test modelled
in Ansys.
Five different numerical models were developed within Ansys
– the different model mesh geometries, outputted from Ansys,
are shown in Figure 4 below:
x Model 1 – Elastic without Reinforcement under UDL and
point load.
x Model 2 – Inelastic without Reinforcement under point
load.
x Model 3 – Inelastic with Reinforcement under point load.
o Singly reinforced.
o Double singly reinforced.
x Model 4 – Inelastic with Reinforcement with links under
point load.
x Model 5 – Inelastic with Reinforcement with links with
web openings under point load.
o Web openings modelled: 100x100mm; 200x100mm;
300x100mm and 300x150mm.
 &LYLO (QJLQHHULQJ 5HVHDUFK LQ ,UHODQG 

Models 1 and 2 – No Reinforcement, No

Openings

(a) Volume geometries for Model 5.

Model 3a – Singly Reinforced, No Openings
2no 10mm bars (lower)

Model 3b – Double-Singly Reinforced, No (b) Applied Boundary Conditions.

Openings
4no 10mm bars (lower)
Model 4 – Fully Reinforced, No Openings
4no 10mm bars (lower)
2no 10mm bars (upper)
8mm stirrups at 150mm
spacing
Figure 5. Image (a) shows for Model 5a, nine separate
geometry volumes created and then glued together to form
one homogenous beam and image (b) shows the applied
boundary conditions of the simply supported conditions
denoted in light blue left and right and hatched in red are
the locations of the applied pressures.
The finite element types used were Solid65 to model the
concrete and Link180 to model the steel reinforcement. Stress
strain characteristic curves for concrete (i.e. Figure 6) and steel
(i.e. Figure 7) were inputted into Ansys.

Model 5 – Fully reinforced with Openings

4no 10mm bars (lower)
2no 10mm bars (upper)
8mm stirrups at 150mm
spacing

Openings
5a – 300x150mm Opening
5b – 300x100mm Opening
5c – 200x100mm Opening
5d – 100x100mm Opening

Figure 4. Geometries of the five different models developed

in Ansys. Figure 6. Ansys output showing the linear/non-linear
stress-strain curve for structural concrete for Model 5a
Building the Ansys Models (݂௖௞ ൌ Ͷʹ‫)ܽܲܯ‬. SIG is stress parameter in SI units of Pa;
EPS is the strain and is dimensionless.
Ansys Mechanical APDL 19.2 was used for this study. For
example, Figure 5 (a) below shows the volume geometries for
Model 5a and Figure 5 (b) shows the mesh detail and applied
boundary conditions. Mesh convergence analysis showed that
an element length of 50mm was sufficient for all models.


&LYLO (QJLQHHULQJ 5HVHDUFK LQ ,UHODQG 

Figure 7. Ansys output showing the stress-strain curve for
the longitudinal reinforcement steel for Model 5a. SIG is
stress parameter in SI units of Pa; EPS is the strain and is
dimensionless.
Figure 8. Screenshot of the concrete material definitions
specifying different important concrete material attributes,
e.g. assigning stress threshold level for cracking.
Figure 8 above shows the important threshold inputs used for
concrete cracking and crushing (inputs in Pa). Poisson ratio for
Figure 9. Solution Controls screens in Ansys. The load
time-step was set for 100 increments and for each
increment, results were logged and recorded.
3 RESULTS
Results Overview
Models 1 – 3 were compared against analytical results (please
refer to Figure 10 and Figure 11 below). The analytical
computations were obtained using pure bending theory for
Model 1 (elastic behaviour with no reinforcement) and for
Model 2 (inelastic behaviour with no reinforcement). For
Model 3, for Service Limit State (SLS), triangular stress block
theory is used and for Ultimate Limit State (ULS), the Whitney
stress theory is used.
Model 4 and 5 were compared against experimental results
obtained from Abdalla et al. [6] – please refer to Figure 12 to
Figure 14.
Models 1 – 3: Analytical Results’ Comparison
As is evident from Figure 10 below, numerical mid-point
deflection results for Model 1 (i.e. elastic behaviour with no
reinforcement) compared very well with pure bending
analytical results.
10000
Ansys Analytical
8000
Force (N)

6000
concrete and steel was given as 0.2 and 0.3 respectively. Yield
strength of the longitudinal steel was given as 400MPa for all 4000
models; and the yield strength for the stirrups was given as
240MPa. The concrete strength of each beam model matched 2000
Abdalla et al. [6] – for all beam models except Models 5a and
5d, ݂௖௞ ൌ ͷͲ‫ ;ܽܲܯ‬for Models 5a and 5d, ݂௖௞ =Ͷʹ‫ ܽܲܯ‬and 0
Ͷ͵‫ ܽܲܯ‬respectively. 0 0.1 0.2 0.3 0.4 0.5 0.6
Mid Deflection (mm)
The solution convergence criteria and increment load step were
set. Figure 9 below shows a fixed 100 increment time step was Figure 10. Model 1 – Ansys mid-span deflection results vs
set up. The solution iterated through the chosen time-step analytical results.
period and results were outputted for every time-step to the
results database. For Model 2 (i.e. inelastic behaviour with no reinforcement),
Ansys mid-span deflection results equated to Model 1 results
Results were obtained and validated against analytical and
above. At ultimate load capacity, when the beam cracked, the
experimental results, as detailed in the next section.


 &LYLO (QJLQHHULQJ 5HVHDUFK LQ ,UHODQG 

section failed; sudden failure occurred due to the absence of

reinforcement. Ansys Experimental
Figure 11 below shows results from Model 3a and showed very 25.00
good correlation with analytical results. Ansys graphical
outputs are also given, showing the stress distributions within
20.00
the beam and reinforcement.
25000

Load (kN)
15.00
Ansys Analytical
20000
10.00
15000
Force (N)

10000 5.00

5000 0.00
0.00 2.00 4.00 6.00 8.00 10.00
0
0.00 10.00 20.00 30.00 40.00
Midspan Deflection of RC Beam (300x150 Opening)
Midspan Stress of Concrete (MPa)
(mm)
Model 3a – Model mid-span horizontal stress results vs analytical results.
Figure 13. Mid-span Deflection (mm) of the RC Beam for
Model 5a (i.e. with 300x150mm Opening).

Ansys Experimental
90
80
Ultimate Load Capacity (kN)

70
Model 3a - Horizontal stress distribution in beam (Units in Pa).
60
50
40
30

Model 3a - Horizontal stress distribution in reinforcement (Units in Pa). 20

Figure 11. Results for Model 3a shown above along with 10

Ansys outputs of the beam and reinforcement horizontal
0
stress distributions.
Model 4 Model 5a Model 5b Model 5c Model 5d
Models 4 – 5: Experimental Results’ Comparison Solid 300x150mm 300x100mm 200x100mm 100x100mm

Ansys Experimental Figure 14. Ansys ultimate load capacity (kN) of RC beams
90.00 with web openings when compared with experimental data.
80.00 Other Ansys Outputs
70.00
Ansys permits the output and plotting of results after every
60.00
timestep. This is very useful as bending stresses and strains
Load (kN)

50.00 may be tracked and analysed within the beam and

40.00 reinforcement, as applied loads are increased. Cracking of the
30.00 RC beam may also be analysed throughout the bending process.
20.00 Figure 15 below provides examples of crack distributions that
10.00 were outputted from Ansys. For example, as is evident from
Figure 15, cracking starts in the areas of high stress (e.g. around
0.00
the edges of the opening); as loading increases, cracking
0.00 5.00 10.00 15.00
propagates throughout the beam with the most significant
Midspan Deflection of Solid RC Beam (mm) cracking occurring in and around the opening, as expected.
Crack distributions may be analysed to determine whether a
Figure 12. Model 4 – Model mid-span deflection results for predominantly flexure or shear failure has occurred.
the solid RC beam compared well with experimental results
from Abdalla et al [6].


 &LYLO (QJLQHHULQJ 5HVHDUFK LQ ,UHODQG 
Crack distribution at
total load of 5700N
Crack distribution at
total load of 8250N
Crack distribution at
ultimate yield (total
load = 22,800N)
Figure 15. Outputs from Ansys showing how cracking
propagates through the RC beam with increased applied
loads.
Results Discussion
In the main, results from this research compared well with
analytical and experimental data. Figures 10 – 14 gives a
graphical summary of the results obtained for the models
analysed. Figure 14 is an important bar chart comparing the
study’s numerical ultimate load capacity results against
experimental ultimate load data from Abdalla et al. [6]; three of
the five RC beams modelled here failed at ultimate loads
closely correlating to experimental ultimate loads from Abdalla
et al. [6]. Summary results from the study are as follows:
x Ansys Solid RC Beam Models vs Analytical Results for
Concrete and Steel - Results substantially within 11.8%.
x Ansys Solid RC Beam Model vs Experimental Results -
Ultimate Load Capacity within 4.3%.
x Ansys Models for 100x100mm and 200x100mm Openings
– numerical Ultimate Load Capacity is greater by 55% and
56% respectively when compared against Abdalla et al.
[6]. Aykac et al. [5] have shown that load capacity reduces
by 35% where square openings are 40% of total height
(e.g. Model 5d) – this finding would more closely match
the numerical Ansys results obtained for Model 5d,
differing by 20%.
x Ansys Models for 300x150mm and 300x100mm Openings
vs Experimental Results - Ultimate Load Capacity within
9.8%.
x Increasing web opening to 60% of overall height of beam
- Ultimate Load Capacity reduces by 75% - correlates
closely with experimental results [6].
x The Ansys models developed for this study closely
concurred with strength-of-materials beam bending
theory, specifically in relation to solid RC beams.
4 CONCLUSIONS
The key findings from this research are given as follows:
x The Ansys solid RC beam models correlated very well
with analytical results.

View publication stats
x The Ansys solid RC beam models showed good correlation
with experimental results.
x Ansys model results for large web openings (300x150 and
300x100) closely matched experimental findings. Ansys
model results for smaller web openings (200x100 and
100x100) differ by 55% and 56% respectively with
experimental findings - please see Results section for
discussion on this finding.
x Increasing the web opening height to 60% of overall beam
height reduces the ultimate load capacity of the beam by
75% when compared with the equivalent solid beam - this
finding was closely consistent with experimental results
[6].
For the most part, this study has shown that Ansys is capable
of accurately modelling RC beams with web opening under
flexural loading. Non-linear material properties for concrete
and steel may be incorporated in Ansys. Crack propagation
under increased flexure loads may be closely analysed.
Openings in RC beams are an everyday occurrence and
therefore, this research has a real practical and useful
engineering application. Future work may explore the
application of FRP sheets to strengthen web openings in RC
beams; a lot of experimental work has more recently been
undertaken in this area. Ansys is to be capable of modelling
the application of thin materials, like FRP sheets, to a geometry
[9]. It would be an interesting extension of this research to
model the use of FRP sheets on RC beams with web openings.
ACKNOWLEDGMENTS
The first author would like to sincerely thank his co-author, Dr.
Suhaib Salawdeh, for his expert help and guidance. The
authors would also like to extend their thanks to the Department
of Civil Engineering at the National University of Ireland,
Galway (NUIG) where this research was carried out. The
second author would like to acknowledge the support of
Science Foundation Ireland through the SFI Industry
Fellowship (Award ID 18/IF/6318).
REFERENCES
[1] Bresler, B., & Scordelis, A. C. (1963). Shear strength of reinforced
concrete beams. In Journal Proceedings, 60(1), 51-74.
[2] Somes, N.F., & Corley, W.G. (1974). Circular openings in webs of
continuous beams. Shear in Reinforced Concrete. American Concrete
Institute, Special Publication SP-44, 359-398.
[3] Mansur, M. A, Lee, Y. F, Tan, K. H, & Lee, S. L. (1991). Tests on RC
Continuous Beams with Openings. Journal of Structural Engineering,
117(6), 1593-1606.
[4] Allam, S. M. (2005). Strengthening of RC beams with large openings in
the shear zone. Alexandria Engineering Journal, 44(1), 59-78.
[5] Aykac, B., Kalkan, I., Aykac, S., & Egriboz, Y. E. (2013). Flexural
behavior of RC beams with regular square or circular web openings.
Engineering Structures, 56, 2165-2174.
[6] Abdalla, H.A ; Torkey, A.M ; Haggag, H.A ; & Abu-Amira, A.F.
(2003). Design against cracking at openings in reinforced concrete
beams strengthened with composite sheets. Composite Structures, 60(2),
197-204.
[7] Amiri, S., & Masoudnia, R. (2011). Investigation of the opening effects
on the behaviour of concrete beams without additional reinforcement in
opening region using Fem method. Australian Journal of Basic and
Applied Sciences, 5(5), 617-627.
[8] Hafiz, B., Ahmed, S., Barua, S., & Chowdhury, S. (2014). Effects of
Opening on the Behavior of Reinforced Concrete Beam, IOSR Journal
of Mechanical and Civil Engineering, 11(2), 52-61.
[9] Hassan, N. Z., Sherif, A. G., & Zamarawy, A. H. (2017). Finite element
analysis of reinforced concrete beams with opening strengthened using
FRP. Ain Shams Engineering Journal, 8(4), 531-537.