See discussions, stats, and author profiles for this publication at: [Link]

net/publication/336150832

Plastic Hinge Behavior and Rotation Capacity in Reinforced Ductile Concrete

Flexural Members

Article in Engineering Structures · October 2019

DOI: 10.1016/[Link].2019.109699

CITATIONS READS
43 2,920

2 authors:

Mandeep Pokhrel Matthew J. Bandelt

New Jersey Institute of Technology New Jersey Institute of Technology
8 PUBLICATIONS 77 CITATIONS 61 PUBLICATIONS 783 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Matthew J. Bandelt on 01 October 2019.

The user has requested enhancement of the downloaded file.

ACCEPTED MANUSCRIPT
Plastic Hinge Behavior and Rotation Capacity in Reinforced Ductile
Concrete Flexural Members
Mandeep Pokhrel and Matthew J. Bandelt

Engineering Structures
Received Date: 24 May 2019
Received in Revised Form: 10 September 2019
Accepted Date: 16 September 2019
Published Online: 1 October 2019

Please cite this article as:

Pokhrel, M. and M.J. Bandelt (2019). "Plastic Hinge Behavior and Rotation Capacity in Reinforced
Ductile Concrete Flexural Members." Engineering Structures, 200: 109699. doi:
10.1016/[Link].109699.

This is a PDF file of a final draft manuscript post-refereeing that has been accepted for publication.
The manuscript has completed the peer review process, but this is not the copyedited version of this
publication. This publication is provided to promote timely dissemination of scholarly and technical
work. Copyright and all rights therein are retained by the authors or copyright holders.

Please visit the following website to download the copyedited and published copy of this publication
from the publisher:
[Link]
 Plastic Hinge Behavior and Rotation Capacity in Reinforced Ductile
Concrete Flexural Members

Mandeep Pokhrela , Matthew J. Bandelta,

a John A. Reif, Jr., Department of Civil and Environmental Engineering, New Jersey Institute of Technology,
Newark, NJ, USA

Abstract

Ductile concrete materials, such as high-performance fiber-reinforced cementitious composites (HPFR-

CCs), are being considered as a new material that can be used in plastic hinge regions of buildings
and bridges to reduce damage as compared to ordinary reinforced concrete. In flexure, a highly
nonlinear plastic strain distribution of the primary tensile reinforcement is observed at dominant
crack locations, leading to significant differences in plastic hinge length and behavior as compared
to ordinary reinforced concrete flexural members. In this study, two-dimensional finite element
models were validated with experimental results using a recently developed bond-slip constitutive
model, which aids in simulating multiple damage states, such as yield and collapse level drift states.
After validation, the modeling approach was used to study the plastic hinge region behavior of re-
inforced HPFRCC flexural members with variations in mechanical properties, boundary conditions,
and geometric properties. The particular areas of interest were reinforcement yielding location,
plastic strain distribution, and curvature localization region. New expressions were proposed to
predict the equivalent plastic hinge length, Lp , using variables such as shear-span, tensile strength
of HPFRCC, reinforcement ratio, yield strength of the reinforcement, boundary condition, and load-
ing scenario. Proposed equations for plastic hinge length of reinforced HPFRCCs are incorporated
into a mechanics-based approach to predict rotational capacity of flexural members. The analytical
approach was also shown to estimate yield and nominal flexural strength with reasonable accuracy.
The proposed methodology is compared with experimental results and expressions for plastic hinge
length found throughout the literature for reinforced concrete and HPFRCC members. The results
of the study will help practicing engineers and researchers simulate the performance of reinforced
HPFRCC flexural members in a computationally efficient manner.
Keywords: Plastic Hinge Length, HPFRCC, Finite Element Analysis, Plastic Rotation, Fracture

∗ Correspondingauthor
Email addresses: mp595@[Link] (Mandeep Pokhrel), bandelt@[Link] (Matthew J. Bandelt)

Preprint submitted to Engineering Structures September 13, 2019

1 1. Introduction and Background

2 In recent years, there has been a gradual shift in seismic design philosophy from traditional force-
3 based seismic design (FBSD) to alternative seismic design principles based on deformation capacity
4 known as performance-based seismic design (PBSD) [1]. In PBSD, a structure is designed to achieve
5 a target performance level under a specified intensity of seismic loading. The performance level
6 of a structure is explicitly defined in terms of metrics such as material strain or inelastic compo-
7 nent rotation that corresponds to physical damage such as yielding, crushing, or fracture of tensile
8 reinforcement [2, 3]. PBSD also requires inelastic modeling of structural components to measure
9 these performance metrics which cannot be completed using elastic modeling techniques typically
10 applied in FBSD approach. Therefore, it is important to accurately quantify inelastic structural
11 component response in order to calculate displacements and rotation demands for performance as-
12 sessment of structures subjected to different intensities of seismic loading. In reinforced concrete
13 structural elements, numerous experimental and numerical studies have been used to understand
14 inelastic behavior in terms of plastic hinge region response, which has resulted in empirical models
15 to predict plastic hinge length as described in Table 1.
16 As PBSD approaches have become more commonplace in analysis and design, engineers have
17 sought out new materials and structural systems to take advantage of material and structural duc-
18 tility [4, 5]. In the area of reinforced concrete, recent developments have been made to improve
19 the mechanical response of concrete materials and have led to the emergence of high-performance
20 fiber-reinforced cementitious composites (HPFRCCs). HPFRCCs are a class of cement-based duc-
21 tile materials exhibiting a pseudo tensile strain-hardening behavior with distributed cracking under
22 uniaxial tension[6, 7]. Under compression, these materials have increased damage tolerance with
23 resistance to spalling [8, 9]. The improved mechanical properties of HPFRCC significantly increases
24 ductility, energy dissipation capacity and damage tolerance of structural members, especially in ex-
25 pected plastic hinge locations under seismic loadings [10]. Reinforced HPFRCC structural members
26 such as beams [11, 12, 13, 14], columns [15, 16], walls[17], and beam-column joints [18], among others,
27 have been physically tested in laboratories under quasi-static monotonic and reversed cyclic load-
28 ing, and found to be a promising alternative to improve structural performance and constructibility
29 compared to conventional reinforced concrete members. Unlike reinforced concrete, predicting the
30 plastic hinge behavior of components using ductile concrete materials has received limited response.
31 Recent literature has shown that the HPFRCC flexural members reinforced with mild steel pre-
32 dominantly fail by fracture of tensile reinforcement which contrasts the usual compression crushing
33 phenomenon in ordinary reinforced concrete members [19, 20]. Composite behavior of reinforced
34 HPFRCC is characterized by higher bond strength [21, 22, 23] which helps resist the formation of

2
35 splitting cracks leading to damage localization in the vicinity of a single transverse crack or few
36 cracks as seen in tension stiffening experiments [9, 24]. Experimental investigations of reinforced
37 HPFRCC beams conducted by Bandelt and Billington [19] shows that the number of flexural cracks
38 and the rate of damage progression in the plastic hinge region, increases with increasing reinforce-
39 ment ratio. Further, the deformation capacity of specimens under cyclic loading were found to be
40 lower than the corresponding monotonic specimens because of a rapid accumulation of plastic strain
41 in the localized crack region with increasing loading cycles. In another study, it was found that the
42 presence of large pulses at the beginning of loading history could increase the accumulation rate of
43 reinforcement strain in HPFRCC flexural members with low reinforcement ratios, and cause fracture
44 of reinforcement at drifts lower than specimens without pulses [25].
45 The plastic hinge region of reinforced HPFRCCs is a complex inelastic region with a highly non-
46 linear interaction between the reinforcement and HPFRCC matrix, characterized by multiple fine
47 distributed crack before yielding of the reinforcement and a small number (usually one or two) of
48 localized dominant cracks after yielding of the reinforcement [19, 26]. The formation and nature of
49 plastic hinge regions in a structural member govern the nominal load carrying capacity, as well as
50 the ultimate displacement, rotation, and energy dissipation capacities under extreme loadings. The
51 equivalent plastic hinge length, generally referred to as “plastic hinge length”, Lp , is the length over
52 which plastic curvature is assumed to be constant, such that the integrated area under the actual
53 plastic curvature is equal to the area under the assumed curvature [27]. This plastic hinge length
54 provides a simplified method to calculate total plastic rotation or deformation capacity of a member
55 undergoing large inelastic deformation using cross-sectional curvature analysis. The physical length
56 of a real plastic hinge region is close to the length over which the reinforcement yields, which will
57 be referred to as reinforcement yielding zone, Ly , in this paper [27, 28].
58 With availability of advance finite element modeling technique such as total strain-based models
59 [29, 30, 31] and recent development of bond-slip constitutive model that captures the complex
60 interaction between reinforcement and HPFRCC matrix [32]; numerical simulations are able to
61 predict damage stages such as yielding of reinforcement, reinforcement fracture, and damage pattern
62 similar to that of actual physical experimentation [32, 33]. In this paper, a finite element-based
63 approach is applied to investigate the plastic hinge region in reinforced HPFRCC members to avoid
64 significant cost and time associated with large scale experimental programs. Numerical simulations
65 are used as a tool to develop expressions of plastic hinge length for different HPFRCC material and
66 structural properties, and the proposed expressions are validated on a large experimental database.

3
 Table 1: Empirical models of plastic hinge length

Reference Expressions Element Loading

Baker [35] k1 k2 k3 d(z/d)1/4 RC Beam Monotonic

Sawyer [36] 0.25d + 0.075z RC Beam Monotonic
√
Corley [37] 0.5d[1 + (0.4/ d)(z/d)] RC Beam Monotonic
Mattock [38] 0.5d + 0.05z RC Beam Monotonic
Paulay and Priestley [27] 0.08L + 0.022fy db RC Beam & Column Monotonic
Naaman et al. [39]† 0.5(1.06 + 0.13ρV )d R-HPFRCC Beam Monotonic
Panagiotakos and Fardis [40] 0.18d + 0.021asl fy db RC Beam & Column Monotonic
0.12d + 0.014asl fy db RC Beam & Column Cyclic
Bae and Bayrak [41] [ {0.3(P/Po ) + 3(As /Ag ) − 0.1} (L/h) + 0.25]h ≥ 0.25h RC Column Cyclic
p
Berry et al. [42] 0.05L + 0.1fy db / fc0 RC Column Cyclic
Tariq et al. [43] (0.3 + 0.18ρ)d R-HPFRCC Beam Cyclic
† 1.5
Xu et al. [44] [10(fc /fy ) {Esh /(Es ρs )} − 0.12]d + 0.238z R-HPFRCC Beam Monotonic
†
Length of reinforcement yielding region

67 2. Existing Studies on Plastic Hinge Region in reinforced HPFRCC

68 Table 1 shows numerous plastic hinge length equations developed for reinforced concrete and
69 reinforced HPFRCCs under different loading conditions, and for various structural elements. The
70 expressions are based on regression analysis of either experimental, numerical, or combined exper-
71 imental and numerical studies carried out with variations in material and geometric properties.
72 Researchers have identified three distinct phenomena that causes the spread of plasticity in plas-
73 tic hinge region, namely: (a) moment gradient related to the increment of moment with length of
74 shear span or flexural depth, (b) tension shift related to shear distortion which invalidates the lin-
75 ear strain distribution assumption in a cross-section, and (c) tensile strain penetration which refers
76 to an increment in length of plasticity due to accumulation of significant strain inside a joint or
77 foundation through the bond-slip mechanism between the longitudinal reinforcement and the sur-
78 rounding matrix [27, 34]. Most of the primary variables that are used to define the plastic hinge
79 length expressions in Table 1, explicitly take into account one or two of these phenomena. For
80 example, length(L), shear span(z), shear span to depth (z/d), or effective depth (d) are associated
81 with moment gradient in a flexural member. Similarly, yield stress (fy ), diameter of longitudi-
82 nal reinforcement (db ), reinforcement ratio (ρ), or compressive strength (fc0 ) are related to strain
83 penetration. Most of the expressions ignore the tension shift phenomenon, assuming its effect to
84 be statistically insignificant, or incorporate this effect implicitly by calibrating the equation with
85 ultimate rotation or displacement obtained through experimental or numerical studies.
86 In regards to HPFRCC plastic hinge length, Naaman et al. [39] conducted an early study to quan-

4
 87 tify the plastic hinge length and used it to compute plastic rotation capacity of reinforced HPFRCC
88 flexural members through experimental testing of 12 reinforced HPFRCC beams containing steel
89 fibers under monotonic loading, and with reinforcement ratios ranging from 0.67% to 1.89%. An
90 empirical equation was proposed for plastic hinge length corresponding to the length of reinforcement
91 yielding zone, Ly (Table 1). The plastic curvature was shown to vary non-linearly from the reinforce-
92 ment yielding section to the section corresponding to ultimate strain( εcu = 0.003). The proposed
93 plastic hinge equation had to be multiplied by a curvature modification factor, β(= 0.56 − 0.16Vf ),
94 to compute the average plastic curvature over the reinforcement yielding zone, Ly , in order to cal-
95 culate plastic rotation capacity, θp , of the flexural member. However, the curvature modification
96 factor was developed for steel fibers used in that particular study and it is unclear if this equation
97 holds true for other types of fiber-reinforcing materials (e.g., polymeric fibers). Further, the fail-
98 ure criteria chosen (i.e., εcu = 0.003) for the calculation of plastic rotation capacity for reinforced
99 HPFRCC beams is too conservative because a recent experimental study has shown that the strain
100 in extreme compression fiber can reach a value up to 0.03 before extensive softening of cementitious
101 matrix begins followed by the failure of the member due to fracture of the tensile reinforcement
102 [45]. Similarly, several other studies have concluded that the reinforced HPFRCC flexural members
103 do not lose significant load-carrying capacity and undergo considerable plastic deformation before
104 failure occurs by the rupture of longitudinal reinforcement [19, 25, 26, 46].
105 Parra-Montesinos and Chompreda [13] conducted an experimental study to investigate the de-
106 formation capacity and shear strength of structural members constructed with steel fiber-reinforced
107 cement composites (FRCC). In order to estimate the plastic rotation capacity under reversed cyclic
108 loading, the authors used an equivalent plastic hinge length, Lp , equal to the “effective depth” of
109 the cross-section; however, this was assumed based on literature by Blume et al. [47] used for the
110 analysis of conventional reinforced concrete buildings under seismic loading.
111 Xu et al. [44] performed a parametric study of reinforced ECC beams under monotonic loading
112 by varying material and cross-sectional properties, and proposed an equation to predict the length
113 of reinforcement yielding zone, Ly (Table 1) . This length is close to the physical plastic hinge
114 length over which the seismic detailing is required for the ductile performance of the plastic hinge
115 region, but it cannot be used directly to compute plastic rotation capacity of a HPFRCC member,
116 unless curvature distribution data is available along the reinforcement yielding zone, Ly . Further,
117 failure of the beams were observed to be by crushing of the compression zone material as most of the
118 beams were singly reinforced beams; however, in ductile seismic design practice, doubly reinforced
119 symmetrical cross-sections are more common in flexural members because of the expected moment
120 reversal under seismic loading [27, 48].

5
121 Tariq et al. [43] used a fiber-based lumped plasticity modeling technique to calibrate experi-
122 mentally tested reinforced ECC beams by minimizing the error in the force-deformation backbone
123 response and collapse level drift. An expression for equivalent plastic hinge length, Lp (Table 1),
124 was proposed that could reasonably predict fracture strain at collapse drift in specimens under dif-
125 ferent types of cyclic loading; however, the equation was based on a limited database of cantilever
126 beams with variation in cross-sectional dimensions and tensile reinforcement ratio only, and only
127 one HPFRCC material was considered.
128 Pokhrel and Bandelt [49, 33] conducted an investigation with variation in mechanical properties,
129 geometric properties, and boundary conditions to identify potential variables affecting deformation
130 capacity and plasticity (i.e., reinforcement yielding zone, Ly ) under monotonic loading. The study
131 showed that there was a substantial decrease in the deformation capacity, length of plasticity (Ly )
132 and length of rebar slippage with increase in tensile strength, ft . Tensile strength influenced these
133 factors as strain concentrated over a shorter gage length of the reinforcement, leading ultimately to
134 early fracture of the tensile reinforcement. In addition to that, the change in boundary conditions
135 from cantilever to simply-supported (i.e., end-hinges to center hinges) resulted in a reduction of
136 deformation capacity in reinforced HPFRCC beams. Cantilever beams had higher deformation
137 capacity because of a larger spread of plasticity inside the joint of the foundation due to strain
138 penetration which increases plastic rotation capacity in ordinary reinforced concrete members as
139 well [27, 34, 42]. This study, however, did not quantify plastic hinge length and did not consider the
140 effect of cyclic loading on the length of plasticity in reinforced HPFRCC members.

141 3. Numerical Simulation Description and Validation

142 3.1. Finite element models

143 The finite element software DIANA FEA Version 10.2 [50] was used to create two dimensional
144 finite element models with variation in material properties, section geometry, and boundary condi-
145 tions. Representative finite element models of simply supported and cantilever beams used in the
146 study are shown in Fig. 1. The test setup and specimen dimensions for the numerical simulations
147 were selected because they are representative of reinforced HPFRCC component experiments tested
148 under monotonic and cyclic loading to high deformations including fracture of the reinforcement
149 [11, 19, 51, 25].
150 Elastic support and loading plates were used at both sides of the beams to prevent inelastic
151 stress concentration at the interface between plates and HPFRCC beam elements. Vertical and
152 lateral springs were modeled as uniaxial springs at the base of the support plates with stiffness to
153 replicate the softness of the plate supports as in the physical experiment [20, 32]. Foundation beams

6
 180 mm

Applied
Elastic Support Displacement

Nonlinear HPFRCC
Applied Steel Truss and
Displacement 10 x 10 mm Line Interface
10 x 10 mm mesh size Steel Truss and mesh size Element
with 130 mm thickness Line Interface Element with 130 mm Shear span (Ls)
thickness

180 Steel Truss

mm Element
Compression
Only Spring
Steel Truss Element
Compression Only
Spring
Shear span (Ls) 380 mm
(a)

Compression 800 mm
Only Spring
(b)

Figure 1: Finite element geometry of (a) simply supported beam (b) cantilever beam

154 with dimension (l × b × h) of 800 mm × 130 mm × 380 mm were used as the fixed base support for
155 the cantilever beams similar to the experimental testing conducted by Frank et al. [20].
156 There are several potential variables that can affect the formation, spread, and length of plastic
157 hinge region in reinforced HPFRCC members. In this study, a total of 72 finite element beam
158 simulations with variation in shear span, reinforcement ratio, tensile and compressive property of
159 HPFRCC, boundary condition, and loading type were considered as discussed in Sec. 3.1.1 - 3.1.5.

160 3.1.1. Shear Span

161 Shear span (Ls ) of a structural member is the length between maximum moment and zero
162 moment. Shear span is directly related to the moment gradient in a plastic hinge region of a
163 structural element. The equations listed in Table 1 for reinforced concrete members show that the
164 equivalent plastic hinge length is directly proportional to the shear span. To investigate the length
165 of plastic hinge regions in reinforced HPFRCC, shear span lengths of 650 mm and 1080 mm were
166 chosen corresponding to shear span-to-depth ratio (Ls /d) of 4.06 and 6.75 respectively. Specimens
167 with a shear span-to-depth ratio of 6.75 are considered to be dominated by flexural deformations;
168 while specimens with a shear span-to-depth ratio of 4.06 are considered to have contribution in
169 deformation from both flexure and shear stresses [48].

170 3.1.2. Reinforcement Ratio

171 Reinforcement ratio is directly related with the maximum moment capacity, ultimate rotation
172 capacity and tensile strain penetration phenomenon in a reinforced concrete or HPFRCC member.

7
173 Symmetrical longitudinal reinforcement on top and bottom with areas of 142 mm2 , 258 mm2 , and 398
174 mm2 corresponding to reinforcement ratios of 0.70%, 1.25% and 1.90% were used. It is recognized
175 that increasing the area of compression steel limits the potential for damage on the compression side
176 of the members; however, symmetrical reinforcement was selected to make comparisons to results
177 from the available database in which the members are generally symmetrically reinforced for seismic
178 applications. All of the beams were modeled with transverse reinforcement of 16 mm2 with uniform
179 spacing of 75 mm.

180 3.1.3. Boundary Conditions

181 Two boundary conditions were considered in this investigation such that plastic hinges could form
182 at the center of a simply supported beam and at the end of a cantilever beam integrally connected
183 to a foundation element. Previous research studies on HPFRCCs were conducted either on simply
184 supported or cantilever elements; however, the effects of end conditions have not been systematically
185 studied to identify the boundary conditions effects in the spread of plasticity in reinforced HPFRCC
186 members.

187 3.1.4. Tensile and Compressive Property

188 In order investigate the effect of tensile strength and compressive strength on the length of the
189 plastic hinges in reinforced HPFRCC members, the mechanical response of three types of materi-
190 als, namely, an Ultra-High Performance Concrete (UHPC), an Engineered Cementitious Composite
191 (ECC) and a Hybrid Fiber Reinforced Concrete (HyFRC) were used in the simulations. Each of these
192 materials has unique mechanical properties that are promising for various structural applications.
193 UHPC, which is increasingly used in bridge construction for improving durability and accelerated
194 construction techniques [52], is an ultra high strength material with tensile and compressive strengths
195 in the range of 5.5-11 MPa [53] and 112-210 MPa [53] respectively. ECC, developed by Li and Le-
196 ung [54] with principles of micro-mechanics such that it exhibits ultra high tensile ductility under
197 uniaxial test, has tensile and compressive strengths in the range of 2.5-5.0 MPa [55] and 30-90 MPa
198 [56], respectively. HyFRC uses a hybridization of fibers across various length scales for durability
199 and seismic applications [57, 9], and has tensile and compressive strengths in the range of 1.5-3.1
200 MPa [58, 59] and 31-45 MPa [24, 58] respectively.
201 The three materials were chosen for this study because of the range of their mechanical prop-
202 erties associated with their distinctive fiber compositions, and their wide use in various large-scale
203 experimental programs. UHPC, which is also known as, Ultra-High-Performance Fiber-Reinforced
204 Concrete (UHP-FRC), generally contains 1%−2.5% by volume fraction (Vf ) of smooth, hooked, or
205 twisted steel fibers with an aspect ratio (lf /df ) of 65−100 [60, 52], whereas ECC mixtures mostly

8
206 contain 2% by volume fraction of polyvinyl alcohol or polyethylene fibers with an aspect ratio of
207 300−350 [13, 58]. HyFRC consists of a hybrid combination of steel macro-fibers and polyvinyl
208 alcohol micro-fibers with a volume fraction of 0.5%−1.3% and 0.2%−0.3%, respectively [59, 58].
209 Macro-fibers in HyFRC mixtures generally have an aspect ratio of 55−80, whereas micro-fibers have
210 an aspect ratio of 200−250 [59, 58]. The fiber types and mixture constituents were not simulated
211 numerically; however, the mechanical response associated with each mixture was included in the
0
212 simulation. Representative tensile strengths (ft ) and compressive strengths (fc ) corresponding to
213 UHPC (8 MPa & 120 MPa), ECC (2.9 MPa & 55 MPa), and HyFRC (1.8 MPa & 45 MPa) were
214 used in the simulation. These mechanical properties were selected based on the uniaxial tension and
215 compression tests conducted by Wille and Naaman [60], Russell and Graybeal [52] and Moreno-Luna
216 [58], whose mix compositions did not use coarse aggregates as a constituent material for UHPC and
217 ECC.

218 3.1.5. Load Type

219 Two types of loading were used on the reinforced members in this study: (a) monotonic loading,
220 and (b) reversed cyclic loading. Monotonic loading was applied using incremental displacement-
221 based approach at a step size of 0.25 mm until the fracture of longitudinal reinforcement or crushing
222 of HPFRCC was observed. Reversed cyclic loading was applied using the FEMA 461 [61] loading
223 protocol such that two successive cycles with the same maximum amplitude of drift was applied.
224 The amplitude of drift was increased by 40% for the next two successive cycles, and this process
225 was repeated until the failure was observed. The first two cycles had maximum amplitude of 0.15%
226 drift, where the drift was obtained by normalizing vertical deflection by shear-span (i.e., ∆/Ls ). A
227 total of 72 finite element models were simulated in this study with variability in the loading scenario,
228 boundary condition, shear span, tensile reinforcement ratio, and HPFRCC type (2 × 2 × 2 × 3 × 3).

229 3.2. Material models and analysis parameters

230 3.2.1. HPFRCC material model

231 A total strain-based fixed-crack model was used as the constitutive model to simulate the behavior
232 of HPFRCC materials [29]. A constant shear retention factor (βs ) of 1% [32] was used to simulate the
233 transfer of shear stress across the cracks. Although there are more refined approaches to simulate the
234 shear transfer mechanism pertaining to fiber bridging action in various types of HPFRCC materials
235 [31], same shear retention factor was used for all three types of HPFRCCs in this study because of the
236 flexure dominant nature of most of the simulated beams [62], and to avoid numerical convergence
237 issue and severe stress locking associated with constant high shear retention factor under large
238 inelastic deformation [63]. The HPFRCC materials were modeled using eight-noded quadratic plane

9
 σ σ
fc’
2 3 𝐺𝑓 𝐺𝑐
ft 𝐴𝑟𝑒𝑎 𝑜𝑓 𝑡𝑟𝑖𝑎𝑛𝑔𝑙𝑒 =
𝐴𝑟𝑒𝑎 =
𝑓𝑟𝑜𝑚 𝑝𝑜𝑖𝑛𝑡𝑠 3 𝑡𝑜 4 ℎ ℎ

E fc′
3
1 4

et0 etp etu e

(a) eco ecc ecu e
(b)
σ u
umax
fy
k
1

E
uf
ey eu e s1 smax s
(c) (d)

Figure 2: (a) Multilinear tensile stress-strain parameters of HPFRCC and contours of principal
tensile strains (b) parabolic compressive stress-strain parameters of HPFRCC (c) uniaxial tensile
stress-strain parameters of steel (d) Bond-slip parameters

239 stress elements. The size of each element was 10 × 10 mm and the thickness was 130 mm. A 3
240 × 3 Gauss integration scheme and quadratic interpolation was used in the finite element numerical
241 formulation. The material tensile response was modeled using an idealized multi-linear stress-strain
242 curve (Fig. 2 (a)), as it has been successfully used by researchers to simulate composite action
243 between steel reinforcement and HPFRCC material [58], and component level response of reinforced
244 HPFRCC flexural members [32, 25, 43, 49].
245 Parameters such as initial stiffness (E), maximum tensile stress (ft ), and tensile strain at the
246 onset of softening (εtp ) were obtained from uniaxial tensile test data [60, 58] and are summarized in
247 Table 2. The ultimate tensile strain (εtu ) was based on the experimentally observed tensile fracture
248 energy and was calculated using a linear softening model as shown in Eq. 1 [58].

Gf 1
εtu = 2 + εtp (1)
h ft
249 The compression response of the HPFRCC materials was modeled using a parabolic stress-strain
250 curve based on compressive fracture energy [64] as shown in Fig. 2 (b). The compression stress-strain
0
251 parameters such as compressive strength (fc ), compressive fracture energy (Gc ), and Poisson’s ratio

10
252 (ν) were based on experimental data [52, 58] and are summarized in Table 2. The initial compressive
253 strain (εco ), compressive strain at maximum compressive strength (εcc ) and the ultimate compressive
254 strain (εcu ) were calculated using the parabolic model as shown in Eq. 2 [50].

0
1 fc
εco = 3 E
0
5 fc (2)
εcc = 3 E
3 Gc 1
εcu = 2 h fc0 + εcc
√
255 In Eq. 1 and 2, h is the crack bandwidth which is equal to A, where A is the area of an
256 individual plane stress element. As only one mesh discretization is presented in this paper, crack-
257 band approach was used to eliminate the problem related with mesh dependency [31, 58, 65]. Other
258 mesh sizes were explored, but were not shown to influence the results due to the crack-band approach.
259 The cyclic tensile and compressive behavior of HPFRCC was represented by secant unloading and
260 reloading path through origin. There are more refined approaches as well [30, 31], but the secant
261 path approach has less convergence issue, and has been previously used by researchers in simulating
262 component level response of reinforced concrete [66] and reinforced HPFRCC [32] with sufficient
263 accuracy. Kesner et al. [8] conducted an experimental investigation to characterize uniaxial cyclic
264 response of HPFRCC material and found out that there was 50% to 80% reduction in tensile fracture
265 energy of HPFRCC material under cyclic loading compared to monotonic loading. Therefore, to
266 accurately simulate inelasticity and damage in HPFRCC under cyclic loading, a damaged fracture
267 energy equal to 50% of the monotonic fracture energy (Gf ) was utilized for simulations under cyclic
268 loading.

269 3.2.2. Steel material model

270 Longitudinal reinforcement was modeled as a three-noded truss element with an element size
271 of 10 mm. A three-point Gauss integration scheme was used for numerical integration of the truss
272 element. Transverse reinforcement was modeled as a two-noded directly integrated truss elements
273 with the same element size as longitudinal reinforcement. The constitutive behavior of longitudinal
274 reinforcement was modeled using the Von Mises plasticity strain hardening model [67] for monotonic
275 loading. The longitudinal reinforcement under cyclic loading was modeled using a modified two-
276 surface plasticity model [68] which includes the Bauschinger effect essential to capture the hysteretic
277 response in beams under cyclic loading [66, 69] as discussed in Sec. 3.3. The tensile stress-strain curve
278 parameters such as modulus of elasticity (E), yield stress (fy ), ultimate stress (fu ), and ultimate
279 strain (εu ) of the longitudinal reinforcement were based on representative uniaxial tensile tests of
280 steel reinforcement conducted by Bandelt and Billington [19] as shown in Fig. 2 (c) and are listed
281 in Table 2. The constitutive behavior of the transverse reinforcement was modeled using the Von

11
282 Mises plasticity model without strain hardening [67], and the tensile stress-strain curve parameters
283 such as modulus of elasticity (E) and yield stress (fy ) were taken from experimental data of Bandelt
284 and Billington [19] and are tabulated in Table 2.

285 3.2.3. Bond-slip model

286 The modeling of a bond-slip mechanism can improve the simulated component level response
287 in terms of strength, stiffness, deformation capacity and damage pattern in reinforced HPFRCC
288 members [32]. The longitudinal reinforcing elements were connected to the HPFRCC elements
289 through a six-noded line interface element. A three-point Newton-Cotes integration scheme was used
290 for the line interface element. The bond-slip model (Fig. 2 (d)) proposed by Bandelt and Billington
291 [32] was used to simulate the interaction between longitudinal reinforcement and HPFRCC. The
292 bond-slip constitutive relationship was developed based on extensive bond-slip experimental testing
293 under monotonic and cyclic loading for ECC material [22, 23], and was extended for other classes
294 of HPFRCC material (e.g., UHPC and HyFRC) by changing parameters such as maximum bond
295 strength (umax ), slip at onset of softening (smax ), bond-slip softening stiffness (k), and residual
296 friction bond strength (uf ). These parameters were obtained from bond-slip test data of Dagenais
297 and Massicotte [70] for UHPC, and similarly for ECC and HyFRC these parameters were obtained
298 from the test data of Bandelt and Billington [19].

299 3.2.4. Analysis Parameters

300 A nonlinear static analysis with an incremental displacement-based loading was used for the
301 simulation. The details about monotonic and cyclic loading have already been discussed in Sec. 3.1.5.
302 There were two types of failure observed in this study: (a) fracture of longitudinal reinforcement,
303 and (b) crushing of HPFRCC material. Fracture was assumed to occur when the strain over a 30
304 mm gage length of the longitudinal reinforcement exceeded an 18% assumed fracture strain based on
305 the approach outlined in Bandelt and Billington [32]. Crushing of compression zone was assumed to
306 occur when the compressive strain in a single HPFRCC element reached the ultimate compressive
307 strain (εcu ) based on compression model shown in Fig. 2 (b). A regular Newton-Raphson scheme
308 was used for equilibrium criteria and a line search algorithm was used for numerical convergence.
309 Convergence was assumed to have occurred at each iteration if either energy, displacement, or force
310 norm did not exceed limiting values of 0.01%, 0.1% and 1%, respectively.

311 3.3. Finite Element Model Verification

312 Using the finite element modeling technique described in Sec. 3.2, two reinforced HPFRCC
313 beams were simulated under monotonic loading (M-1.3 and ECC-M-1.3) and two were simulated

12
 Table 2: Bond-slip, HPFRCC and steel material properties

Property Notation Unit UHPC ECC HyFRC Longitudinal Transverse

Steel Steel

Maximum bond strength umax [MPa] 10.4 8.20 7.60 - -

Slip at onset of softening smax [mm] 1.0 (0.6) 1.0 (0.6) 1.0 (0.6) - -
Bond-slip softening stiffness k [MPa/mm] 1.0 (5.0) 1.0 (5.0) 1.0 (5.0) - -
Residual friction bond strength uf [MPa] 3.8 (2.2) 3.0 (1.8) 2.8 (1.6) - -
Tensile strength ft [MPa] 8.0 2.9 1.8 - -
Strain at crack initiation εt0 [%] 0.0191 0.0169 0.0104 - -
Strain at onset of softening εtp [%] 0.20 0.75 0.30 - -
Tensile fracture energy Gf [MPa-mm] 19 (9.5) 6.1 (3.1) 5.3 (2.7) - -
0
Compressive strength fc [MPa] 120 55 45 - -
Compressive fracture energy Gc [MPa-mm] 180 53 61 - -
Young’s modulus of elasticity E [Gpa] 42 17 17 200 205
Poisson’s ratio ν [mm/mm] 0.18 0.15 0.15 0.30 0.30
Shear retention factor βs [%] 1.0 1.0 1.0 - -
Yield strength fy [MPa] - - - 455 690
Ultimate strength fu [MPa] - - - 675 -
Strain at ultimate strength εu [%] - - - 16 -

Note: The values inside parenthesis are only used for simulation under cyclic loading.
All the other values are same for simulations under either monotonic or cyclic loading.

314 under reversed cyclic loading (C-1.3 and ECC-0.95-F). While detail description of the specimens
315 can be found in the literature [19, 71, 25], some important material and geometric properties of
316 the specimens are summarized in Table 3. Specimens M-1.3 and C-1.3 were simply supported test
317 specimens, whereas ECC-M-1.3 and ECC-0.95-F were tested under cantilever test setup. It can be
318 observed from Fig. 3 that the initial stiffness, load carrying capacity, and drift capacity at two damage
319 levels (i.e., yield and collapse) are captured reasonably well through the numerical simulation for all
320 the cases. However, there is over prediction of load at reinforcement fracture in specimen ECC-M-1.3
321 because the necking region of steel reinforcement was not modeled in the numerical simulation due
322 to numerical convergence issues [32, 25, 43]. The failure mode in all of the numerically simulated
323 members was due to the fracture of the longitudinal reinforcement which is in agreement with the
324 experimental observations (Fig. 4). Under monotonic loading, fracture was observed at one major
325 flexural crack location in the physical specimen M-1.3 (Fig. 4 (a)), whereas fracture was simulated
326 at two major flexural cracks in finite element model (Fig. 4 (b)) because of the nearly symmetrical
327 damage localization on either side of mid-span in the numerical simulation.
328 Since the plastic hinge region is associated with a high level of inelastic damage and cracking, it is
329 important to closely simulate the damage pattern. Fig. 4 shows the comparison of damage pattern
330 observed in test specimens and the principal tensile strain contours obtained through numerical

13
 Table 3: Material and geometrical properties of validation specimens

0
Notation Loading Test Ls Cross Section HPFRCC Longitudinal Reinforcement (ρ = ρ )
0
Type Setup [mm] b[mm] d[mm] h[mm] ft [MPa] fc [MPa] fy [MPa] db [mm] n As [mm2 ]

M-1.3 [19] Monotonic SSB 690 130 160 180 2.2 46 455 13 2 258
ECC-M-1.3 [71] Monotonic CLB 760 130 160 180 4.0 45 455 13 2 258
C-1.3 [19] Cyclic SSB 690 130 160 180 2.2 46 455 13 2 258
ECC-0.95-F [25] Cyclic CLB 810 170 170 200 4.0 45 455 13 2 258

SSB = Simply supported beam and CLB = Cantilever beam

90 35

30
75
Applied Load [kN]

Applied Load [kN]

25
60
20
45
15

30
10
ECC-M-1.3 [Exp]
M-1.3 [Exp]
15 M-1.3 [Num]
ECC-M-1.3 [Num]
5
Yield Yield
Fracture Fracture
0 0
0 3 6 9 12 15 0 3 6 9 12 15 18

(b) Δ/Lshear-span [%]

(a) Δ/Ls [%] [%]
Δ/Lshear-span Δ/Ls [%]

90 40
75
30
60
45 20
Applied Load [kN]

30
Applied Load [kN]

10
15
0 0
-15
-10
-30
-45 -20
-60 C-1.3 [Exp] ECC-0.95-F [Exp]
C-1.3 [Num] -30 ECC-0.95-F [Num]
-75 Yield Yield
Fracture Fracture
-90 -40
-12 -9 -6 -3 0 3 6 9 12 -12 -9 -6 -3 0 3 6 9 12

(c) Δ/Lshear-span [%] Δ/Ls [%][%]

Δ/Ls [%] (d) Δ/Lshear-span

Figure 3: Experimental and simulated applied force vs drift response of (a) M-1.3 [19] (b) ECC-M-1.3
[71] (c) C-1.3 [19] and (d) ECC-0.95-F [25]

14
 Major Flexure-Shear Cracks (2)
Reinforcement Fracture

(a) Major Flexural Cracks (2)

Major Flexure-Shear Cracks (2)

Reinforcement Fracture Major Flexural Cracks (2)

(b)
Reinforcement Fracture
Minor Flexure-Shear Cracks (3) Minor Flexure-Shear Cracks (3)

(c) Major Flexural Crack(1)

Reinforcement Fracture
Major Flexure-Shear Cracks (3) Major Flexure-Shear Cracks (3)

Major Flexural Cracks (2)

(d)

Figure 4: (a) Experimental damage pattern in M-1.3 [19] (b) simulated principal tensile strain
contours in M-1.3 (c) experimental damage pattern in C-1.3 [19] and (d) simulated principal tensile
strain contours in C-1.3 before the failure

15
331 simulation, just before failure. Principal tensile strain contours, indicative of damage and cracking
332 in the beams, are associated with varying levels of strain as shown in Fig. 2 (a). The contours are
333 based on the stress-strain relationship of the HPFRCC in tension, and represents elastic behavior,
334 pseudo-strain hardening, softening, and dominant crack formation. In the case of monotonically
335 loaded test specimen M-1.3 (Fig. 4 (a) and (b)), the primary crack patterns (i.e., major flexural
336 cracks and flexure-shear cracks) are in good agreement with the numerical simulation crack patterns,
337 indicating the capability of the finite element modeling technique to simulate damage patterns.
338 Under reversed cyclic loading, it can be seen that the amount of damage increased in both test
339 specimen and simulation compared to the monotonically loaded specimen. There was a single major
340 flexural crack in the test specimen C-1.3 (Fig. 4 (c)), whereas there were two major flexural crack in
341 the simulated crack pattern (Fig. 4 (d)) because of the nearly symmetrical response on either side
342 of the mid-span in the numerical simulation. However, reinforcement fracture was only simulated
343 at one side of the beam because of the increased effect of the asymmetric boundary condition on
344 the accumulation of reinforcement strain under large cyclic drifts. The flexure-shear crack widths
345 were over predicted in Fig. 4 (d) and is one of the limitations of this numerical simulation technique;
346 however, the location and type of cracks (flexure-shear) and reinforcement rupture location were
347 closely simulated under reversed cyclic loading.

348 4. Plastic Hinge Region in Reinforced HPFRCC

349 The plastic hinge region in ordinary reinforced concrete structural element consists of the rein-
350 forcement yielding zone, concrete crushing zone, and curvature localization zone [72]; however, in
351 reinforced HPFRCC structural member, the plastic hinge region comprises of reinforcement yield-
352 ing zone, tensile strain concentration zone, and curvature localization zone. HPFRCC compression
353 crushing was observed in two finite element models out of seventy-two finite element simulations
354 carried out in this study, which is significantly less than what would be expected in reinforced con-
355 crete members. The failure of seventy simulated HPFRCC specimens occurred due to fracture of
356 the tensile reinforcement due to concentration of inelastic strain at dominant cracks in HPFRCC
357 members as explained experimentally in the tension stiffening experiments conducted by Moreno
358 et al. [24]. Additionally, under monotonic or cyclic flexural loading, several experimental studies
359 have shown that the failure of reinforced HPFRCC beams occurs due to fracture of longitudinal re-
360 inforcement rather than crushing of compression zone material, unlike ordinary reinforced concrete
361 [11, 13, 19, 20, 25]. The details of reinforcement yielding zone and tensile strain concentration zone
362 are discussed in Sec. 4.1 and discussion regarding curvature localization zone is provided in Sec. 4.2.

16
363 4.1. Strain Distribution
364 Tensile strain in the longitudinal reinforcement was extracted at collapse level drifts to understand
365 how reinforcement strain varied along the shear-span of the structural member. The collapse level
366 drift is the drift level at which the strain in the longitudinal reinforcement exceeded the assumed
367 fracture strain of 18% as previously discussed for finite element verification purposes in Sec. 3.3.
368 Fig. 5 shows a plot of reinforcement tensile strain in UHPC, ECC, and HyFRC under monotonic
369 and cyclic loading along the length of the members. The tensile strain plot for both simply supported
370 beams (Fig. 5 (a) & (b)) and cantilever beams (Fig. 5 (c) & (d)) have been shown for representative
371 simulations with a shear-span (Ls ) equal to 650 mm and reinforcement ratio (ρ) equal to 1.25%.
372 The length of the tensile longitudinal reinforcement over which the strain exceeded the yield
373 strain of steel is defined here as the reinforcement yielding zone (Ly ). It can be observed from the
374 numerical simulation results , that regardless of the loading type, Ly increased with a decrease in
375 tensile strength and matrix toughness (i.e., (Ly )HyF RC > (Ly )ECC > (Ly )U HP C ) at the collapse
376 level for both types of boundary conditions. Comparisons between the length of reinforcement
377 yielding under monotonic and cyclic loading for the same setup and HPFRCC material shows that
378 Ly is longer under cyclic loading than monotonic loading in eight out of twelve simulation results.
379 The experimental study carried out by Bandelt and Billington [19] also showed similar results (i.e.,
380 large number of distributed cracks and longer length of damage) under cyclic loading compared to
381 monotonic loading for identical simply supported beam specimens.
382 Reinforcement tensile strain plots at various drift levels revealed that the strain increment was
383 higher along reinforcement hardening length (εs > εsh ) compared to the other regions. The length
384 of this hardening region is defined herein as tensile strain concentration zone (Lsc ). Plots of rein-
385 forcement strain (Fig. 5 (a) & (b)) shows that the tensile strain concentration length, Lsc , is shorter
386 under cyclic loading compared to monotonically loaded simply-supported beam specimens. Investi-
387 gation of principal tensile strain contour revealed that the number of dominant cracks were lower
388 under cyclic loading such that damage localized more rapidly, and reinforcement strain increased
389 exponentially within a short length in cyclically loaded simply supported beams. In the case of
390 cantilever beams (Fig. 5 (c) & (d)), analysis shows that the region of tensile strain concentration
391 was found to be marginally higher for monotonically loaded specimens compare to those cyclically
392 loaded. The reason for a longer Lsc under monotonic loading in the cantilever beam setup was due
393 to a higher number of dominant cracks as compared to specimens loaded cyclically.

394 4.2. Curvature Distribution

395 The curvature of a cross-section was calculated using Eq. 3. This equation is based on the
396 assumption that the plane section perpendicular to the axis of bending will remain perpendicular

17
 MONOTONIC CYCLIC
20 20
Fracture Strain
Simply Supported Beams

Fracture Strain

Reinforcement Tensile Strain [%]

18 18 UHPC
Reinforcement Tensile Strain [%]

UHPC ECC
16 ECC HyFRC
16
HyFRC
14 14
12 12
10 10
8 8
Ly = 310 Lsc= 20
6 Lsc = 210 6
4 Ly = 230 4
2 2
0 0
300 400 500 600 700 800 900 1000 1100 300 400 500 600 700 800 900 1000 1100
(a) Distance from Support [mm] (b) Distance from Support [mm]
750 750
Fracture Strain Fracture Strain
700 700 Foundation Level
Foundation Level
650 UHPC
Distance from Base [mm]

UHPC
Distance from Base [mm]

650 ECC
Cantilever Beams

600 Ly = 250 ECC

600
Ly= 250 HyFRC
HyFRC
550 Lsc= 240 550 Lsc= 230
500 500
450 450
400 400
350 350
300 300
250 250
200 200
150 150
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
(c) Reinforcement Tensile Strain [%] (d) Reinforcement Tensile Strain [%]

Figure 5: Longitudinal reinforcement strain vs. distance from support in simply supported beams
under (a) monotonic and (b) cyclic loading; distance from base vs. longitudinal reinforcement strain
in cantilever under (c) monotonic and (d) cyclic loading for simulations with Ls = 650 mm &
ρ = 1.25% at collapse level

18
397 after the flexural bending.

εs
φ= (3)
d−c
398 where, εs is the strain in tensile longitudinal reinforcement, d is the distance from tensile longitudinal
399 reinforcement to the extreme compression fiber, and c is the distance from neutral axis to the extreme
400 compression fiber. Fig. 6 shows plots of curvature distribution in UHPC, ECC, and HyFRC under
401 monotonic and cyclic loading just prior to collapse (i.e., reinforcement fracture). The curvature
402 distribution plot for both simply supported beams (Fig. 6 (a) & (b)) and cantilever beams (Fig. 6
403 (c) & (d)) have been shown for representative simulations with shear-span (Ls ) equal to 650 mm
404 and reinforcement ratios (ρ) equal to 1.25%. After yielding of the tensile reinforcement, section
405 curvatures along a small region near the dominant crack section increases exponentially compared
406 to other regions of the shear-span. This region of sharp curvature increase contributes more towards
407 plastic deflection of the member and is referred to as curvature localization zone (Lcl ). The curvature
408 localization phenomenon has been observed experimentally in reinforced HPFRCC beams [39]. For
409 the purposes of the discussion in this study, curvature localization length, Lcl , is defined as the length
410 over which a curvature value of 2.5×10−4 mm−1 was exceeded as shown in Fig. 6. The curvature
411 localization phenomenon was found to be most pronounced in UHPC, followed by ECC, and finally
412 HyFRC under both types of loading and boundary conditions. This is due to relatively shorter
413 length of strain concentration zone, Lsc , in UHPC followed by a longer Lsc in ECC and the longest
414 Lsc in HyFRC as discussed in Sec. 4.1.
415 Fig. 6 (a) & (b) show that the curvature distribution in two sets of simply supported beams
416 (UHPC and ECC) are highly localized under cyclic loading than under monotonic loading (i.e.,
417 Lcl,cyclic < Lcl,monotonic ). However, all sets of cantilever beams (Fig. 6 (c) & (d)) subjected to
418 cyclic loading, exhibit a more uniform curvature distribution compared to the monotonically loaded
419 beams (i.e., Lcl,cyclic > Lcl,monotonic ). The difference between curvature localization in the two
420 boundary conditions is due to the transfer of post-yield strain into the foundation through bond-slip
421 mechanism between longitudinal reinforcement and HPFRCC matrix. This phenomenon results into
422 a more uniform plastic curvature distribution aided by foundation softening due to large cyclic load
423 reversal, which allows the damage to propagate inside the foundation region of the cantilever beam.
424 This causes the formation of longer curvature localization zone in cantilever beams under cyclic
425 loading compared to the monotonic loading. Yuan and Wu [73] conducted an extensive study to
426 investigate the effect of different loading scenarios on the plastic hinge length of reinforced concrete
427 flexural member with foundations and found that the length of curvature localization zone was
428 longer in members under cyclic loading compared to the monotonic loading scenario due the effect

19
429 of reversed load cycling.

430 5. Equivalent plastic hinge length (Lp )

431 The ultimate displacement, rotation, or ductility capacity of a reinforced HPFRCC structural
432 member can be calculated using linear strain distribution assumptions (i.e., pure flexural behavior)
433 and the concept of equivalent plastic hinge length. This mathematical approach has been extensively
434 used for the analysis of conventional reinforced concrete member [27]. For example, if the equivalent
435 plastic hinge length (Lp ) is known, the total (or ultimate) rotation capacity of a reinforced concrete
436 or HPFRCC structural element can be calculated using Eq. 4.

1
θu = θy + θp = φy Ls + (φu − φy )Lp (4)
2

437 In Eq. 4, φy and φu are the curvatures at the critical section of the structural member at yield
438 and collapse. The equivalent plastic hinge length, Lp , is defined as the fictitious length near the
439 critical section of a structural member over which the plastic curvature is assumed to be constant,
440 such that the integrated area under the actual plastic curvature is equal to the area under the
441 idealized plastic curvature [27]. This mathematical approach of concentrating the inelasticity in a
442 small region is analogous to a mechanical hinge, and therefore, widely used in industry for lumped
443 plasticity based modeling of structural components in buildings and bridges [5, 74]. Further, the
444 equivalent plastic hinge length lumps the effects of tension shift due to shear cracking, bond-slip
445 phenomenon and tension-stiffening in the plastic hinge region implicitly and is the conventional way
446 of calculating displacement parameters at collapse level through hand calculations [27, 40].

447 5.1. Theoretical Formulation

448 Equivalent plastic hinge length was computed by numerical integration and structural mechan-
449 ics. The theoretical formulation can be understood using a cantilever beam with shear-span, Ls ,
450 subjected to a lateral load, P , as shown in Fig. 7 (a). The collapse level curvature distribution of the
451 cantilever would be highly non-linear (Fig. 7 (b)) with a significant portion of longitudinal reinforce-
452 ment near the section of maximum moment exceeding the yield stress. This curvature distribution
453 at collapse can be idealized and divided into two regions: elastic curvature and plastic curvature
454 (Fig. 7 (c)). Thus, the total rotation capacity of the cantilever can be calculated as the sum of
455 the yield rotation, θy , and plastic rotation, θp , as shown in Fig. 7(d). Using numerical simulation,
456 the yield rotation can be calculated by integrating the curvature distribution, at the yielding of the
457 reinforcement, along the shear-span and length of the reinforcement embedded inside the foundation

20
 MONOTONIC CYCLIC
0.005
UHPC
Simply Supported Beam

UHPC
ECC
0.003 ECC
HyFRC
0.004 HyFRC

Curvature [1/mm]
Curvature [1/mm]

0.002 0.003

0.002
Lcl = 10
0.001
Lcl = 60
Lcl = 20
Lcl = 240
0.001
Lcl = 410
Lcl = 350

0 0.000
300 400 500 600 700 800 900 1000 1100 300 400 500 600 700 800 900 1000 1100
(a) Distance from Support [mm] Distance from Support [mm]
(b)
750 750
Foundation Level Foundation Level
700 700 UHPC
UHPC
ECC
650 ECC 650
Distance from Base [mm]

HyFRC
600 HyFRC 600 Lcl = 60
Cantilever Beam
Distance from Base [mm]

Lcl = 50
550 550 Lcl = 180
Lcl = 130
500 Lcl = 200
Lcl = 180 500
450 450
400 400
350 350
300 300
250 250
200 200
150 150
0 0.001 0.002 0.003 0 0.001 0.002 0.003
(c) Curvature [1/mm] (d) Curvature [1/mm]

Figure 6: Curvature vs. distance from support in simply supported beams under (a) monotonic and
(b) cyclic loading; curvature vs. distance from base in cantilever under (c) monotonic and (d) cyclic
loading for simulations with Ls = 650 mm & ρ = 1.25% at collapse level

21
 Δ𝑦 Δ𝑝
𝑃

Idealized Elastic
Curvature
Elastic Curvature 𝜃𝑦 𝜃𝑝
𝜙𝑦 𝑧
𝐿𝑠 𝜃𝑦 = 1Τ2 ∗ 𝜙𝑦 ∗ 𝐿𝑠

Ultimate Curvature
𝜙𝑢 𝑧 𝜃𝑝 = (𝜙𝑢 − 𝜙𝑦 ) ∗ 𝐿𝑝
Idealized
𝐿𝑦 𝑑𝑧 Plastic
𝐿𝑝
Curvature
𝑧
𝜙𝑦′ 𝜙𝑝′ 𝜙𝑦 𝜙𝑢 − 𝜙𝑦 𝐿𝑠𝑝
𝜙𝑢′ (d)
(a) (b) (c)

Figure 7: Theoretical basis of equivalent plastic hinge length (a) cantilever beam with lateral load
(b) actual curvature distribution (c) idealized elastic and plastic curvature (d) yield and plastic
rotation or deflection

458 (area below the elastic curvature in Fig. 7 (b)). The plastic rotation can be calculated by integrat-
459 ing the plastic curvature (φp (z) = φu (z) − φy (z)) along the reinforcement yielding zone using Eq. 5
460 (shaded region in Fig. 7 (b)).

Z Ly
θp = [φu (z) − φy (z)]dz (5)
0

461 Where φu (z) and φy (z) are the section curvatures of the beam at the collapse and yield levels.
462 Using the equal area principle, equivalent plastic hinge length can be calculated by dividing the
463 plastic rotation (obtained from Eq. 5) by a constant plastic curvature (i.e., φp = φu − φy ) as shown
464 in Eq. 6.

θp
Lp = (6)
(φu − φy )
465 Where φu and φy are the curvatures at the critical section of the beams at collapse level and yield
466 level, which were calculated using Eq. 7 and 8 based on the “plane section hypothesis”, uniaxial
467 stress-strain relationship, equilibrium of forces and strain compatibility as described in Sec. 6.1 and
468 6.2.

εsu
φu = (7)
d − cu

22
 εsy
φy = (8)
d − cy
469 In Eq. 7 and 8, εsy is the yield strain in tensile longitudinal reinforcement, εsu is the ultimate
470 strain in tensile longitudinal reinforcement, d is the distance from tensile longitudinal reinforcement
471 to the extreme compression fiber, cu and cy are the distances from neutral axis to the extreme
472 compression fiber at collapse level and yield level, respectively. Unlike ordinary concrete which
473 is analyzed based on crushing of the cementitious material, the collapse level curvature, εsu , of
474 HPFRCC members are calculated based on reinforcement fracture strains as this is the common
475 failure mechanism.

476 5.2. Expression for Lp in Reinforced HPFRCC

477 An new expression was developed based on the 72 numerical simulations. The developed expres-
478 sion proposed in this study is based on existing expressions (Table 1) for reinforced concrete and
479 reinforced HPFRCC, and a previous study carried out by Pokhrel and Bandelt [33, 49] to identify
480 variables controlling deformation capacity of reinforced HPFRCC flexural members at collapse. The
481 new equivalent plastic hinge length (Lp ) expression combines four predictor variables: shear-span
482 (Ls ), tensile longitudinal reinforcement ratio (ρ), tensile strength of HPFRCC material (ft ), and
483 yield stress of reinforcement (fy ) as given in Eq. 9.

ρfy
Lp = αLs + β (9)
ft
484 Where α and β are the model coefficients determined through regression. The first term in the
485 expression shows the dependence of plastic hinge length on moment gradient with changing shear-
486 span. The second term shows the dependence of plastic hinge length on (1) tensile strain penetration
487 inside the foundation or joint (Lsp shown in Fig. 7 (c)), and (2) strain concentration and curvature
488 localization within a shear-span due to change in tensile strength of the HPFRCC material and
489 reinforcement ratio. It can be observed from the new plastic hinge length expression that among
490 all the variables associated with the mechanical property of HPFRCCs, tensile strength (ft ) is the
491 primary variable governing the equivalent plastic hinge length in reinforced HPFRCC members and
492 other factors such as tensile strain at onset of softening(εtp ) do not primarily impact the equivalent
493 plastic hinge length.
494 The model coefficients α and β in Eq. 9 were obtained by minimizing the sum of the squares of
495 residuals using regression analysis for four scenarios, considering variation in boundary conditions
496 and loading scheme. Two main sets of equations were developed to account for load variation;
497 one set under monotonic loading scenario and another set under cyclic loading scenario. Eq. 10 was

23
498 formulated using numerical results for one-half of the simply supported beams (Lp,center ) and the full
499 length of cantilever beams (Lp,joint ) under monotonic loading. Eq. 11 was similarly developed under
500 cyclic loading. The numerical simulation results of simply supported models were used to formulate
501 center hinge equations and the simulation results of cantilever models were used to formulate joint
502 hinge equations, which incorporate post-yield strain penetration inside the joint.

Lp,center = 0.02Ls + 0.24 ρfy

ft
Monotonic Loading (10)
Lp,joint = 0.03Ls + 0.27 ρfy

ft

Lp,center = 0.01Ls + 0.11 ρfy

ft
Cyclic Loading (11)
Lp,joint = 0.03Ls + 0.39 ρfy

ft

503 5.3. Comparison with Existing Lp

504 Fig. 8 shows the comparison of numerically simulated plastic hinge length values with the values
505 obtained through the different equations. It can be seen that statistical parameters such as the
506 mean ratio (µRAT IO ) of derived expression-to-numerically simulated Lp value is greater than 1.44
507 (i.e., most of the values lie above and/or left of the mean line). This implies that the existing Lp
508 expressions overestimate the equivalent plastic hinge length, which can result in over-prediction of
509 plastic rotation capacity in reinforced HPFRCC structural members as demonstrated in Sec. 6.3
510 when the proposed expression is validated against a large database of experiments. Further, there
511 is large dispersion in the Lp values predicted by the existing empirical models as indicated by
512 the coefficient of variation values (47% ≤ CoV ≤ 69%). The over-prediction and scatter can be
513 attributed to the following: (1) six of the empirical equations (Fig. 8(a-f)) were developed using
514 reinforced concrete experimental data rather than reinforced HPFRCC specimens; (2) most of the
515 expressions were developed considering only one type of loading scheme (e.g., Fig. 8(a-d) & (g)
516 were developed using monotonically loaded specimens); (3) Lp expression for reinforced HPFRCC
517 members developed by Naaman et al. [39] was based on reinforcement yielding zone which is larger
518 than equivalent plastic hinge length and cannot be used for computation of displacement parameters
519 at collapse level; and (4) Lp expression for reinforced HPFRCC members proposed by Tariq et al.
520 [43] was based on a limited database of six reinforced ECC specimens without variation in HPFRCC
521 type, boundary condition and loading scheme. The new equation developed in this study predicts
522 equivalent plastic hinge length with high accuracy (µRAT IO = 1.0) and significantly less prediction
523 scatter (15% ≤ CoV ≤ 29%) compared to the existing models under both type of loading as shown
524 in Fig. 8 (i). However, it should be noted that the proposed plastic hinge length expression is valid
525 at the collapse level and that the plastic rotation capacity of structural element at intermediate

24
526 damage states (e.g., between yield and collapse), if calculated using this equation, would provide an
527 nonconservative estimate.

528 6. Section Analysis

529 6.1. Chord Rotation and Moment at Yield

530 The analytical model to calculate chord rotation and moment at yield using cross-section analysis
531 is based on the extension of commonly used Euler-Bernoulli beam theory used in the analysis of
532 reinforced concrete members. The assumptions of the analytical formulation are listed below:

533 1. The distribution of strain across the cross-section is assumed to be linear (Euler-Bernoulli
534 hypothesis)
535 2. The strain in the reinforcement and HPFRCC are assumed to be equal at the same layer
536 (perfect bond assumption)
537 3. The internal stresses are computed from the strain through the uniaxial stress-strain relation-
538 ship for both steel and HPFRCC (uniaxial material model)

539 The analytical model with the cross-section parameters required to calculate chord rotation and
540 moment at yield is shown in Fig. 9 (a-c). The internal stress distribution shown in Fig. 9 (c) is
541 based on principal stresses at the integration points across the critical section of the finite element
542 models at yield level. The flowchart shown in Fig. 10 (a) shows the procedure to calculate yield
543 curvature and moment of the member using simplified mechanics. The cross-section and material
544 properties are used as the input parameters for the analytical framework. The section is assumed
545 to have reached its yield capacity when the strain in the tensile reinforcement reaches the yield
546 strain. Using the assumptions stated earlier, cross-sectional forces are computed and the neutral
547 axis depth is iterated until the section equilibrium condition is satisfied. It can be observed in Fig. 9
548 (c), HPFRCC contributes significantly to the tensile component (i.e., Tto + Tt ) of the section forces,
549 which is ignored in the analysis of reinforced concrete. Once the neutral axis depth of the critical
550 section is found, the section yield curvature can be determined using Eq. 8. The yield rotation (θy )
551 of the member can be computed using one of the several elastic-beam deflection procedure such as
552 moment-area method as defined in Eq. 12.

1
θy = φy Ls (12)
2

25
 350 350 350
Cyclic Cyclic Cyclic
300 Monotonic 300 Monotonic 300 Monotonic
1956 [mm]

[mm]

Corley 1966 [mm]

[rad]

Corley 1966 [rad]

250
1956 [rad]

250 250

1964
200

1964
200 200
Lp, Baker

Sawyer
Sawyer
150 150 150
LpBaker

µC,RATIO : 2.75 µC,RATIO : 2.86 µC,RATIO : 2.52

LpLp,
100
CoVC : 65% Lp, 100 100
CoVC : 64%
Lp
CoVC : 69%
50 µM,RATIO : 1.80 50 µM,RATIO : 1.85 50 µM,RATIO : 1.65
CoVM : 54% CoVM : 56% CoVM : 54%
0 0 0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
(a) LpSimulation
Simulation [rad]
(b) (c) LpSimulation
Simulation [rad]
Lp, [mm] Lp, Lp, [mm]
Lp Simulation[rad]
Simulation
[mm]
350

[rad]
350
Cyclic Cyclic 350
µC,RATIO : 4.98
[rad]

[mm]
300 Monotonic 300 Monotonic
CoVC : 67%
[mm]

300
[mm]

20012001
µM,RATIO : 4.82
[rad]

250
1992

250 250

and Fardis
19671967

CoVM : 52%
1992

200
Priestley

200

Panagiotakos
200
Mattock

Paulay

150 150
Mattock

150

Panagiotakos
µC,RATIO : 3.34 µC,RATIO : 5.23
Lp,and

100
CoVC : 64%
LpLp,

100
CoVC : 66% 100
LpPaulay

50 µM,RATIO : 2.17 50 µM,RATIO : 3.40 Cyclic

LpLp,
50
CoVM : 54% CoVM : 49%
0
Monotonic
0 0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
0 50 100 150 200 250 300 350
(d) Lp, (e) Lp,
Lp Simulation [mm]
[rad] (f)
Lp Simulation
Simulation
[mm]
[rad] Simulation Lp, Lp
Simulation [mm]
[rad]
Simulation
350 350 350
Cyclic Cyclic Cyclic
300 Monotonic 300 Monotonic 300 Monotonic
[mm]
[rad]

10 & 11 [rad]
2018[rad]
1996 [mm]

[mm]

250 250 250

New Equation
et al.1996

et al. 2019

200 200 200

Lp, Tariq
Naaman

150 150 150

Lp,Equation

µC,RATIO : 2.18
Naaman

µC,RATIO : 2.83 µC,RATIO : 1.00

LpTariq

100
CoVC : 65% 100 CoVC : 61% 100 CoVC : 29%
Lp,

µM,RATIO : 1.44
Lp

µM,RATIO : 1.92 µM,RATIO : 1.00

Lp

50 50 50
CoVM : 50% CoVM : 47% CoVM : 15%
0 0 0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
(g) Lp
Lp,Simulation
Simulation[rad]
[mm] (h) Lp, Simulation
Lp [mm]
Simulation [rad]
(i) Lp,Lp
Simulation [rad]
Simulation [mm]

Figure 8: Comparison of simulated plastic hinge length with expressions from (a) Baker [35] (b)
Sawyer [36] (c) Corley [37] (d) Mattock [38] (e) Paulay and Priestley [27] (f) Panagiotakos and
Fardis [40] (g) Naaman et al. [39] (h) Tariq et al. [43] (i) current study (Eq. 10 & 11)

26
553 The yield moment capacity (My ) can be computed through moment equilibrium of the cross-
554 section as shown in Eq. 13.

0 0 0 0 1 c 1 2dto (h − c − dto )
My = (fs −Eh εs )As (d−d )+ fc bc(d− )− ft bdto (d−c− )−ft b(h−c−dto )( −ce )
2 3 2 3 2
(13)
0
555 Where fs is the stress in compressive reinforcement, Eh is modulus of elasticity of HPFRCC,
0 0
556 εs is the strain in compressive reinforcement, As is the area of compressive reinforcement, d is the
0
557 effective depth of cross section, d is the distance from top compressive fiber to center of compressive
558 reinforcement, fc is the stress at top compressive fiber, b is the width of cross section, h is the height
559 of cross section, c is the neutral axis depth, ft is the tensile strength of HPFRCC, dto is the distance
560 from neutral axis to the strain layer corresponding to εto , and ce is the distance from bottom tensile
561 fiber to center of tensile reinforcement.
562 To validate the proposed analytical formulation, alternate values of chord rotations were cal-
563 culated using the numerical simulation results, where the strain data at integration points on the
564 tension and compression sides were extracted to compute curvature values along the shear-span of
565 the beams at yield. The chord rotation was calculated by numerically integrating curvature distri-
566 bution along the length of the beam and in the case of cantilever setup, the contribution of pre-yield
567 strain penetration was taken into account by integrating curvature inside the joint as stated in Eq.14.

Z Ls
θy = [φy (z)]dz (14)
0
568 The yield moment was calculated from numerical simulation by extracting the load carrying
569 capacity at yield and using principles of statics (i.e., My,CLB = Py Ls or My,SSB = (Py /2)Ls ).
570 Fig. 11 (a-b) shows the comparison of the yield rotation and moment computed using the analytical
571 framework in comparison to the numerical simulation results. It can be observed that the mean ratio
572 (µRAT IO ) of analytical-to-numerically simulated value of θy is equal to 0.85 under cyclic loading and
573 0.90 under monotonic loading (i.e., most of the values lie below or right of the mean line). This
574 implies that the rotation values are marginally under-predicted using the proposed analytical model
575 which can be attributed to pre-yield strain penetration which has not been incorporated into the
576 analytical model, as Ls has been used to calculate the yield rotation without including the additional
577 length inside the joint. This was done to generalize the analytical formulation based on shear-span
578 length for any location of plastic hinge (i.e., center hinge or joint hinge connected). The yield
579 moment was found to be well predicted using the analytical model with the mean ratio (µRAT IO ) of
580 analytical-to-numerically simulated value of My equal to 0.91 under cyclic loading and 0.93 under
581 monotonic loading, and with minimal prediction scatter (CoV ≤ 3%). The analytical model predicts

27
 𝐴′𝑠 𝜀𝑐 𝑓𝑐
𝜀𝑠′ 𝐶𝑠
𝜙𝑦 𝑐 𝐶𝑐
𝑑 𝜀𝑡𝑜 𝑑𝑡𝑜 𝑇𝑡𝑜
𝑇𝑡
𝜀𝑠𝑦 𝑇𝑦

(a) 𝐴𝑠 (b) 𝜀𝑡 (c) 𝑓𝑡

𝐴′𝑠 𝜀𝑐 𝑓𝑐 0.85𝑓𝑐′
𝜀𝑠′ 𝐶𝑠
𝑐 𝑎 𝐶𝑐
𝜙𝑢
𝑑

𝑓𝑢 𝑇𝑢
𝜀𝑠𝑓 : 0.18
(d) 𝐴𝑠 (e) (f) (g)

Figure 9: Analytical model at yield level (a) cross section (b) strain distribution (c) stress distribution
with section forces; and at collapse level (d) cross section (e) strain distribution (f) stress distribution
(g) equivalent Whitney’s stress block and sectional forces

28
 Start

Start Section and material properties

𝑏, 𝑑 , 𝐴𝑠 , 𝐴′𝑠 , 𝐸𝑠 , 𝑓𝑦 , 𝑓𝑢 , 𝐸ℎ , 𝑓𝑐′
Section and material properties
𝑏, 𝑑, 𝐴𝑠 , 𝐴′𝑠 , 𝐿𝑠 , 𝐸𝑠 , 𝑓𝑦 , 𝐸ℎ , 𝑓𝑡 , 𝜀𝑡𝑜 Tensile reinforcement strain
𝜙𝑢 → 𝜀𝑠𝑓 = 0.18 𝑎𝑛𝑑 𝑀𝑛 → 𝜀𝑠𝑢
Tensile reinforcement yield’s
𝜀𝑠𝑦 = 𝑓𝑦 /𝐸𝑠
Assume neutral axis depth (𝑐)

Assume neutral axis depth (𝑐)

Calculate strain assuming linear strain distribution
𝜀𝑐 , 𝜀𝑠′
Iterate neutral axis depth

Calculate strains assuming linear strain distribution

Iterate neutral axis depth

𝜀𝑐 , 𝜀𝑠′ , 𝜀𝑡𝑜
Calculate stress in steel using uniaxial stress-strain model
𝑓𝑢 , 𝑓𝑠′
Calculate stress using uniaxial stress-strain models
𝑓𝑐 , 𝑓𝑠′ , 𝑓𝑡 Assume Whitney’s stress block for HPFRCC in compression
0.85𝑓𝑐′
Calculate tensile and compressive forces
𝑇𝑦 , 𝑇𝑡𝑜 , 𝑇𝑡 , 𝐶𝑐 , 𝐶𝑠 Calculate tensile and compressive forces
𝑇𝑢 , 𝐶𝑐 , 𝐶𝑠

NO Check equilibrium of forces ?

𝑇𝑦 + 𝑇𝑡𝑜 + 𝑇𝑡 − 𝐶𝑐 − 𝐶𝑠 = 0 NO
Check equilibrium of forces ?
𝑇𝑢 − 𝐶𝑐 − 𝐶𝑠 = 0
YES
Yield Curvature and Moment YES
𝜙𝑦 , 𝑀𝑦 Ultimate Curvature and Nominal Moment
𝜙𝑢 , 𝑀𝑛

End
(a) (b) End

Figure 10: Flowchart to calculate sectional (a) yield curvature and yield moment (b) ultimate
curvature and nominal moment

29
 0.02 50
C-UHPC C-UHPC
C-ECC C-ECC
C-HyFRC C-HyFRC
40
[rad] [rad]

[kNm]
0.015 M-UHPC M-UHPC

[kNm]
M-ECC M-ECC
θy,Analytical

M-HyFRC M-HyFRC

My,Analytical
30

My,Analytical
0.01
ϴy,Analytical

20

µC,RATIO : 0.85 µC,RATIO : 0.91

0.005
CoVC : 25% 10 CoVC : 3%
µM,RATIO : 0.90 µM,RATIO : 0.93
CoVM : 17% CoVM : 2%
0 0
0 0.005 0.01 0.015 0.02 0 10 20 30 40 50
(a) ϴθy,Simulation
y,Simulation [rad]
[rad] (b) My,Simulation
M [kNm]
y,Simulation [kNm]
0.25 50
M-UHPC C-UHPC
M-ECC C-ECC
0.2 M-HyFRC 40 C-HyFRC
[kNm]

C-UHPC M-UHPC
[kNm]
θu,Analytical[rad]
[rad]

C-ECC M-ECC
C-HyFRC
Mn,Analytical

0.15 30 M-HyFRC
ϴu,Analytical

Mn,Analytical

0.1 20
µC,RATIO : 0.94
µC,RATIO : 1.03
CoVC : 11%
0.05 CoVC : 32% 10
µM,RATIO : 0.92
µM,RATIO : 0.97
CoVM : 12%
CoVM : 14%
0 0
0 0.05 0.1 0.15 0.2 0.25 0 10 20 30 40 50
(c) ϴu,Simulation
θu,Simulation[rad]
[rad] (d) MMn,Simulation [kNm]
n,Simulation [kNm]

Figure 11: Comparison of analytical vs. simulated (a) yield rotation (b) yield moment (c) ultimate
rotation and (d) nominal moment

30
582 yield rotation and moment with reasonable accuracy and can be useful to estimate the yield point
583 of a moment-rotation curve (i.e., My , θy ) through hand calculations for reinforced HPFRCC flexural
584 members. It should be noted that the tensile strength of HPFRCC material is not assessed through
585 a standard test procedure and subsequently its value can vary widely based on the type of test
586 specimen or setup used [8, 75]. The plastic hinge length expression and analytical equation of yield
587 moment requires the value of HPFRCC tensile strength as an input parameter, therefore a sensitivity
588 analysis was carried out to evaluate the impact of such variation. It was found that a 10% change in
589 tensile strength resulted in a 6% change in plastic hinge length, and a 2% change in yield moment.

590 6.2. Chord Rotation at Collapse and Nominal Moment

591 The chord rotation at the collapse level will be referred herein as the ultimate rotation (θu ) and
592 the maximum moment capacity of the reinforced HPFRCC beams will be referred to as nominal
593 moment (Mn ). To calculate ultimate rotation and nominal moment, a simplified mechanics-based
594 approach was used with the same underlying assumptions as used in Sec. 6.1 (i.e., plane section
595 hypothesis, strain compatibility, and uniaxial material model ). The analytical model with the cross-
596 section parameters required to calculate ultimate chord rotation is shown in Fig. 9 (d-g). The
597 analytical model to compute nominal moment is similar to Fig. 9 (d-g), except that the controlling
598 strain criteria in the tensile steel is equal to ultimate strain (εs = εsu ). This criteria is based on the
599 experimental study which suggest that at the peak load (corresponding to nominal moment), crack
600 localization and inelastic strain concentration leads to fully hardened tensile reinforcement [19, 25].
601 The internal stress distribution shown in Fig. 9 (f) is based on principal stresses at the integration
602 points across the critical section of the finite element models. The sequential algorithm to calculate
603 ultimate curvature at a critical section and nominal moment capacity of the member is shown in
604 Fig. 10 (b). The input parameters required for the framework includes cross-sectional and material
605 properties. The cross-section is assumed to have reached the collapse level when the strain in tensile
606 reinforcement reaches the fracture strain (εsf ) of 18% and nominal level when the strain in tensile
607 reinforcement reaches the ultimate strain (εsu ). The finite element simulation was able to predict
608 experimental collapse at 18% fracture strain under both monotonic and cyclic loading in Sec. 3.3.
609 Thus, a constant strain of 18% has been adopted in both finite element simulation and the analytical
610 framework to calculate the ultimate chord rotation. The fixed strain criterion provides simplicity and
611 consistency for hand calculation under different loading scheme, similar to the ultimate compressive
612 failure strain of 0.003 used in reinforced concrete to calculate nominal flexural strength [76]. One
613 of the significant differences of the sectional stress distribution in Fig. 9 (f) with that of Fig. 9 (c),
614 is nearly zero contribution of HPFRCC in tensile component of section stresses in Fig. 9 (f) due to
615 the formation of localized flexural cracks at the critical section. The internal parabolic compressive

31
616 stress block at the critical section shown in Fig. 9 (f) is simplified using an equivalent Whitney’s
0
617 stress block, which is defined as a uniformly distributed compressive stress of 0.85fc applied along
618 the depth “a” from extreme compression fiber as shown in Fig. 9 (g). The equivalent Whitney’s
619 stress block has been previously used in computation of nominal flexural strength of steel fiber
620 reinforced concrete [77, 78] with reasonable agreement to physical experimentation result. Using
621 the three fundamental assumptions stated in Sec. 6.1, cross-section forces were computed and the
622 neutral axis depth was iterated until the sectional equilibrium condition is satisfied. The nominal
623 moment (Mn ) can be computed using Eq. 15 through the equilibrium of sectional moment.

0 0 0 0 0 a
Mn = (fs − 0.85fc )As (d − d ) + 0.85fc ba(d − ) (15)
2
0 0
624 Where fs is the stress in compressive reinforcement, As is the area of compressive reinforcement,
0
625 d is the effective depth of cross section, d is the distance from top compressive fiber to center of
0
626 compressive reinforcement, fc is the maximum compressive stress of HPFRCC material, b is the
627 width of cross section, and a = β1 c is the height of the equivalent Whitney’s stress block. The
628 value of β1 depends on the maximum compressive stress of HPFRCC used; with a minimum value
0 0
629 of 0.65 (if fc ≥ 56 MPa), a maximum value of 0.85 (if fc ≤ 28 MPa) and linear variation between
0
630 the two extreme values (if 28 MPa < fc < 56 MPa) [76]. The equation to calculate nominal
631 moment (Mn ) is similar to that of doubly reinforced concrete, but there will be significant numerical
632 difference in Mn due to the higher force associated with the hardened tensile reinforcement (Tu )
633 which substantial increases the sectional moment arm. There are several versions of the analytical
634 model (Fig. 9) proposed by researchers to calculate nominal flexural strength (Mn ) in reinforced
635 steel fiber reinforced concrete (SFRC) [77, 78] and UHPC [79] with variations in internal stress
636 distribution and failure criteria; however, in this study a common formulation of nominal flexural
637 strength is being proposed for broad variants of HPFRCC with a wide range tensile and compressive
638 properties.
639 Ultimate rotation capacity of a reinforced HPFRCC flexural member can be expressed as the
640 sum of yield (θy ) and plastic rotation (θp ) as defined in Eq. 4 [27]. The elastic rotation (θy ) of
641 the member can be computed using Eq. 12 by following the procedure described in Sec. 6.1. After
642 determining the neutral axis depth at the collapse drift, ultimate curvature (φu ) of a critical section
643 can be determined using Eq. 7. Finally, using the new expression of equivalent plastic hinge length
644 (i.e., Eq. 10 & Eq. 11) and section curvatures (φy and φu ) determined earlier, the plastic rotation
645 was calculated by re-arranging Eq. 6.
646 To verify the proposed analytical formulation, values of ultimate chord rotations were calculated
647 based on the numerical simulation results. Strain data at integration points on tension and compres-

32
648 sion sides were extracted to compute curvature values along the shear-span of the beams at collapse
649 level. The chord rotation was calculated by numerically integrating curvature distribution along
650 the shear span of the beam (in the case of cantilever setup, the contribution of post-yield strain
651 penetration was taken into account by integrating curvature inside the joint) at the collapse level
652 using Eq.16.

Z Ls
θu = φu (z)dz (16)
0

653 The nominal moment was calculated from numerical simulation by extracting the maximum load
654 carrying capacity and using the principle of statics (i.e., Mn,CLB = Pn Ls or Mn,SSB = (Pn /2)Ls ).
655 Fig. 11 (c-d) shows the comparison of the ultimate rotation and nominal moment computed using
656 the analytical framework and the numerical simulation technique. It can be observed that the
657 mean ratio (µRAT IO ) of the analytical-to-numerically simulated value of θu is equal to 1.03 under
658 cyclic loading and 0.97 under monotonic loading, indicating a very accurate estimation of ultimate
659 rotation. However, there is slightly higher scatter in the rotation values as indicated by the coefficient
660 of variation values (CoVC = 32% and CoVM = 14%). The nominal moment was found to be
661 well predicted using analytical model with the mean ratio (µRAT IO ) of analytical-to-numerically
662 simulated value of Mn equal to 0.94 under cyclic loading and 0.92 under monotonic loading with
663 significantly less scatter as indicated by the low coefficient of variation (CoV ≤ 12%). This simplified
664 mechanics-based analytical framework can be utilized to estimate maximum moment capacity and
665 collapse level rotation capacity of reinforced HPFRCC flexural members through hand calculations.

666 6.3. Verification with Experimental Database

667 To validate the proposed analytical framework, a database of experimentally tested HPFRCC
668 beams was compiled, whose failure mode was due to fracture of tensile longitudinal reinforcement.
669 The database is composed of 37 cyclically tested HPFRCC beams and 13 monotonically tested
670 HPFRCC beams, with cantilever and simply supported test setups. Due to nature of the evaluation
671 process, experimental data was limited to reinforced ductile concrete specimens which reached their
672 full displacement capacity and failed in flexure due to reinforcement fracture. Validation of the
673 analytical framework was completed using 12 UHPC specimens, 33 ECC specimens and five High-
674 Performance Fiber-Reinforced Concrete (HPFRC) specimens. HPFRC is another class of HPFRCC
675 which generally contains 1.5% by volume fraction of steel fibers and coarse aggregates of maximum
676 1/2” size [80, 81]. The important material and cross-section properties required for the sectional
677 analysis are listed in Table 4. The experimental yield moment (My ), nominal moment (Mn ) and

33
678 ultimate drifts (∆u /Ls ) of beams are presented in the last three columns of the table, which are
679 used to verify the accuracy of the analytical framework.
680 To calculate flexural strength at yield, the section models and flowchart shown in Fig. 9 (a-c) &
681 10 (a) were used along with Eq. 13 as discussed in Sec. 6.1. Similarly, to compute nominal flexural
682 strength, the section models and flowchart shown in Fig. 9 (e-f) & 10 (b) were used along with
683 Eq. 15 as discussed in Sec. 6.2.
684 Fig. 12 (a) shows the comparison of analytically calculated moment with the experimental mo-
685 ment at yield level. The statistical metrics (µC,RAT IO = 0.89 and µM,RAT IO = 0.95) show that the
686 analytical method marginally under-predicts the yield moment values with small dispersion (CoV ≤
687 10%) in the estimated values. The under-prediction of yield moment, even when considering the
688 tensile stress component of HPFRCC, is hypothesized due to localized early strain hardening in
689 the tensile reinforcement (at the critical section) of the reinforced HPFRCC test specimens, before
690 the occurrence of global yielding at the component level [24]. The under-prediction will be more
691 pronounced in sections with multiple layers of tensile reinforcement because the outermost layer
692 of tensile steel yields early and starts to harden before yielding of the inner layer of tensile steel.
693 Fig. 12 (b) shows the comparison of analytically calculated nominal moment with the experimental
694 values. The nominal moment was found to be well predicted using the analytical model with the
695 mean ratio (µRAT IO ) of analytical-to-experimental value of Mn equal to 1.02 under cyclic loading
696 and 1.05 under monotonic loading with less scatter as indicated by the low coefficient of variation
697 (CoV ≤ 15%). The proposed analytical formulation was compared with the standard code-based
698 approach (i.e., ACI Committee 318 [76]) to calculate the moment capacity at both damage levels
699 (i.e., yield and nominal) and it was found to improve the prediction capability by more than 25%.
700 Fig. 13 shows the comparison of analytically computed ultimate rotation capacity (e.g., θu,Baker1956 )
701 with experimentally observed ultimate rotation (θu,Experimental ) of the reinforced HPFRCC database
702 compiled in Table 4. The analytical framework described in Sec. 6.2 along with Eq. 4 was used to
703 calculate the ultimate rotation, using various existing expressions for equivalent plastic hinge length
704 (e.g., θu,Baker1956 was calculated using Lp,Baker ) listed in Table 1. The experimental value of ul-
705 timate rotation (θu,Experimental ) was computed through the ultimate drifts (∆u /Ls ) listed in the
706 last column of Table 4. The mean ratio (µRAT IO ) of analytical-to-experimentally observed θu value
707 is greater than 1.41 for monotonic specimens and 1.38 for cyclic specimens (Fig. 13 (a-h)). This
708 suggests that the use of the existing equations of Lp into the proposed analytical framework tend to
709 overestimate the ultimate rotation capacity in reinforced HPFRCC flexural members. Further, there
710 is significant scatter in the prediction as indicated by the coefficient of variation values (CoV ) ranging
711 from 31% to 51% under both types of loading scenario. The possible reasons of over-prediction and

34
712 scatter can be attributed to the limitation of existing plastic hinge length expressions as described
713 in Sec. 5.3.
714 The proposed analytical framework with the incorporation of new equivalent plastic hinge length
715 expression exhibits better prediction capability (µC,RAT IO = 0.96 and µM,RAT IO = 0.75) and less
716 scatter (CoVC = 22% and CoVM = 29%) as indicated by the statistical measures under both types of
717 loading scheme (Fig. 13 (i)). The improvement in the estimation is due to the use of new Lp equation
718 tailored for reinforced HPFRCC under different loading scenarios, boundary condition, shear span,
719 tensile strength, reinforcement ratio and yield strength. The use of the Lp expression proposed by
720 Tariq et al. [43] provides the least over-prediction under cyclic loading, among the existing equations,
721 because it was developed based on cyclically loaded reinforced ECC test specimens whose failure
722 occurred through fracture of tensile reinforcement, which is the dominant failure mechanism observed
723 in reinforced HPFRCC members. Further, the current database contains 34 cantilever specimens
724 out of a total 37 specimens, which is the same type of test setup used by Tariq et al. [43] to formulate
725 the Lp expression. The use of the new Lp expression proposed in this study can give marginally
726 conservative ultimate rotation capacity of reinforced HPFRCC which is good for practical purposes.

35
 Table 4: Experimental database with yield moment, nominal moment and ultimate drifts

0 0 0
References Specimen HPFRCC Fiber Setup Load Failure Ls Height Width Depth Ls /d E fc ft fy = fy fu ρ ρ fys φv sv My Mn ∆u /Ls
Name Type Type Type Type Mode [mm] [mm] [mm] [mm] [mm/mm] [MPa] [MPa] [MPa] [MPa] [MPa] [%] [%] [MPa] [mm] [mm] [kNm] [kNm] [%]

Fischer and Li [11]? ECC-0.57-0 ECC PE CL C RF 500 100 100 82.0 6.10 18000 80 6.0 410 620 1.73 1.73 315 3 25 6.0 8.0 14.0
ECC-0-0 ECC PE CL C RF 500 100 100 85.0 5.88 18000 80 6.0 410 620 1.67 1.67 316 0 0 6.0 8.3 15.0

Parra-Montesinos† PE2.0-0-0.6 ECC PE DCL C RF 685 250 100 230.0 2.98 10026 56 3.5 440 670 0.62 0.62 - 0 0 22.1 22.8 7.0
and Chompreda [13] PE2.0-0-1.1 ECC PE DCL C RF 685 250 100 230.0 2.98 7992 44 3.5 440 730 1.12 1.12 - 0 0 33.1 34.7 10.0

Olsen and Billington [80] No. 2 P-Bar HPFRC SF CL C RF 914 457 70 416.9 2.19 33794 52 4.0 662 745 0.19 0.19 696 3 76 31.0 33.4 2.7
V-Bar-2 HPFRC SF CL C RF 914 457 70 415.4 2.20 33794 52 4.0 469 634 0.33 0.33 696 3 76 42.1 42.3 3.8
Taper HPFRC SF CL C RF 914 457 70 415.4 2.20 33794 52 4.0 469 634 0.33 0.33 696 3 76 43.9 44.7 2.5

Yuan et al. [82]? S4-ECC-0.69-350 ECC PVA CL C RF 1400 400 200 344.5 4.06 18500 50 5.0 359 542 1.37 1.37 408 8 100 139.6 179.9 6.4
S5-ECC-0.69-525 ECC PVA CL C RF 1400 400 200 344.5 4.06 18500 50 5.0 359 542 1.37 1.37 408 8 100 136.6 175.8 6.0
S6-ECC-1.04-350 ECC PVA CL C RF 1400 400 200 344.5 4.06 18500 50 5.0 359 542 1.37 1.37 408 8 100 135.1 167.6 5.3
†
Tavallali et al. [81] UC4-F HPFRC SF DCL C RF 610 254 406 203.0 3.00 31176 44 4.4 669 807 1.23 1.23 469 10 51 160.3 173.7 12.0
UC2-F HPFRC SF DCL C RF 610 254 406 203.0 3.00 30095 41 4.1 669 807 1.23 1.23 469 10 102 149.4 164.6 11.0

Yoo and Yoon [26]? S13-0.94% UHPC SSF 4P-SSB M RF 900 220 150 179.1 5.02 46733 212 5.8 495 743 0.94 0.53 491 10 80 36.3 39.3 5.8
S13-1.50% UHPC SSF 4P-SSB M RF 900 220 150 177.5 5.07 46733 212 5.8 510 765 1.49 0.54 491 10 80 49.5 55.8 5.7
S19-0.94% UHPC SSF 4P-SSB M RF 900 220 150 179.1 5.02 46881 210 8.5 495 743 0.94 0.53 491 10 80 35.1 42.0 5.6
S19-1.50% UHPC SSF 4P-SSB M RF 900 220 150 177.5 5.07 46881 210 8.5 510 765 1.49 0.54 491 10 80 46.5 56.3 7.3
S30-0.94% UHPC SSF 4P-SSB M RF 900 220 150 179.1 5.02 46773 210 8.0 495 743 0.94 0.53 491 10 80 36.0 43.2 8.9
S30-1.50% UHPC SSF 4P-SSB M RF 900 220 150 177.5 5.07 46773 210 8.0 510 765 1.49 0.54 491 10 80 47.4 56.1 8.1
T30-0.94% UHPC TSF 4P-SSB M RF 900 220 150 179.1 5.02 46972 232 9.5 495 743 0.94 0.53 491 10 80 35.1 43.5 7.3
T30-1.50% UHPC TSF 4P-SSB M RF 900 220 150 177.5 5.07 46972 232 9.5 510 765 1.49 0.54 491 10 80 50.4 60.3 9.1

Frank et al. [71]† ECC-1.30-M ECC PVA CL M RF 760 178 127 149.3 5.09 7000 44 4.0 455 675 1.36 1.36 690 3 75 19.7 23.9 16.0

36
ECC-1.30-MF ECC PVA CL C RF 760 178 127 149.3 5.09 7000 44 4.0 455 675 1.36 1.36 690 3 75 19.8 23.2 11.9

Bandelt and M-0.54 ECC PVA 4P-SSB M RF 457 230 130 209.5 2.18 16000 47 2.2 395 620 0.52 0.52 690 3 75 15.9 17.1 11.0
Billington [19]† C-0.54 ECC PVA 4P-SSB C RF 457 230 130 209.5 2.18 16000 47 2.2 395 620 0.52 0.52 690 3 75 15.8 16.4 3.4
M-0.70 ECC PVA 3P-SSB M RF 685 180 130 159.5 4.29 16000 47 2.2 445 690 0.68 0.68 690 3 75 14.4 17.4 8.4
C-0.70 ECC PVA 3P-SSB C RF 685 180 130 159.5 4.29 16000 47 2.2 445 690 0.68 0.68 690 3 75 14.4 15.5 4.8
M-1.3 ECC PVA 3P-SSB M RF 685 180 130 158.0 4.34 16000 47 2.2 455 675 1.26 1.26 690 3 75 22.2 28.3 12.0
C-1.3 ECC PVA 3P-SSB C RF 685 180 130 158.0 4.34 16000 47 2.2 455 675 1.26 1.26 690 3 75 22.7 25.6 8.5
M-2.0 ECC PVA 3P-SSB M RF 685 180 130 156.5 4.38 16000 47 2.2 440 625 1.96 1.96 690 3 75 30.8 41.1 16.0

Hung and Chueh [46] B-2R-1SF UHPC SF CL C RF 1850 350 250 309.5 5.98 33300 148 6.2 703 907 1.98 1.98 477 13 75 342.3 351.5 8.0
B-2R-2SF UHPC SF CL C RF 1850 350 250 309.5 5.98 34110 152 6.2 703 907 1.98 1.98 477 13 75 342.3 360.8 9.0
B-1.4R-2SF UHPC SF CL C RF 1850 350 250 309.5 5.98 33030 147 6.2 703 907 1.32 1.32 477 13 75 259.0 240.5 7.0
B-1.4R-2LF UHPC SF CL C RF 1850 350 250 309.5 5.98 32220 143 8.0 703 907 1.32 1.32 477 13 75 259.0 249.8 8.0

Frank et al. [25]† ECC-0.73-F ECC PVA CL C RF 760 178 127 150.8 5.04 7000 44 4.0 445 690 0.74 0.74 690 3 75 11.8 13.3 11.9
ECC-0.95-F ECC PVA CL C RF 813 203 165 161.3 5.04 7000 44 4.0 455 675 0.97 0.97 690 3 38 23.6 24.2 11.9
ECC-1.00-F ECC PVA CL C RF 813 203 165 162.8 4.99 7000 44 4.0 445 690 1.06 1.06 690 3 38 27.2 27.6 11.9
ECC-1.30-F ECC PVA CL C RF 760 178 127 149.3 5.09 7000 44 4.0 455 675 1.36 1.36 690 3 75 19.8 23.4 11.9
ECC-1.40-F ECC PVA CL C RF 813 203 165 161.3 5.04 7000 44 4.0 455 675 1.45 1.45 690 3 38 32.9 33.9 16.9
ECC-1.50-F ECC PVA CL C RF 813 203 165 159.8 5.09 7000 44 4.0 440 625 1.51 1.51 690 3 38 31.5 34.7 16.7
 0 0 0
References Specimen HPFRCC Fiber Setup Load Failure Ls Height Width Depth Ls /d E fc ft fy = fy fu ρ ρ fys φv sv My Mn ∆u /Ls
Name Type Type Type Type Mode [mm] [mm] [mm] [mm] [mm/mm] [MPa] [MPa] [MPa] [MPa] [MPa] [%] [%] [MPa] [mm] [mm] [kNm] [kNm] [%]
†
Frank et al. [25] ECC-0.73-SP ECC PVA CL CSP RF 760 178 127 150.8 5.04 7000 44 4.0 445 690 0.74 0.74 690 3 75 12.2 13.4 8.5
ECC-0.95-SP ECC PVA CL CSP RF 813 203 165 161.3 5.04 7000 44 4.0 455 675 0.97 0.97 690 3 38 24.4 24.6 11.9
ECC-1.00-SP ECC PVA CL CSP RF 813 203 165 162.8 4.99 7000 44 4.0 445 690 1.06 1.06 690 3 38 28.5 28.5 11.9
ECC-1.30-SP ECC PVA CL CSP RF 760 178 127 149.3 5.09 7000 44 4.0 455 675 1.36 1.36 690 3 75 20.5 23.7 16.9
ECC-1.40-SP ECC PVA CL CSP RF 813 203 165 161.3 5.04 7000 44 4.0 455 675 1.45 1.45 690 3 38 34.1 34.6 16.9
ECC-1.50-SP ECC PVA CL CSP RF 813 203 165 159.8 5.09 7000 44 4.0 440 625 1.51 1.51 690 3 38 33.7 35.5 16.7
ECC-0.73-LP ECC PVA CL CLP RF 760 178 127 150.8 5.04 7000 44 4.0 445 690 0.74 0.74 690 3 75 13.7 14.1 6.1
ECC-0.95-LP ECC PVA CL CLP RF 813 203 165 161.3 5.04 7000 44 4.0 455 675 0.97 0.97 690 3 38 25.2 26.1 11.9
ECC-1.00-LP ECC PVA CL CLP RF 813 203 165 162.8 4.99 7000 44 4.0 445 690 1.06 1.06 690 3 38 27.6 28.7 11.9
ECC-1.30-LP ECC PVA CL CLP RF 760 178 127 149.3 5.09 7000 44 4.0 455 675 1.36 1.36 690 3 75 20.5 23.3 16.9
ECC-1.50-LP ECC PVA CL CLP RF 813 203 165 159.8 5.09 7000 44 4.0 440 625 1.51 1.51 690 3 38 32.5 36.1 16.7

Note: SF = Steel fibers, SSF = Smooth Steel Fibers, TSF = Twisted Steel Fibers
CL = Cantilever, DCL = Double cantilever, 3P-SSB = 3 Point-Simply supported beam, 4P-SSB = 4 Point-Simply supported beam
C = Cylic load, M = Monotonic load, CSP = Cyclic load with initial small pulse, CLP = Cyclic load with initial large pulse
RF = Reinforcement Fracture, † Yield load computed based on load deflection response in the reference, ? Yield load value provided in the reference

37
 175 200
C-UHPC C-UHPC
C-ECC 175 C-ECC
150
C-HPFRC C-HPFRC

[kNm]
M-UHPC 150 M-UHPC
y,Analytical [kNm]
[kNm]

125

Mn,Analytical [kNm]
M-ECC M-ECC
125
MMy,Analytical

100

Mn,Analytical
100
75
75

50
µC,RATIO : 0.89 µC,RATIO : 1.02
50 CoVC : 14%
CoVC : 9%
25 µM,RATIO : 0.95 25
µM,RATIO : 1.05
CoVM : 10% CoVM : 10%
0 0
0 25 50 75 100 125 150 175 0 25 50 75 100 125 150 175 200
My,Experimental [kNm]
My,Experimental [kNm] (b) MMn,Experimental[kNm]
[kNm]
n,Experimental
(a)

Figure 12: Comparison of analytical vs. experimental (a) yield moment (b) nominal moment

727 7. Conclusions

728 The unique characteristics of reinforced HPFRCC structural member under flexural loading, such
729 as high resistance to splitting cracks, less rebar slippage, damage localization in a single or few flexural
730 cracks, inelastic strain concentration within a small region of rebar, and failure by the fracture of
731 tensile longitudinal reinforcement in the plastic hinge region [26, 19, 25, 32, 49] motivated this
732 research study. A detailed numerical investigation of the plastic hinge region in reinforced HPFRCC
733 flexural members with variations in mechanical properties, geometric properties, and loading scheme
734 was conducted.
735 The strain distribution along the length of HPFRCC flexural members showed that the length
736 of reinforcement yielding zone, Ly , increased with decrease in tensile strength of HPFRCC material,
737 and Ly was longer under cyclic loading compared to monotonic loading in majority of simulations.
738 The length of tensile strain concentration zone, Lsc , was found to be shorter under cyclic loading
739 compared to monotonic loading in simply supported beams, whereas Lsc was found to be marginally
740 longer under cyclic loading compared to monotonic loading in cantilever beams. The investigation of
741 curvature localization zone, Lcl , revealed that under cyclic loading, the length of curvature localiza-
742 tion zone was less compared to monotonic loading in simply supported beams, whereas in cantilever
743 beams the trend of Lcl was found to be reverse (i.e., Lcl,monotonic < Lcl,cyclic ).
744 An improved expression for equivalent plastic hinge length for reinforced HPFRCC flexural mem-
745 bers was proposed, which explicitly takes into account the change in moment gradient, strain pen-
746 etration effect and influence of inelastic strain concentration. The effectiveness of new expression

38
 0.4 0.4 0.4
C-UHPC C-UHPC C-UHPC
C-ECC 0.35 C-ECC C-ECC
0.35 0.35
C-HPFRC C-HPFRC C-HPFRC

[rad]

[rad]
[rad]

[rad]
M-UHPC
[rad]

M-UHPC

[rad]
0.3 M-UHPC 0.3 0.3
M-ECC M-ECC M-ECC

1964
1956

1966
0.25 0.25 0.25

1964
1956

1966
θu, Sawyer
θu, Baker

θu, Corley
0.2 0.2 0.2

ϴu,Sawyer
ϴu,Baker

ϴu,Corley
0.15 µC,RATIO : 1.68 0.15
µC,RATIO : 1.59 0.15 µC,RATIO : 1.59
0.1 CoVC : 47% 0.1 CoVC : 40% 0.1 CoVC : 51%
µM,RATIO : 1.68 µM,RATIO : 1.66 µM,RATIO : 1.56
0.05 0.05 0.05
CoVM : 35% CoVM : 37% CoVM : 33%
0 0 0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
(a) ϴθu,Experimental[rad]
[rad] (b) ϴθu,Experimental [rad] (c) ϴθu,Experimental
u,Experimental [rad] u,Experimental [rad]
u,Experimental [rad]

0.4 0.4

[rad]
[rad]

0.4
C-UHPC
[rad]

0.35 C-ECC 0.35

[rad]
0.35
19921992

C-HPFRC
[rad]

2001
0.3 M-UHPC 0.3
[rad]

2001
0.3
M-ECC µC,RATIO : 2.77
Priestley

µC,RATIO : 2.55

and Fardis
19671967

0.25 0.25
CoVC : 32% 0.25
Priestley

θu, Panagiotakos
CoVC : 31%
µM,RATIO : 3.21
θu, Mattock

0.2 0.2
0.2 µM,RATIO : 4.33
CoVM : 39%
andand
ϴu,Mattock

µC,RATIO : 1.98 0.15 CoVM : 39%

ϴu,Panagiotakos
0.15 0.15
C-UHPC
θu, Paulay

0.1 CoVC : 47% 0.1 C-ECC C-UHPC

ϴu,Paulay

C-HPFRC 0.1 C-ECC

µM,RATIO : 1.99 0.05 M-UHPC C-HPFRC
0.05
CoVM : 35% M-ECC 0.05 M-UHPC
0 M-ECC
0 0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
θu,Experimental [rad]
(d) ϴθu,Experimental [rad]
u,Experimental [rad]
(e) ϴu,Experimental [rad] (f) θu,Experimental [rad]
ϴu,Experimental [rad]
0.4 0.4
C-UHPC 0.4
C-UHPC
C-UHPC
1996 [rad]

0.35 C-ECC 0.35 C-ECC

0.35 C-ECC
[rad]

[rad]
[rad]
2019[rad]

C-HPFRC
[rad]

C-HPFRC C-HPFRC
0.3 M-UHPC 0.3 M-UHPC
0.3 M-UHPC
al. 1996

& 11

M-ECC M-ECC
0.25 M-ECC
0.25
& 11
[Link].2019

0.25
1010
et al.

0.2 0.2
θu,Equation

0.2
Naaman et

ϴu,Equation
ϴu,Naaman

0.15 µC,RATIO : 1.86 0.15 0.15 µC,RATIO : 0.96

u,Tariq

µC,RATIO : 1.38
CoVC : 47% CoVC : 22%
0.1 0.1 CoVC : 40% 0.1
ϴθu,

µM,RATIO : 1.84 µM,RATIO : 0.75

µM,RATIO : 1.41
θu,

0.05 0.05
CoVM : 33% 0.05
CoVM : 29%
CoVM : 34%
0 0 0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

(g) ϴu,Experimental
θu,Experimental
[rad]
[rad] (h) ϴθu,Experimental [rad] (i) ϴθu,Experimental [rad]
u,Experimental [rad]
u,Experimental [rad]

Figure 13: Comparison of analytical vs. experimental ultimate rotation capacity using equivalent
plastic hinge expressions of (a) Baker [35] (b) Sawyer [36] (c) Corley [37] (d) Mattock [38] (e) Paulay
and Priestley [27] (f) Panagiotakos and Fardis [40] (g) Naaman et al. [39] (h) Tariq et al. [43] (i)
Equations 10 & 11

39
747 was evaluated through comparison with existing expressions and in combination with the analytical
748 method to estimate rotation capacity at collapse level.
749 An analytical framework based on section analysis and component level structural analysis was
750 developed to calculate flexural strength and rotational parameters at multiple damage levels for rein-
751 forced HPFRCC flexural members. The proposed framework was validated through an experimental
752 database of 50 reinforced HPFRCC test specimens subjected to monotonic and cyclic loading. The
753 ultimate chord rotation computed using the analytical framework and new plastic hinge length ex-
754 pression showed better prediction capability (µC,RAT IO = 0.96 and µM,RAT IO = 0.75) compared to
755 the use of existing plastic hinge length equations. The flexural strength prediction under both types
756 of loading scenario were found to be well predicted (0.89 ≤ µC,RAT IO ≤ 1.02 and 0.95 ≤ µM,RAT IO ≤
757 1.05) with prediction scatter not higher than 15%.
758 The proposed equivalent plastic hinge length expression for reinforced HPFRCC flexural members
759 can be conveniently used by practicing engineers to model structural components employing “fiber-
760 based fixed length hinge model” in commercial analysis/design softwares such as ETABS, STAAD
761 Pro, SAP2000, Perform 3D, OpenSEES, etc. The developed framework provides an effective tool
762 to determine moment capacity at yield and nominal levels for steel reinforced HPFRCC section,
763 irrespective of the class of HPFRCC used. Further, the analytical model can be effectively used to
764 calculate rotation values of the backbone curve of “concentrated plastic hinge models” such as rigid
765 plastic hinge or non-linear spring hinge of structures constructed with HPFRCC in plastic hinge
766 region.

767 Acknowledgments

768 The authors gratefully acknowledge the support of John A. Reif, Jr., Department of Civil and
769 Environmental Engineering at New Jersey Institute of Technology.

770 References

771 [1] M. Priestley, Performance based seismic design, in: Proceedings of 12th world conference on
772 earthquake engineering, 2831, Auckland, New Zealand, 2000.

773 [2] ASCE 41, Seismic evaluation and retrofit of existing buildings, Technical Report ASCE/SEI
774 41-13, American Society of Civil Engineers, Reston, Virginia, USA, 2014.

775 [3] ATC, Seismic performance assessment of buildings (Volume 1-Methodology), Technical Report
776 FEMA-P-58-1, Federal Emergency Management Agency, Washington, DC, USA, 2012.

40
777 [4] V. C. Li, Can concrete be bendable?, American Scientist 100 (2012) 484–493.

778 [5] J. Moehle, Y. Bozorgnia, N. Jayaram, P. Jones, M. Rahnama, N. Shome, Case studies of the
779 seismic performance of tall buildings designed by alternative means, Technical Report PEER
780 Report 2011/05, Pacific Earthquake Engineering Research Center, Berkeley, California, USA,
781 2011.

782 [6] A. Naaman, H. Reinhardt, High Performance Fiber Reinforced Cement Composites 2:
783 (HPFRCC 2), Proceedings of the Second International Workshop ‘High Performance Fiber
784 Reinforced Cement Composites’, RILEM Proceedings Volume 31, E&FN Spon, Ann Arbor,
785 USA, June 11-14, 1995.

786 [7] V. Li, On engineered cementitious composites (ECC) a review of the material and its applica-
787 tions, Journal of Advanced Concrete Technology 1 (2003) 215–230.

788 [8] K. Kesner, S. Billington, K. Douglas, Cyclic response of highly ductile fiber-reinforced cement-
789 based composites, ACI Materials Journal 100 (2003) 381–390.

790 [9] G. Jen, W. Trono, C. P. Ostertag, Self-consolidating hybrid fiber reinforced concrete : Devel-
791 opment , properties and composite behavior, Construction and Building Materials 104 (2016)
792 63–71.

793 [10] G. J. Parra-Montesinos, High-performance fiber-reinforced cement composites: An alternative

794 for seismic design of structures, ACI Structural Journal 102 (2005) 668–675.

795 [11] G. Fischer, V. Li, Effect of matrix ductility on deformation behavior of steel-reinforced ECC
796 flexural members under reversed cyclic loading conditions, ACI Structural Journal 99 (2002)
797 781–790.

798 [12] B. A. Canbolat, G. J. Parra-Montesinos, J. K. Wight, Experimental study on seismic behavior

799 of high-performance fiber-reinforced cement composite coupling beams, ACI Structural Journal
800 102 (2005) 159–166.

801 [13] G. Parra-Montesinos, P. Chompreda, Deformation capacity and shear strength of fiber-
802 reinforced cement composite flexural members subjected to displacement reversals, Journal
803 of Structural Engineering 133 (2007) 421–431.

804 [14] R. Lequesne, G. Parra-Montesinos, J. Wight, Seismic Behavior and Detailing of High-
805 Performance Fiber-Reinforced Concrete Coupling Beams and Coupled Wall Systems, Journal
806 of Structural Engineering 139 (2013) 1362–1370.

41
807 [15] S. Billington, J. Yoon, Cyclic Response of Unbonded Posttensioned Precast Columns with
808 Ductile Fiber-Reinforced Concrete, ASCE Journal of Bridge Engineering 9 (2004) 353–363.

809 [16] M. Panagiotou, W. Trono, G. Jen, P. Kumar, C. P. Ostertag, Experimental Seismic Response
810 of Hybrid Fiber-Reinforced Concrete Bridge Columns with Novel Longitudinal Reinforcement
811 Detailing, ASCE Journal of Bridge Engineering (2014).

812 [17] K. Kim, G. Parra-Montesinos, Behavior of HPFRCC Low-Rise Walls Subjected to Displacement
813 Reversals, in: A. Naaman, H. Reinhardt (Eds.), High Performance Fiber Reinforced Cement
814 Composites 4 (HPFRCC 4), RILEM Publications S.A.R.L, Cachan Cedex, France, 2004, pp.
815 505–515.

816 [18] G. Parra-Montesinos, S. Peterfreund, S. Chao, Highly Damage-Tolerant Beam-Column Joints

817 Through Use of High-Performance Fiber-Reinforced Cement Composites, ACI Structural Jour-
818 nal 102 (2005) 487–495.

819 [19] M. J. Bandelt, S. L. Billington, Impact of reinforcement ratio and loading type on the defor-
820 mation capacity of high-performance fiber-reinforced cementitious composites reinforced with
821 mild steel, Journal of Structural Engineering 142 (2016) 04016084.

822 [20] T. E. Frank, M. D. Lepech, S. L. Billington, Experimental testing of reinforced concrete and
823 reinforced ECC flexural members subjected to various cyclic deformation histories, Materials
824 and Structures 50 (2017) 232.

825 [21] S. Chao, A. Naaman, G. Parra-Montesinos, Bond behavior of reinforcing bars in tensile strain-
826 hardening fiber-reinforced cement composites, ACI Structural Journal 106 (2009) 897–906.

827 [22] M. J. Bandelt, S. L. Billington, Bond behavior of steel reinforcement in high-performance

828 fiber-reinforced cementitious composite flexural members, Materials and Structures 49 (2016)
829 71–86.

830 [23] M. J. Bandelt, T. E. Frank, M. D. Lepech, S. L. Billington, Bond behavior and interface
831 modeling of reinforced high-performance fiber-reinforced cementitious composites, Cement and
832 Concrete Composites 83 (2017) 188–201.

833 [24] D. M. Moreno, W. Trono, G. Jen, C. Ostertag, S. L. Billington, Tension stiffening in reinforced
834 high performance fiber reinforced cement-based composites, Cement and Concrete Composites
835 50 (2014) 36–46.

42
836 [25] T. E. Frank, M. D. Lepech, S. L. Billington, Experimental testing of reinforced ecc beams
837 subjected to various cyclic deformation histories, Journal of Structural Engineering 144 (2018)
838 04018052.

839 [26] D.-Y. Yoo, Y.-S. Yoon, Structural performance of ultra-high-performance concrete beams with
840 different steel fibers, Engineering Structures 102 (2015) 409–423.

841 [27] T. Paulay, M. Priestley, Seismic Design of Reinforced Concrete and Masonry Buildings, John
842 Wiley, New York, 1992.

843 [28] İ. Kazaz, Analytical study on plastic hinge length of structural walls, Journal of Structural
844 Engineering 139 (2012) 1938–1950.

845 [29] P. H. Feenstra, J. Rots, A. Arnesen, J. Teigen, K. Hoiseth, A 3D constitutive model for concrete
846 based on a co-rotational concept, in: D. Borst, Bicanic, Mang, Meschke (Eds.), Computational
847 Modelling of Concrete Structures, Proceedings of EURO-C 1998, Balkema, Rotterdam, 1998,
848 pp. 13–22.

849 [30] T. Han, P. Feenstra, S. Billington, Simulation of highly ductile fiber-reinforced cement-based
850 composite components under cyclic loading, ACI Structural Journal 100 (2003) 749–757.

851 [31] C. Hung, S. El-Tawil, Hybrid Rotating/Fixed-Crack Model for High-Performance Fiber-
852 Reinforced Cementitious Composites, ACI Materials Journal 107 (2010) 568–576.

853 [32] M. J. Bandelt, S. L. Billington, Simulation of deformation capacity in reinforced high-

854 performance fiber-reinforced cementitious composite flexural members, Journal of Structural
855 Engineering 144 (2018) 04018188.

856 [33] M. Pokhrel, M. J. Bandelt, Simulation of reinforced HPFRCC deformation capacity under
857 flexure- and shear-dominated stress states, in: G. Meschke, B. Pichler, J. G. Rots (Eds.),
858 Proceedings of Computational Modelling of Concrete and Concrete Structures (Euro-C 2018),
859 CRC Press, Bad Hofgastein, Austria, 2018, pp. 633–640.

860 [34] E. M. Hines, J. I. Restrepo, F. Seible, Force-displacement characterization of well-confined

861 bridge piers, ACI Structural Journal 101 (2004) 537–548.

862 [35] A. L. L. Baker, The ultimate load theory applied to the design of reinforced & prestressed
863 concrete frames, Concrete Publications Ltd., London, 1956.

864 [36] H. A. Sawyer, design of concrete frames for two failure stages, ACI Structural Journal (1964)
865 405–437.

43
866 [37] W. Corley, Rotational capacity of reinforced concrete beams, Journal of the Structural Division
867 92 (1966) 121–146.

868 [38] A. H. Mattock, Discussion of rotational capacity of reinforced concrete beams, Journal of the
869 Structural Division 93 (1967) 519–522.

870 [39] A. Naaman, P. Paramasivam, G. Balazs, Z. Bayasi, J. Eibl, L. Erdelyi, N. Hassoun,

871 N. Krstulovic-Opara, V. Li, G. Lohrmann, Reinforced and prestressed concrete using hpfrcc
872 matrices, in: A. Naaman, H. Reinhardt (Eds.), High Performance Fiber Reinforced Cement
873 Composites 2 (HPFRCC 2), Proceedings of the Second International RILEM Workshop, E&FN
874 Spon, London, UK, 1996, pp. 291–347.

875 [40] T. B. Panagiotakos, M. N. Fardis, Deformations of reinforced concrete members at yielding

876 and ultimate, ACI Structural Journal 98 (2001) 135–148.

877 [41] S. Bae, O. Bayrak, Plastic hinge length of reinforced concrete columns, ACI Structural Journal
878 105 (2008) 290–300.

879 [42] M. P. Berry, D. E. Lehman, L. N. Lowes, Lumped-plasticity models for performance simulation
880 of bridge columns, ACI Structural Journal 105 (2008) 270.

881 [43] H. Tariq, E. A. Jampole, M. J. Bandelt, Fiber-hinge modeling of engineered cementitious

882 composite flexural members under large deformations, Engineering Structures 182 (2019) 62–
883 78.

884 [44] L. Xu, J. Pan, C. Lu, W. Yin, Development mechanism of plastic hinge in reinforced engineered
885 cementitious composite beams under monotonic loading, Structural Concrete (2018) 1–15.

886 [45] Y. Shao, S. L. Billington, Predicting the two predominant flexural failure paths of longitudi-
887 nally reinforced high-performance fiber-reinforced cementitious composite structural members,
888 Engineering Structures 199 (2019) 109581.

889 [46] C.-C. Hung, C.-Y. Chueh, Cyclic behavior of uhpfrc flexural members reinforced with high-
890 strength steel rebar, Engineering Structures 122 (2016) 108–120.

891 [47] J. A. Blume, N. M. Newmark, L. H. Corning, Design of multistory reinforced concrete buildings
892 for earthquake motions, volume 4, Portland Cement Association Chicago, 1961.

893 [48] J. K. Wight, Reinforced Concrete: Mechanics and Design, Pearson, Hoboken, New Jersey, 2015.

44
894 [49] M. Pokhrel, M. J. Bandelt, Material properties and structural characteristics influencing defor-
895 mation capacity and plasticity in reinforced ductile cement-based composite structural compo-
896 nents, Composite Structures 224 (2019) 111013.

897 [50] DIANAFEA, Diana release 10.2, [Link] 2017.

898 [51] A. Hemmati, A. Kheyroddin, M. K. Sharbatdar, Plastic hinge rotation capacity of reinforced
899 hpfrcc beams, Journal of Structural Engineering 141 (2013) 04014111.

900 [52] H. G. Russell, B. A. Graybeal, Ultra-high performance concrete: A state-of-the-art report for
901 the bridge community, Technical Report PUBLICATION NO. FHWA-HRT-13-060, Federal
902 Highway Administration, McLean, Virginia, USA, 2013.

903 [53] B. A. Graybeal, Material property characterization of ultra-high performance concrete, Tech-
904 nical Report PUBLICATION NO. FHWA-HRT-06-103, Federal Highway Administration,
905 McLean, Virginia, USA, 2006.

906 [54] V. Li, C. Leung, Steady-state and multiple cracking of short random fiber composites, Journal
907 of Engineering Mechanics 118 (1992) 2246–2264.

908 [55] V. C. Li, S. Wang, C. Wu, Tensile strain-hardening behavior of polyvinyl alcohol engineered
909 cementitious composite (pva-ecc), ACI Materials Journal 98 (2001) 483–492.

910 [56] V. C. Li, Engineered Cementitious Composites (ECC) - Material, Structural, and Durability
911 Performance, CRC press, 2008.

912 [57] P. Kumar, G. Jen, W. Trono, D. Lallemant, M. Panagiotou, C. Ostertag, Self compacting
913 hybrid fiber reinforced concrete composites for bridge columns, Technical Report PEER Report
914 2011/106, Pacific Earthquake Engineering Research Center, University of California, Berkeley,
915 California, USA, 2011.

916 [58] D. Moreno-Luna, Tension stiffening in reinforced high performance fiber reinforced cement based
917 composites, Ph.d., Stanford University, 2014.

918 [59] J. Blunt, C. Ostertag, Performance-based approach for the design of a deflection hardened
919 hybrid fiber-reinforced concrete, Journal of engineering mechanics 135 (2009) 978–986.

920 [60] K. Wille, A. Naaman, Fracture energy of uhpfrc under direct tensile loading, in: e. a. B.H. Oh
921 (Ed.), Fracture Mechanics of Concrete and Concrete Structures, Proceedings of FraMCoS-7,
922 Jeju, Korea, 2010, pp. 65–72.

45
923 [61] FEMA, Interim Testing Protocols for Determining the Seismic Performance Characteristics of
924 Structural and Nonstructural Components, Technical Report FEMA 461, Federal Emergency
925 Management Agency, Washington, DC, USA, 2007.

926 [62] N. Mitra, L. N. Lowes, Factors influencing analytical continuum simulation of three-point bend
927 test of a concrete notched beam, in: Proceedings of Fourteenth World Conference of Earthquake
928 Engineering, Beijing, China, 2008.

929 [63] J. G. Rots, J. Blaauwendraad, Crack models for concrete, discrete or smeared ? Fixed, multi-
930 directional or rotating ?, Technical Report, 1989.

931 [64] P. H. Feenstra, Computational Aspects of Biaxial Stress in Plain and Reinforced Concrete,
932 Ph.D. thesis, Delft University of Technology, 1993.

933 [65] S. L. Billington, Y. Shao, T. E. Frank, M. J. Bandelt, D. M. Moreno, Simulation of reinforced

934 ductile cement-based composite beams under cyclic loads, in: G. Meschke, B. Pichler, J. G.
935 Rots (Eds.), Proceedings of Computational Modelling of Concrete and Concrete Structures
936 (Euro-C 2018), CRC Press, Bad Hofgastein, Austria, 2018, pp. 805–812.

937 [66] W.-P. Kwan, S. Billington, Simulation of structural concrete under cyclic load, Journal of
938 Structural Engineering 127 (2001) 1391–1401.

939 [67] R. V. Mises, Mechanics of solid bodies in plastically deformable state, News from the Society
940 of Sciences on Gottingen, Mathematical-Physical Class 1913 (1913) 582–592.

941 [68] C. Shen, E. Mizuno, T. Usami, A generalized two-surface model for structural steels under
942 cyclic loading, Doboku Gakkai Ronbunshu 10 (1993) 23–33.

943 [69] T. E. Frank, Response of reinforced engineered cementitious composite flexural members sub-
944 jected to various cyclic deformation histories, Ph.d., Stanford University, 2017.

945 [70] M.-A. Dagenais, B. Massicotte, Cyclic behavior of lap splices strengthened with ultra high
946 performance fiber-reinforced concrete, Journal of Structural Engineering 143 (2016) 04016163.

947 [71] T. Frank, M. Lepech, S. Billington, Effect of deformation history on steel-reinforced hpfrcc flex-
948 ural member behavior, in: Proceedings of the 7th International Workshop on High Performance
949 Fiber Reinforced Cement Composites (HPFRCC-7), volume 1, Stuttgart, Germany, 2015.

950 [72] X.-M. Zhao, Y.-F. Wu, A. Leung, Analyses of plastic hinge regions in reinforced concrete beams
951 under monotonic loading, Engineering Structures 34 (2012) 466–482.

46
 952 [73] F. Yuan, Y.-F. Wu, Effect of load cycling on plastic hinge length in rc columns, Engineering
953 Structures 147 (2017) 90–102.

954 [74] M. L. Marsh, I. G. Buckle, E. Kavazanjian Jr, LRFD Seismic Analysis and Design of Bridges
955 Reference Manual, Technical Report PUBLICATION NO. FHWA-NHI-15-004, National High-
956 way Institute, Washington, DC, USA, 2014.

957 [75] B. A. Graybeal, F. Baby, Development of direct tension test method for ultra-high-performance
958 fiber-reinforced concrete., ACI Materials Journal 110 (2013).

959 [76] ACI Committee 318, Building code requirements for structural concrete (ACI 318-14) and com-
960 mentary, Technical Report ACI 318-14, American Concrete Institute, Farmington Hills, Michi-
961 gan, USA, 2014.

962 [77] R. Swamy, A. Saad, Deformation and ultimate strength in flexure of reinforced concrete beams
963 made with steel fiber concrete 78 (1981) 395–405.

964 [78] ACI Committee 544, Design Considerations for Steel Fiber Reinforced Concrete, Technical Re-
965 port ACI 544.4R-88, American Concrete Institute, Farmington Hills, Michigan, USA, 1988.

966 [79] S. Aaleti, B. Petersen, S. Sritharan, Design Guide for Precast UHPC Waffle Deck Panel Sys-
967 tem, including Connections, Technical Report PUBLICATION NO. FHWA-HIF-13-032, Federal
968 Highway Administration, Washington, DC, USA, 2013.

969 [80] E. Olsen, S. Billington, Cyclic Response of Precast High-Performance Fiber-Reinforced Con-
970 crete Infill Panels, ACI Structural Journal 108 (2011) 51–60.

971 [81] H. Tavallali, A. Lepage, J. M. Rautenberg, S. Pujol, Concrete beams reinforced with high-
972 strength steel subjected to displacement reversals., ACI Structural Journal 111 (2014) 1037–
973 1048.

974 [82] F. Yuan, J. Pan, Z. Xu, C. Leung, A comparison of engineered cementitious composites versus
975 normal concrete in beam-column joints under reversed cyclic loading, Materials and structures
976 46 (2013) 145–159.

47

View publication stats