Analysis Reinforced Concrete (RC) Beams Nonlinear - FEA
Analysis Reinforced Concrete (RC) Beams Nonlinear - FEA
Analysis Reinforced Concrete (RC) Beams Nonlinear - FEA
Prepared By:
David R. Dearth, P.E.
"Analysis of Reinforced Concrete (RC) Beams using Nonlinear Finite Element Techniques"
The purpose of this article is to develop an understanding of the arithmetic involved in the design solution to
analyzing reinforced concrete (RC) beams. This article provides a single source review to outline the steps
necessary to perform analysis of RC beams beginning at the linear elastic region into the nonlinear analyses
and to ultimate failure.
In this analysis no prior knowledge of the cracked RC beam is assumed. The analytical results are compared
to actual test data from a RC beam tested under closely monitored laboratory conditions. A finite element
model is also constructed to produce a simplified nonlinear analysis of the RC beam.
Introduction
During the process of designing reinforced concrete beams structural engineers typically estimate the general
sizing of the beam using conventional hand equations. These conventional hand analysis approaches involve
using linear elastic equations to compute equivalent, or transformed, cross sectional properties. The linear
elastic approaches have been utilized for many years and have for the most part been very successful. These
elastic equations are limited to estimating the onset of RC beam cracking of the concrete and to some extent
also approximating ultimate failure of the RC beam after initial cracking.
When it is desired to calculate the regions between initial cracking and ultimate failure, nonlinear analysis
techniques are required. When nonlinear approaches are desired, finite element analysis (FEA) techniques are
employed. Before considering taking on the task of calculating the nonlinear response of RC beams, engineers
should have at least a working knowledge of how to perform a conventional linear analysis using pencil, paper
and a calculator. When tasked with performing the nonlinear analysis one most likely will look at a sample
tutorial problem and simply follow the same steps with their particular problem of interest substituting
instructions from the sample tutorial. In essence this simply becomes a case of parroting the steps outlined in
the sample tutorial problem without fully understanding what is going on.
So how might one develop confidence in performing these types of nonlinear analysis problems? The best
way is to locate a sample real life problem with a known, documented solution that one can be work through
using hand calculations and also develop a FEA model.
So the question is.... How can one relate the physical observations witnessed in the environmental test lab
to virtual testing calculated using nonlinear FEA techniques? Or how one can simulate actual physical
testing of RC beams using computer analyses software?
The best approach would be to locate some actual test data. When it comes to verifying the analytical results
from analysis of RC beams there is very little documented information showing results from actual physical
testing under tightly controlled laboratory conditions.
Sample Problem: RC Beam from Buckhouse Testing (1997) Marquette University 1,2
A search through the available engineering literature found comprehensive, documented data of actual
physical testing under tightly controlled laboratory conditions of several RC beams performed by Foley and
Buckhouse1. Wolanski2 provides analytical correlation to the laboratory testing with detailed finite element
analysis of the 1997 Buckhouse1 RC beam tests. The investigations performed by Foley and Buckhouse1 are
cited in several other technical papers addressing FEA of RC beams.
Figure 1a shows a sketch of the1997 Buckhouse RC beam geometry, loading and boundary conditions.
Figure 1b shows the layout for the internal reinforcement, rebar & stirrups.
Figure 1a
RC Beam Tested at Marquette University
Figure 1b
RC Beam Reinforcement Layout
Linear Elastic – Uncracked Approach: The linear elastic uncracked method assumes tension stress in
the concrete remains below the cracking limit. Tension stresses are assumed linear elastic and fully
effective in an uncracked concrete section. This method is used to calculate the state of stress and
deflections when the RC beam structure is subjected to normal anticipated service load conditions.
Elastic – Cracked Approach: The elastic cracked method assumes concrete tension stress has exceeded
cracking limits and neglects any concrete tension stress. Linear elastic compressive stresses are balanced
by tension stresses in the reinforcement.
Ultimate – Cracked Approach: The ultimate cracked method assumes a simplified yielding stress
criterion. For ultimate load carrying strength capability, tension stress in the concrete is assumed
nonexistent and maximum compressive strain is assumed to equal ε c = 0.003. The balancing tensile
loading is assumed fully carried by the steel reinforcement with the steel at yield.
Stage 1: Linear Elastic Conditions – Normal Service Life & Initial Crack Stress
During normal service life conditions no cracking of the concrete is assumed. Stresses and deflections of RC
beams can be performed using a conventional linear elastic approach. For normal service life conditions it is
assumed that stress in the RC remain in the linear elastic range and the only difficulty is computing the sections
properties, EcI, for the RC beam. In computing the section properties, Young’s modulus, E c, is taken as the
28-day strength value. The inertia, I, for the composite section of concrete and steel reinforcement is computed
using the conventional “transformed section method”.
Analysis for the uncracked RC beam can be performed by treating the RC beam as a composite assembly of
concrete and steel reinforcement. A conservative approach is to neglect the stiffness contribution from the
reinforcement and consider the gross section properties of the concrete only. When the steel reinforcement is
included, it is assumed the reinforcement steel maintains intimate contact with the surrounding concrete. With
intimate contact or bonding maintained, the steel reinforcement and concrete will maintain the same strain
compatibility during loading. So long as the maximum tensile stress in the concrete remains below the
maximum the tensile capacity stress of the concrete, fr, (fr is also referred to as the modulus of rupture) then
the RC beam will act as a conventional composite assembly.
To maintain strain compatibility between the steel reinforcement and surrounding concrete, the steel
reinforcement having a greater modulus of elasticity, Esteel > Econcrete, experiences a greater magnitude in stress
distribution across this composite section than does the concrete. The ratio in stresses between the steel and
concrete is called the modular ratio, n. Where n = Esteel / Econcrete. The transformed section properties for the
composite assembly of concrete and steel rebar is computed by replacing the steel area by an equivalent
concrete area.
Figure 2 shows a cross section of the concrete and rebar for the 1977 Buckhouse RC beam tested. For the
cross section shown in Figure 2, the transformed cross sectional properties for the composite assembly is I tr
= 5,138 in4. Appendix “A” lists detailed arithmetic to compute the properties listed in Figure 2.
Figure 2
RC Beam Linear Elastic Composite Section Properties
As a general rule of thumb concrete tension stress is approximately 1/10 the compressive values. For the
allowable compressive stress for concrete used, Wolanski2 listed a value of f’c = 4,800 psi. ACI 318 9.5.2.33
computes tension stress or modulus of rupture stress, fr, computed based on the maximum compressive stress
using the following:
𝑓 = 7.5 𝑓 = 7.5 4,800 = 520 𝑝𝑠𝑖
The onset of initial cracking of the concrete is computed per ACI 9.5.2.3 using the gross section properties.
The threshold cracking stress is computed from ACI 318 equations for the rupture stress, fr. For the gross
section the calculated cracking moment, Mcr, and corresponding equivalent loading Pcr = 4,680 lbs(1.). The
equivalent linear elastic deflections for this applied loading = 0.050”.
For Figure 2 transformed section properties the cracking moment, M cr_tr, and corresponding equivalent
loading Pcr_tr = 5,080 lbs. The equivalent linear elastic deflections for this applied loading = 0.052”.
(1.) The calculated value of 4,680 lb is within 4% of the average physical observed loading to crack initiation equal to 4,500 lb.
quoted by Foley and Buckhouse1.
Stage 2: Elastic Cracked Section - Balanced State of Stress; Concrete & Rebar
Cracks begin to form when the tensile stress in the concrete exceed the maximum capacity of the concrete to
react tension stress at the modulus of rupture, fr. When the maximum tensile stress in the concrete exceeds
modulus of rupture, fr, the cross section is assumed to be "cracked" and all the tensile stress is assumed to be
carried by the steel reinforcement.
When a flexural crack occurs, it begins at the tension face when the tensile capacity is exceeded and the crack
propagates upward until the concrete is in compression. The section properties change as the crack propagates,
causing the increased tensile stress forcing the crack upwards toward the compression region. Equilibrium is
achieved once the crack stops propagating. Compressive stresses are still assumed to remain in the elastic
region.
For the cross section shown in Figure 3, the cross sectional properties for the composite assembly is Icracked =
1,116 in4. Appendix A lists detailed arithmetic to compute the properties listed in Figure 3. The value of Icrack
it used to compute the instantaneous effective inertia, Ieff per ACI 318 9.5.2.3. This effective inertia is used
for computing deflections after crack initiation.
Figure 3
RC Beam Elastic Cracked Section Properties
To compute ultimate failure, cracked bending moment, a simplified yielding stress criterion is assumed. For
ultimate load carrying strength capability, tension stress in the concrete is assumed nonexistent and maximum
compressive strain is assumed to equal εc = 0.003. This magnitude of compressive strain is representative of
concrete with compressive strength from 2,000 < f‘c < 6,000 psi. The balancing tensile loading is assumed
carried by the steel reinforcement with the steel material at yield stress.
Figure 4 shows the cracked Whitney equivalent stress block cross section. The calculated ultimate moment
capacity ϕMu = 826,740 in-lbs. Using this calculated moment value, ϕMu, an equivalent ultimate loading Pu,
= 13,780 lbs(2.) can be calculated. Appendix A lists detailed arithmetic to compute the ultimate moment ϕM u
values for Figure 4. The equivalent deflections at this applied ultimate loading applied loading = 0.548”.
Figure 4
RC Beam Ultimate Cracked Section
(2.) In the Foley and Buckhouse1 testing a theoretical ultimate load capacity value equal to 14,600 lb. is quoted. No documentation
could be found on how this 14,600 lb. value was computed; 13,780 lb. is within 6% of 14,600 lb. Foley and Buckhouse 1
noted the following: “The moment due to the dead weight of the beam was subtracted from Mult to give the moment capacity
of the beam due to superimposed live load”
The effective inertia, Ieff, is calculated after crack initiation according to ACI 318 9.5.2.3 (3.). Figure 5 shows
a comparison of measured deflections at the center line of the control beam C11,2 to the computed deflections
using ACI 3183 hand equations.
𝑀 𝑀
𝐼 =( ) ∗𝐼 + 1− ∗𝐼
𝑀 𝑀
Figure 5
Comparison of ACI 318 Hand Calculations to 1977 Buckhouse Laboratory Test Data
(3.) When crack initiation calculations are performed using the transformed section properties, Itr, ACI 318 9.5.2.3 instead of
gross section properties, the effective inertia results in slightly lower deflections.
Applied Analysis & Technology, Inc. Page 7 of 20
Analysis of RC Beams using Nonlinear Finite Element Techniques
For comparison purposes it was decided to duplicate as closely as possibly the RC beam test article and FEA
model definition described by Foley and Buckhouse 1 and Wolanski2. FEA programs can be utilized to address
the nonlinear characteristics confronted when analyzing RC beams. Figures 6a & 6b show isometric views
of the RC beam geometry per Figures 1a & 1b. Due to the symmetry of loading and geometry, the full RC
beam can be idealized using quarter symmetric idealization as shown in Figure 6b; symmetric boundary
conditions (constraints) are denoted.
Figure 6a
Full RC Beam Geometry
Figure 6b
Quarter Symmetric RC Beam Geometry
The concrete is idealized using 3D solid elements. Young’s modulus of elasticity is computed using ACI
318 8.5.1. Modulus of elasticity of the concrete is …
The stress-strain curve data for the concrete is shown in Figure 7. The Wolanski2 analysis used a Poisson’s
ratio for concrete ν= 0.3. It is recognized that a Poisson’s value of ν= 0.18 to 0.2 may be more representative
for concrete and therefore Poisson’s ratio ν= 0.2 is used herein. [Poisson’s ratio at 0.3 or 0.2 resulted in no
discernable difference in the deflection results.]
Figure 7
Concrete Compressive Stress-Strain Data from Wolanski 2
The non-linear concrete cracking formulation used is called “Buyukozturk” model. The typical strain-
softening relationship of concrete shown in Figure 8.1 is idealized as shown in Figure 8.2. In Figure 8.2
the area under the tension-softening region represents fracture energy G f. When tension-softening, Es, is not
included, material loses all load-carrying capacity; stress goes to zero upon cracking. Assuming the
characteristic length for the RC concrete beam equals the depth of the beam, hc = 18 inches. Then fracture
energy Gf can be calculated from the following:
𝐸 𝑓
𝐺 = ℎ 1− = 0.62 𝑙𝑏/𝑖𝑛
𝐸 2𝐸
Steel Reinforcement
The steel reinforcement (rebar & stirrups) is idealized using Rod/Truss elements with the node points defined
each rebar element sharing common nodes with the concrete solids. This approach is called discrete
idealization of rebar with the concrete. The steel material is defined using the nonlinear stress-strain curve
data listed below. The nonlinear material properties are entered using von Mises yield criteria.
Figure 9 shows the layout for rebar and stirrups (5.) per Figure 1b.
Figure 9
Quarter Symmetric RC Beam Rebar & Stirrups
(5.) In the analysis performed by Wolanski2 an additional stirrup is shown at the mid-span location of the beam. The nonlinear
FEA model developed herein was processed with and without the additional mid-span stirrup. The net results were a
negligible difference in the solutions.
Foley and Buckhouse1 lists the ultimate recorded loading at failure equal to 16,300 lbs. To ensure uniform
deflections at the load points, individual concentrated loading is distributed as shown in Figure 10.
Figure 10
Concentrated Nodal Loading Distribution
Figure 11 shows a comparison of measured deflections at the center line of the control beam C1 of Foley and
Buckhouse1 to the computed deflections from the FEA model.
Figure 11
Comparison of Results to 1977 Buckhouse Laboratory Test Data
Notes:
(6.) The nonlinear FEA solution contains only 17 output steps using adaptive load stepping
(7.) In the analysis performed by Wolanski2 the iteration parameters were adjusted during selected load steps to ensure the
analytical results better fit the experimental data. Having prior knowledge of the solution to the nonlinear response is not
what is generally available to analysts attempting to predict the response of beams before they are built. For the analysis
outlined herein, no prior knowledge of the solution is assumed and it was decided to perform the analysis by applying the
full ultimate loading and letting the program solution determine what happens in between zero load and full ultimate loading.
Figure 12 shows maximum principal stress contour plot of the concrete at the onset of crack initiation. As
indicated on the contour plot legend, the load step increment to the onset of cracking is “Incr =11, Time
=0.32150” of total loading. The applied loading corresponding to Figure 12 is 16,300 x 0.32150 = 5,240
lbs. This value is within +3% of the hand calculations using the composite properties for the transformed
section (concrete & rebar) Pcr_tr = 5,080 lbs. The corresponding computed stress value of 490 psi is within
6% of the maximum allowable tension stress, or rupture stress fr = 520 psi defined per ACI 318 9.5.2.3.
Figure 12 illustrates the concrete stress distribution at the last linear-elastic load step before cracks begin to
develop.
Figure 12
Maximum Principal Stress Contour Plot
Last Load Step Prior to Crack Propagation
The progressive pictures shown in Figure 13 illustrate typical propagation of the concrete cracks by
displaying Vector plots of Resultant Crack Strain.
Figure 13
Crack Propagation – Resultant Crack Strain
1.) This report demonstrates the use of linear and nonlinear finite element analysis techniques to analyze
RC beams. It is concluded that FEM codes can be utilized with a high degree of confidence in
performing these types of analysis investigations.
2.) In this report it was desired to simulate the element mesh density and procedures outlined in the
Wolanski2 analysis for purposes of direct comparison. It is recommended extended analysis be
performed using fine grid mesh density to improve the solution accuracy.
3.) In this report, the effects of creep, shrinkage, moisture intrusion, thermal cycling, fatigue, etc…
on concrete have not been included. It is recommended to extend analysis to include these
effects.
References
1.) Christopher M. Foley and Evan R. Buckhouse, “Strengthening Existing Reinforced Concrete Beams for
Flexure Using Bolted External Structural Steel Channels”, Structural Engineering Report MUST-98-1,
January 1998.
2.) Anthony J. Wolanski, B.S., “Flexural Behavior of Reinforced and Prestressed Concrete Beams Using
Finite Element Analysis”, Master’s Thesis, Marquette University, Milwaukee, Wisconsin May, 2004.
3.) ACI 318-08, “Building Code Requirements for Structural Concrete and Commentary” ACI Manual of
Concrete Practice, Part 3, American Concrete Institute, Detroit, MI, 1992.
LIST OF SYMBOLS
Pcr_gross = loading to crack initiation gross cross sectional properties concrete only
Pcr_tr = loading to crack initiation using transformed cross sectional properties
Pu = maximum loading to ultimate using Whitney stress block cross sectional properties
Appendix A
𝑓 = 4,800 𝑝𝑠𝑖
Using ACI 318 8.5.1 modulus of elasticity of the concrete is computed as:
The tensile capacity stress of the concrete, fr, is defined using ACI 318 9.5.2.3. This value fr is also referred
to as the modulus of rupture.
𝑓 = 7.5 𝑓 = 520 𝑝𝑠𝑖 Concrete Cracking Stress for normal weight concrete
𝑏 = 10 𝑖𝑛𝑐ℎ𝑒𝑠, Base
ℎ = 18 𝑖𝑛𝑐ℎ𝑒𝑠, Height
𝑏 ∗ℎ
𝐼 _ = = 4,860 𝑖𝑛𝑐ℎ𝑒𝑠
12
Per ACI 9.5.2.3 the crack initiation moment based on concrete gross section properties
(𝑓 ∗ 𝐼 _ )
𝑀 _ = = 280,582 𝑖𝑛 − 𝑙𝑏𝑠
ℎ
( )
2
_
𝑃 _ = = 4,677 𝑙𝑏𝑠 Loading to crack initiation
𝐸
𝑛= = 7.343
𝐸
Where "n" is modular ratio of Esteel/Econcete. Transform area of Steel to equivalent or effective area of
concrete, As_eff
The location of the centroid of area for the effective composite section, concrete & steel rebar is:
(𝐴 ∗𝑦 _ + 𝐴_ ∗ 𝑑 )
𝑦 = = 9.238 𝑖𝑛𝑐ℎ𝑒𝑠
(𝐴 +𝐴 _ )
The transformed composite area moment of inertia is computed using parallel axis theorem
𝐼 = 𝐼 _ + 𝐴 𝑦 _ −𝑦 + 𝐴 _ (𝑑 −𝑦 ) = 5,138 𝑖𝑛𝑐ℎ𝑒𝑠
Compute the equivalent loading, P lbs, to just exceed the maximum allowable concrete tension stress to
initiate first cracking.
Using bending equation σcr_tr = (Mcr_tr *yc)/Itr, where Mcr_tr = Pcr_tr*60 in-lbs
(𝑓 ∗ 𝐼 )
𝑀 _ = = 304,686 𝑖𝑛 − 𝑙𝑏𝑠
(ℎ − 𝑦 )
_
𝑃 _ = = 5,078 𝑙𝑏𝑠 Loading to crack initiation using transformed section properties
The corresponding stress in the steel rebar at this loading is σ rebar = n(Mcr_tr*yrbar)/Itr
𝑀 _ ∗ (𝑑 −𝑦 )
𝜎_ =𝑛∗ = 2,727 𝑝𝑠𝑖
𝐼
When the maximum tensile stress in the concrete exceeds modulus of rupture, fr, the cross section is
assumed to be "cracked" and all the tensile stress is assumed to be carried by the steel reinforcement. The
compressive stress in the remaining concrete is assumed to remain elastic. Calculate the location of the
neutral axis for the cracked section from the top of the beam, "c crack".
𝑏
−𝐴 _ + 𝐴 _ +4 𝐴 _ (𝑑 )
2
𝑐 = = 3.969 𝑖𝑛𝑐ℎ𝑒𝑠
𝑏
2
2
The moment of inertia of this transformed area w.r.t. the neutral axis for "cracked" section is calculated
using the following for single reinforcement RC section:
[𝑏 (𝑐 ) ]
𝐼 = + 𝐴 _ (𝑑 −𝑐 ) = 1,116 𝑖𝑛𝑐ℎ𝑒𝑠
3
For ultimate load carrying strength capability tension stress in the concrete is assumed nonexistent and
maximum compressive strain is assumed to equal ε c = 0.003. The magnitude of compressive strain is
representative of concrete with compressive strength from 2,000 < f'c < 6,000 psi. The balancing tensile
loading is assumed fully carried by the steel reinforcement with the steel material at yielding at f s_ty.
Calculate the location of the neutral axis for the cracked section from the top of the beam, "c crack".
Moment Reduction factor ϕu set equal to 1.0 to compute Ultimate moment, 𝜙 = 1.0
𝑎
𝑐 = = 1.688 𝑖𝑛𝑐ℎ𝑒𝑠
𝛽
Calculate Deflections from Elastic Moment of Inertia Calculations for Cracked Section
At the computed Ultimate Moment capacity, effective inertia is calculated using ACI 318 9.5.2.3. To be
conservative, the gross section properties, Ic_gross, and concrete modulus, Ec, are used.
_ _
𝐼 = ( ) ∗𝐼 _ + 1− ∗𝐼 = 1,263 𝑖𝑛𝑐ℎ𝑒𝑠
(4 ∗ 𝑎 − 3 ∗ 𝐿 )
𝑦 = (𝑃 ∗ 𝑎 ) ∗ = −0.572 𝑖𝑛𝑐ℎ𝑒𝑠
24 ∗ 𝐸 ∗ 𝐼
Note: When the "cracked" section properties are used in place of "effective" section properties deflections at
ultimate loading equal -0.647".
(4 ∗ 𝑎 − 3 ∗ 𝐿 )
𝑦 _ = 𝑃 _ ∗𝑎 ∗ = −0.050 𝑖𝑛𝑐ℎ𝑒𝑠
24 ∗ 𝐸 ∗ 𝐼 _
(4 ∗ 𝑎 − 3 ∗ 𝐿 )
𝑦 _ = 𝑃 _ ∗𝑎 ∗ = −0.052 𝑖𝑛𝑐ℎ𝑒𝑠
24 ∗ 𝐸 ∗ 𝐼