Construction and Building Materials 175 (2018) 519–530

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Characterizing the pulse velocity and electrical resistivity changes in

concrete with piezoresisitive smart cement binder using Vipulanandan
models
C. Vipulanandan ⇑, N. Amani
Center for Innovative Grouting Materials and Technology (CIGMAT), Department of Civil and Environmental Engineering, University of Houston, Houston, TX 77204-4003,
United States

h i g h l i g h t s

New concrete with smart cement binder is a piezoresistive bulk sensor.

Electrical resistivity is the monitoring parameter.

Resistivity changes during curing were higher than the pulse velocity changes.

Vipulanandan composite, curing and piezoresistive models predicted the behavior.

a r t i c l e i n f o a b s t r a c t

Article history: In this study the behavior of concrete made using piezoresisitive smart cement as the binder was inves-
Received 18 January 2018 tigated to test and model the concrete as a bulk sensing material. The coarse aggregate content in the
Received in revised form 4 April 2018 concrete was varied up to 75% (by volume). The concrete property changes during curing were tested
Accepted 24 April 2018
using the ultrasonic pulse velocity and compared it to the electrical resistivity, since it can be easily
adopted for real-time monitoring. The initial (immediately after mixing) compressive pulse velocity for
the piezoresistive smart cement (binder only) was 1050 m/s and it increased to 1490 m/s with the addi-
Keywords:
tion of 75%, a 42% increase in the initial pulse velocity. After 28 days of curing, the pulse velocity of smart
Concrete
Smart cement
cement increased from 3520 m/s to 4750 m/s with the addition of 75% aggregate, a 35% increase in the
Ultrasonic pulse velocity pulse velocity. Also the addition of coarse aggregates increased the initial electrical resistivity of the
Electrical resistivity smart cement composite as well as the long term electrical resistivity. The initial electrical resistivity
Piezoresistivity of smart cement was 1.02 Om which increased nonlinearly to 3.74 Om. with the addition of 75% aggre-
Compressive strength gate, a 267% increase in the initial electrical resistivity. After 28 days of curing, the electrical resistivity of
Modeling smart cement was 14.14 Om and with 75% aggregate it increased to 61.24 Om, a 333% increase in the
electrical resistivity. Applicability of the mixture theory to predict the resistivity of the concrete from
the constituents was verified and a new composite resistivity model was developed. Also
Vipulanandan p-q curing model was used to predict the resistivity changes in the concrete with the cur-
ing time. The piezoresistivity of the smart cement without and with 75% aggregate after 28 days of curing
were 204% and 101% at the peak compressive stresses of 21.7 MPa and 12.4 MPa respectively. The reduc-
tion in the piezoresistivity at peak compressive stress was due to not only the reduction of smart cement
content in the composite but also the strength. Compared to the compressive failure strain of 0.3%, the
resistivity change for the concrete with 75% gravel after 28 days of curing was over 336 times
(33,600%) higher making the concrete with the smart cement binder a highly sensing bulk material.
The composite behaviors (with curing time and applied stress) were modeled using the Vipulanandan
pulse velocity, concrete resistivity and piezoresistivity models and compared with the current models
used in the literature. Based on the coefficient of determination (R2) and root mean square error
(RMSE), Vipulanandan models predicted the experimental results very well.
Ó 2018 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.
E-mail address: cvipulanandan@uh.edu (C. Vipulanandan).

https://doi.org/10.1016/j.conbuildmat.2018.04.196
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
520 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530
1. Introduction
Cement composites are one of the durable construction materi-
als mainly composed of cement, aggregates, water and additives
based on the applications. Concrete with high aggregate content
in the cement composite can be used in the construction of differ-
ent structures such as roads, houses, bridges, pipes, dams, canals,
storage, missile silos and nuclear waste containment. To attain
the required levels of safety and durability of such structures, mix-
ing proportions and especially aggregate content must be adjusted
according to application in order to achieve mechanical require-
ments which will significantly affect the performance during its
life time [1]. In preparing the concrete and cement slurries, the
water-to-cement ratios have been varied from 0.38 to 0.6 based
on the mixing method, constituents of the concrete mix and appli-
cations [2–5]. There are many different testing techniques such as
ultrasound, fiber optic, electronic microscopy, X-ray diffraction,
thermography and vibro-thermography have been used to study
the aging of cement composites and for damage detection [6]. Say-
ers and Grenfell [7] used the ultrasonic method to characterize the
early age hydration of the cement and evaluated the Poisson’s ratio
of the slurries which decreased from the initial value of 0.5 for the
fluid to values characteristic of a porous solid which was in the
range of 0.25. Studies showed that using ultrasonic method with
longitudinal wave, the location and size of the delamination in
the slab can be evaluated [8]. Zeng et al. [9] used this method to
characterize the development of bond slip in the concrete-
encased composite. However, this method is difficult to adopt in
some field conditions where accessibility becomes an issue in deep
foundations, wells, dams, canals and pipes.
Research studies on the electrical resistivity of the cement com-
posites over the past two decades have shown that out of all the dif-
ferent methods of characterizing the cement composites, electrical
resistivity is one of the highly sensitive and economical nondestruc-
tive method to monitor the serviceability of the cement composites
throughout the entire service life [10]. Several studies have proved
the sensibility of this method on monitoring any chemical or phys-
ical changes in concrete during its life time (Table 1). Ramezanian-
pour et al. [11] suggested surface resistivity as an indicator of
concrete chloride penetration resistance for a wide range of concrete
compositions. Azhari and Banthia [12] worked on conductive
cement-based materials which are also piezoresistive and proved
that the electrical resistivity responses of such cement-based sen-
sors with applied load are nonlinear and rate-dependent. Vipu-
lanandan et al. [13] characterized modified cementitious and
polymer composites using electrical resistivity measurements. Also
newly developed smart cement studies have shown up to 400%
increase in piezoresistivity which is defined as the changes in the
electrical resistivity of the materials at peak stress [3–5,14]. Electri-
cal resistivity can be a sensitive method for characterization of the
concrete which can be affected by the curing conditions, concrete
composition and cement type [15–19]. Chu and Chen [20] provided
a correlation between real-time damage and resistivity of concrete
under the condition of a static load (Table 1). They also showed
the exponential growth of concrete resistivity and residual damage
under the condition of a cyclic load. As previously mentioned, aggre-
gates play a vital role in concrete strength, and the effects on electri-
cal resistivity should be considered while employing concrete
resistivity for various monitoring purposes. Aggregate content and
size and type of aggregates are parameters which have been docu-
mented to affect the electrical resistivity of the concrete [21,22].
Hou et al. [1] studied the effect of up to 30% Coarse aggregates based
on the total weight on electrical resistivity of the concrete using AC
measurements of 1 Hz frequency (Table 1). These studies have
shown that adding coarse aggregates increased the bulk resistivity.
2. Objective
The overall objective of this study was to compare the changes
in compressive pulse velocity and electrical resistivity with curing
time and the piezoresistive behavior of concrete with up to 75%
aggregate (by volume) and smart cement binder. The specific
objectives are as follows:
1) Effect of adding aggregates on the initial and long-term
pulse velocity, electrical resistivity and compressive
strength of smart cement composites.
2) Investigate the effect of adding up to 75% gravel (based on
the total volume of cement composite) on the smart cement
composite curing and piezoresistive behavior up to 28 days.
3) Modeling the behavior of concrete with smart cement
binder.
3. Materials and methods
3.1. Sample preparation
In this study table top blenders were used to prepare the
cement and concrete specimens. Cement (Portland cement Type
1 (ASTM C 150)) specimens were prepared using smart cement
with water-cement ratio of 0.38 to minimize the bleeding and
blended for at least 10 min [3]. Concrete specimens were prepared
using 25%, 50% and 75% coarse aggregates based on the total vol-
ume of cement composite. Sieve analysis (ASTM C136) was done
to determine the gradation of aggregate and the gradation is
shown in Fig. 1. The median diameter [27], which also represents
d50 (ASTM) the size of 50% of the particles less than 4.2 mm [28].
After mixing, the cement and concrete were placed in 100 mm
height and 50 mm diameter cylindrical molds with two conductive
flexible wires 1 mm in diameter (representing the probes) were
placed 50 mm apart vertically to measure the electrical resistance.
The specimens were cured up to 28 days under relative humidity
of 90%. At least three specimens were test under each condition
and the average values are presented in the figures, tables and
discussion.
3.2. Density
In order to determine the density of smart cement and concrete
composites with different percentages of aggregate contents mass
and volume were measured immediately after placing the mixers
in the molds. The dimensions of the specimen were measured
using a Vernier Caliper.
3.3. Pulse velocity (ASTM C597)
The ultrasonic pulse velocity measurements were made by lon-
gitudinal waves using 150 kHz transducers. Smart cement compos-
ite mix was place in a cylindrical mold and the ultrasonic
compression pulse produced by the transmitter was passed
through the specimen and the arrived time was measured accu-
rately up to 0.01 ms, using a commercial available device and a
transducer as a receiver which was embedded on the other side
of the mold.
3.4. Electrical resistivity
Two different devices were used to measure the changes in the
electrical resistivity of the smart cement and concrete immediately
after mixing up to the time they solidify.
 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530 521

Table 1
Summary of electrical studies on cement composites.

Reference Objective Additives Methods of Remarks

Saleem et al. [23]
Shi [24]
Vipulanandan et al. [5]
Gao et al. [25]
Ramezanianpour et al. [11]
Azhari et al. [12]
Chang et al. [26]
Moreno et al. [18]
Sengul [22]
Chu and Chen [20]
Hou et al. [1]
Remarks
Effect of moisture and Crushed limestone and
different dune sand
contaminations
Effect of mixing Silica fume, coal fly ash
proportions and blast furnace slag
Mechanical properties Fine aggregates with
were correlated to the 6% carbon fibers
electrical resistivity
Piezoresistive sensing Fine and coarse
under static loading aggregates, sodium
dodecyl sulfate,
antifoam and up to
2.5% (vol) carbon
nanofiber
Correlation of Metakaolin, Tuff,
compressive strength Pumice, rice husk ash,
and permeability of silica fume and fine
concrete with electrical and coarse aggregate
resistivity
Piezoresistive sensing Silica fume,
under dynamic and superplasticizer,
static loading methylcellulose,
defoamer, 15% carbon
fiber (and 1% multi-
walled carbon
nanotubes (by volume)
Effect of temperature Coarse aggregate, sand,
and mixing on superplasticizer and up
electrical resistivity to 0.75% (by mass of
cement) carbon fiber
Effect of curing time Fly ash, silica fume, fine
and regime and coarse aggregate
Effect of aggregate size Fly ash, blast furnace
and type, slag, silica fume,
environmental crushed limestone
condition sand and coarse
aggregates and
superplasticizer
Damage detection Sand and gravel,
carbon fiber up to 1.5%
(by volume)
Effect of coarse Sand, granite and
aggregates(size, type marble
and quantity)
Piezoresistive, Use of carbon fibers in
mechanical and some studies
chemical studies of
cement composites
using electrical
methods
Measurement
Four-probe method 97
Hz Frequency
ASTM C1202
Two probe method.
Percolation theory was
used to determine the
critical fiber content.
Four probe method
Four probe method
(surface resistivity)
Four probe method AC
Electrical
measurement at 100
kHz frequency
Four probe method DC
Electrical
measurement
Four probe method
(surface resistivity)
Two-probe method
with 1 Hz Frequency
and ASTM C 1760
Four probe method
Two-probe method1
Hz Frequency
Mostly four-probe
method and DC
measurements
1) Four-probe method
2) AC measurements with low frequency
3) No piezoresistivity
4) No carbon fiber content
1) Method of measurements not
mentioned
2) DC measurements
3) No piezoresistivity
4) No carbon fiber content
1) Two probe Method
2) DC measurements
3) No Piezoresistivity.
4) 1% to 6% carbon fiber content and crit-
ical fiber content was 0.7%
1) Four-probe method
2) AC/DC measurements not mentioned
3) Varying piezoresistivity with different
carbon nanofiber content from 25%
to +10%
4) Carbon nanofiber content up to 2.5%
1) Four-probe method
2) AC/DC measurements not mentioned;
Surface measurement
3) No piezoresistivity
4) No carbon fiber content
1) Four-probe method
2) AC measurements at 100 kHz
3) Negative piezoresistivity with maxi-
mum change of 50%
4) Carbon fiber content of 15% and car-
bon nanofiber content of 1%
1) Four-probe method
2) DC measurements
3) No piezoresistivity
4) Carbon fiber content up to 0.75%
1) Four-probe method
2) AC/DC measurements not mentioned;
Surface measurement
3) No piezoresistivity
4) No carbon fiber content
1) Two-probe method
2) AC measurements with low frequency
3) No piezoresistivity
4) No carbon fiber content
1) Four-probe method
2) AC/DC measurements not mentioned
3) Negative piezoresistivity
4) Carbon fiber content up to 1.5%
1) Two-probe method
2) AC measurements with 1 Hz
frequency
3) No piezoresistivity
4) No carbon fiber content
1) Mostly four-probe method
2) Mostly DC measurements or AC with
low frequency
3) Mainly negative piezoresistivity
4) Carbon fiber content varied from
0.75% to 15%
522 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530

3.6. Compression test (ASTM C39)

The cylindrical specimens were capped and tested at a predeter-

mined controlled displacement rate. Tests were performed in the
Tinious Olsun machine at a controlling the displacement rate to
0.125 mm per minute. In order to measure the strain, a commer-
cially available extensometer (accuracy of 0.001% strain) was used.
During the compression test, the change in resistance was measured
continuously using the LCR meter. Two probe method with alterna-
tive current (AC) at 300 kHz frequency was used in order to mini-
mize the contact resistances [14]. The change in resistance was
monitored using the two-probe method, and the parameter in Eq.
(2) was used relate the changes in resistivity to the applied stress.
Fig. 1. Particle size distribution of the coarse aggregate.

3.7. Modeling
3.4.1. Conductivity probe
A commercially available conductivity meter was used to mea- 3.7.1. Vipulanandan pulse velocity-curing time model
sure the conductivity (inverse of electrical resistivity). The conduc- In order to represent the development of the pulse velocity of
tivity measuring range was from 0.1 lS/cm to 1000 mS/cm, the hardening cement and concrete with curing time, the newly
representing a resistivity of 100,000 Xm to 0.01 Xm respectively. developed Vipulanandan Pulse Velocity-Curing Time Model was
used and the relationship is as follows:
2 3
3.4.2. Digital resistivity meter
The digital resistivity meter measured the resistivities in the 6 t
7
V p  V p0 ¼ ðV pc  V p0 Þ6
4
tc 7
 ðp pq3 Þ 5 ð3Þ
range of 0.01–400 O-m. Both of the electrical resistivity devices
were calibrated using the standard solutions of sodium chloride q3 þ ð1  p3  q3 Þ tc þ p3 tc
t t 3 3

(NaCl).
where V p0 is the initial pulse velocity of the cement composite slur-
ries, V pc is the ultimate pulse velocity of the cement composites, tc is
3.5. Electrical resistance
the curing time of the smart cement composites which was 28 days
in this study, and the material parameters p3 and q3 will be influ-
LCR meter (inductance (L), capacitance (C), and resistance (R))
enced by the material constituents, density and other properties.
was used to monitor the electrical resistance of the specimens dur-
ing the curing time. Two wire method with AC at 300 kHz fre-
3.7.2. Mixture theory
quency was used in order to minimize the contact resistances
The simplest of the mixing rules in a composite is the series and
[29]. During the initial stage of curing both the electrical resistivity
parallel addition of the resistances representing the gravel and
(q) electrical resistance (R) were measured to determine the
cement binder [30]. The series equation describes a stack of the
parameter K based on the Eq. (1).
resistances of the two constituents. The electrical resistivity of


L the composite (q) can be calculated as follow:
R¼q ¼ qK ð1Þ


A VS VG
q ¼ qS þ qG ð4Þ
where L is the distance between two points where resistance is VG þ VS VG þ VS
measured, A is the cross-sectional area through which the current The parallel equation describes a stack of the two resistive
is flowing. The ratio L/A is called the geometry factor (nominal materials in parallel. Hence, the electrical resistivity of the com-
Kn) and used for conductive materials. In this study, electrical resis- posite (q) can be calculated as follow:
tance (R) and electrical resistivity (q) were measured independently


during the initial curing period and the effective calibration factor 1 1 VS 1 VG
¼ þ ð5Þ
(Ke) for the materials used in this study (insulators) were deter- q qS VG þ VS qG VG þ VS
mined experimentally. For the smart cement and concrete Parame- where V G is the aggregate volume content, V S is the cement slurry
ter Ke became stable (constant) in two to three hours. The volume content, qG is the electrical resistivity of the dry gravel
Parameter Ke was more than double than the Parameter Kn for which was about 900 Om and qS is the electrical resistivity of the
the specimens tested. cement slurry and was 1.02 Om immediately after mixing.
Normalized change in resistivity Dq with the changing condi-
tions can be represented as follows: 3.7.3. Vipulanandan composite resistivity model
Dq DR In order to quantify the effect of gravel content on the initial
¼ ð2Þ
q R electrical resistivity of the smart cement composite, following com-
posite resisitivity model was used and the relationship is as follows:
The modified cement material is represented in terms of resis- 2 3
tivity (q) and the changes due to stress will be quantified to eval-  
6 VG 7
uate the sensitivity of the material. 6 V þV
 G S 7
6 7
6 VG 7
6 VG þVS 7
3.5.1. Two wire method q  qS ¼ ðqG  qS Þ6
6
c
  7
7
6 0 1 0  1 p pq
4 7
The change in resistance was measured using the two probe 6 VG VG 4 4 7
6 7
4q4 þ ð1  p4  q4 Þ@  A þ p4 @ A
VG þVS VG þVS
method with the LCR meter. To minimize the contact resistances, 5
VG VG
the resistance was measured at 300 kHz using two-wire method. VG þVS
c
VG þVS
c

This configuration was first calibrated using the same liquid ð6Þ
(cement slurry) to determine the parameter K in Eq. (1).
 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530 523

where q is the electrical resistivity in Om for the composite mix- 3.7.6.2. Case 2: special bulk material – resistance only. Case 2 is a
ture and VG is the gravel volume content and VS is the slurry volume special case of Case 1 in which the capacitance of the bulk material
content. The electrical resistivity of the dry gravel ðqG Þ was 900 (Cb) is assumed to be negligible. The total impedance of the equiv-
Om, p4 and q4 are model parameters and the maximum gravel vol- alent circuit for Case 2 (Z2) is
 
ume content, VGVþVG
S
, was 1.
c
2Rc ðrÞ 2xR2c C c ðrÞ
Z 2 ðrÞ ¼ Rb ðrÞ þ j ð12Þ
3.7.4. Vipulanandan curing model 1þx 2 R2 C 2
c c 1 þ x2 R2c C 2c
In order to represent the electrical resistivity development of
the cement, Vipulanandan Curing model was used [23] and the When the frequency of the applied signal is very low, x ? 0, Z2
relationship is as follows: = Rb + 2Rc, and when it is very high, x ? 1, Z2 = Rb.
2 3 The shape of the curves shown in Fig. 2 is very much influenced
 
by material response and the two probes used for monitoring. Test-
1 6 7
tþt 0
1 6 t min þt 0
 7
¼ 6 7 ð7Þ ing of smart cement and concrete indicated that Case 2 repre-
q qmin 4     p1pþq1 5 sented their behaviors and hence the bulk material properties
tþt 0 tþt 0
q1 þ ð1  p1  q1 Þ t min þt0
þ p1 tmin þt0
1
can be represented by resistivity and characterized at a frequency
of 300 kHz using the two probes.
where q is the electrical resistivity in Om, qmin is the minimum
electrical resistivity in Om, tmin is the time corresponding to the
minimum electrical resistivity ðqmin Þ, t represents the curing time, 4. Results and discussion
t 0 is the model parameter influenced by the initial resistivity and
p1 and q1 are time-dependent model parameters as follows: 4.1. Density
t
p1 ¼ p10 þ ð8Þ The average density of smart cement (water-to-cement ratio of
AþBt
0.38) without any gravel was 1.97 g/cm3. The density of the silica
t aggregates were 2.65 g/cm3 and the density of cement was 3.14
q 1 ¼ q 10 þ ð9Þ g/cm3. The density of 25%, 50% and 75% gravel smart cement com-
A0 þ B0  t
posite increased to 2.31 g/cm3, 2.19 g/cm3 and 2.04 g/cm3 and rep-
In which p10 , A, B, q10 , A0 and B0 are model parameters. resented an increase of 17.3%, 11.2% and 3.6% respectively. The
increase in density of the concrete compared to the cement slurry
3.7.5. Vipulanandan piezoresistivity model was due the addition of higher density aggregates.
In order to represent the piezoresistive behavior of the hard-
ened cement, Vipulanandan Piezoresistivity Model [10,13] was 4.2. Ultrasonic pulse velocity
used and the relationship is as follows:


Dq

rmax  ððDqqÞÞ Pulse velocity of the smart cement composites measured during
the initial 28 days of curing.
r¼ q 0
  ð10Þ
Immediately after mixing the initial pulse velocity of the smart

Dq

Dq
 p2pþq2
ðqÞ ðqÞ 2
cement with a density of 1.97 g/cm3 and 0.26% air content was
q2 þ ð1  p2  q2 Þ  þ p2 
ðDqqÞ0 ðDqqÞ0 1050 m/s which is less than the pulse velocity of water (1500 m/
where rmax is the maximum stress, ðDq=qÞ0 is the piezoresistivity of s) due to air voids in the mortar [31].
the hardened cement under the maximum stress and p2 and q2 are
model parameters influenced by the material properties. 4.2.1. Long term pulse velocity
4.2.1.1. Smart cement. The pulse velocity of the smart cement was
3.7.6. Impedance model 3520 m/s after 28 days of curing. Parameters p3 and q3 for the
Vipulanandan et al. [29] studied different possible equivalent model were 3.174 and 0.0237 respectively (Fig. 3, Table 2).
circuits for composite materials with two probes measurement
and found appropriate equivalent circuits to represent materials.

3.7.6.1. Case 1: general bulk material – capacitance and resistance. In

the equivalent circuit for Case1, the contacts were connected in
series, and both the contacts and the bulk material were repre-
sented using a capacitor and a resistor connected in parallel. In
the equivalent circuit for Case 1, Rb and Cb are resistance and
capacitance of the bulk material, respectively; and Rc and Cc are
resistance and capacitance of the contacts, respectively. Both con-
tacts are represented with the same resistance (Rc) and capacitance
(Cc), as they are identical. Total impedance of the equivalent circuit
for Case 1 (Z1) can be represented as:
( )
Rb ðrÞ 2Rc ðrÞ 2xR2c C c ðrÞ xR2b C b ðrÞ
Z 1 ðrÞ ¼ þ j þ
1 þ x2 R2b C b2
1 þ x2 R2c C c 2
1 þ x2 R2c C c 2
1 þ x2 R2b C 2b
ð11Þ
where x is the angular frequency of the applied signal. When the
frequency of the applied signal is very low, x ? 0, Z1 = Rb + 2Rc,
and when it is very high, x ? 1, Z1 = 0. Fig. 2. Different impedance–frequency models for composite materials.
524 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530

Fig. 4. Model parameters of p-q model for evaluating the pulse velocity of the smart
Fig. 3. Development of pulse velocity of smart cement composites during 28 days. cement composites.

4.2.1.2. Smart cement composite.

i. 25% Gravel: The pulse velocity of the 25% gravel smart
cement increased by 7% compared to the smart cement to
3760 m/s after 28 days of curing. Parameters p3 and q3 for
the model were 3.021 and 0.0249 respectively.
ii. 50% Gravel: The pulse velocity of the 50% gravel smart
cement increased by 18% compared to the smart cement to
4150 m/s after 28 days of curing. Parameters p3 and q3 for
the model were 2.942 and 0.0259 respectively.
iii. 75% Gravel: The pulse velocity of the 75% gravel smart
cement increased by 35% compared to the smart cement to
4750 m/s after 28 days of curing. Parameters p3 and q3 for
the model were 2.851 and 0.0268 respectively.

Fig. 5. Impedance characterization of the smart cement composites immediately

As shown in Fig. 4, both p3 and q3 were gravel dependent
after mixing.
parameters and the relationship is as follow:


VG
p3 ¼ ð0:0044  105 Þ þ 3:1597; R2 ¼ 0:98 ð13Þ
VG þ VS


VG
q3 ¼ ð4  105 Þ þ 0:0024; R2 ¼ 0:99 ð14Þ
VG þ VS

4.3. Electrical resistivity

4.3.1. Impedance vs frequency relations

Investigation of the impedance versus frequency relationship
tested immediately after mixing and also after 28 days of curing
for smart cement composites is shown in Figs. 5 and 6. The
observed shape of the curve represents the Case 2, indicating that Fig. 6. Impedance characterization of the smart cement composites after 28 days of
the bulk material can be represented by resistance. curing.

4.3.2. Initial resistivity 4.3.2.2. Smart cement composite.

Initial electrical resistivity of smart cement composites are i. 25% Gravel: The average initial electrical resistivity of the
shown in Fig. 7. 25% gravel smart composite increased by 12% to 1.14 Om.
ii. 50% Gravel: The average initial electrical resistivity of the 50%
gravel smart composite was 1.73 Om, a 70% increment
4.3.2.1. Smart cement. The average initial electrical resistivity of the which is due to increment of gravel content in the composite.
smart cement was 1.02 Om.

Table 2
Model parameters of p-q model for evaluating the pulse velocity of the smart cement composites during 28 days of curing.

Smart Cement Composite p3 q3 tc (days) Vp0 (m/s) Vpc (m/s) R2 RMSE (m/s)
No Gravel 3.174 0.0237 28 1050 3520 0.99 40
25% Gravel 3.021 0.0249 28 1160 3760 0.99 20
50% Gravel 2.942 0.0259 28 1320 4150 0.99 10
75% Gravel 2.851 0.0268 28 1490 4750 0.99 50
 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530
Fig. 7. Initial electrical resistivity of smart cement composites.
iii. 75% Gravel: The average initial electrical resistivity of the
75% gravel smart composite increased by 267% to 3.74
Om. this extraordinary increment is due to high gravel con-
tent of the composite.
Figs. 8 and 9 show the initial and 28 days cured electrical resis-
tivity of the smart cement composite with different aggregate con-
tent respectively. At low aggregate content in the cement
composites, the electrical resistivity was mainly due to electrical
resistivity of cement. With the aggregate content increase dramatic
increment in electrical resistivity was observed. Mixture theory
with parallel circuit better predicted the resistivity than the series
Fig. 8. Initial electrical resistivity of smart cement composites with different gravel
contents (volume).
Fig. 9. Electrical resistivity of smart cement composites with different gravel
contents (volume) after 28 days of curing.
525
circuit and R2 and RMSE are summarized in Table 3. The RMSE for
the parallel circuit after 1 day and 28 days curing were 125 Om and
5 Om respectively. The Resistivity-gravel content model predicted
this trend accurately with model parameters p4 and q4 of 3.1 and
3.2 respectively and the coefficient of variation (R2) of 0.99 and
RMSE was 23 Om. The parameters p4 and q4 were 11.5 and 19.5
and the coefficient of variation (R2) of 0.99 and RMSE was 2 Om
for 28 days cured smart cement composites (Table 3.
4.3.3. Resistivity during setting
Electrical resistivity of a concrete is determined mainly by the
porosity and conductive ion concentration in the pore solution.
From the standpoint of conductivity, concrete can be regarded as
a two-component composite material, pore solution and solid
phase (aggregate + hydration products + unhydrated binders)
[32]. During the setting of the cement, the capillary porosity is con-
stant and changes in the pore solution resistivity leads to deter-
mine the evolution of the slurry resistivity [33]. As shown in
Fig. 10, the pore solution resistivity decreased initially and reached
a minimum resistivity of qmin at specific time of t min which is due to
increment of ionic concentration in pore solution. By preceding the
hydration, production of C-S-H network caused later increment in
bulk paste resistivity [34].
4.3.3.1. Smart cement. The minimum electrical resistivity of the
smart cement after 90 min of mixing was 0.79 Om (Table 4,
Figs. 10 and 11).
4.3.3.2. Smart cement composite.
i. 25% Gravel: The minimum electrical resistivity of the 25%
gravel smart composite increased by 22% to 0.96 O.m. The
time corresponds to the minimum resistivity of 25% gravel
smart composite reduced by 15–75 min compare to no
gravel smart cement composite.
ii. 50% Gravel: The minimum electrical resistivity of the 50%
gravel smart composite increased by 79% to 1.41 Om. The
time corresponds to the minimum resistivity of 50% gravel
smart composite reduced by 15–75 min compare to no
gravel smart cement composite.
iii. 75% Gravel: The minimum electrical resistivity of the 75%
gravel smart composite increased by 339% to 3.46 Om. The
time corresponds to the minimum resistivity of 75% gravel
smart composite reduced by 30–60 min compare to no
gravel smart cement composite.
4.3.4. Long term resistivity
After the setting time, hardened cement composite has a com-
plete connected network which leads to form percolated path of
C-S-H causing high resistivity due to its continuous gel micropores.
Later, volume fraction of C-S-H will play the main role in changes
of the resistivity of the cement composites [33] (See Table 5).
4.3.4.1. Smart cement. After 28 days of curing, the electrical resis-
tivity of smart cement was 14.14 Om (Fig. 12).
4.3.4.2. Smart cement composites.
i. 25% Gravel: After 28 days of curing the electrical resistivity of
25% gravel smart cement composite increased by 20% to
17.00 Om.
ii. 50% Gravel: After 28 days of curing the electrical resistivity of
50% gravel smart cement composite increased by 79% to
25.26 Om.
iii. 75% Gravel: After 28 days of curing the electrical resistivity of
75% gravel smart cement composite increased by 333% to
61.24 Om.
526 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530

Table 3
Model parameters of electrical resistivity of smart cement composites with different gravel contents.

Mixture Theory (Series) Mixture Theory (Parallel) Resistivity-Gravel Content Model

qS qG R 2
RMSE (Om) qS qG R 2
RMSE (Om) p4 q4 R2 RMSE (Om)
Initial Measurements
1.0 900 11.0 664 1.0 900 0.57 125 3.1 3.2 0.99 23
After 28 Days of Curing
13.3 900 0.49 354 13.3 900 0.99 5 11.5 19.5 0.99 2

Fig. 10. Development of electrical resistivity of smart cement composites during

the initial 150 min of curing. Fig. 11. Development of electrical resistivity of smart cement composites during
the initial 24 h of curing.

As shown in Fig. 13, Curing Model parameters p10 and q10 are iii. 75% Gravel: The RI24 of the 75% gravel smart composite
linearly dependent to gravel percentage. The resistivity index decreased by 13–478% compared to the smart cement with
(RI24) varied with the gravel content (Fig. 14) and was linearly no gravel.
related to the compressive strength (Fig. 15).
VG 4.4. Property correlations
p10 ¼ 0:0201 þ 1:19; R2 ¼ 0:99 ð15Þ
VG þ VS
Monitoring the changes in electrical resistivity and compressive
VG 2 pulse velocity of the smart cement composite with different gravel
q10 ¼ 0:0026 þ 0:75; R ¼ 0:92 ð16Þ
VG þ VS content and also during curing time, it can be concluded that elec-
trical measurement is more sensitive to changes due to aggregate
content and hydration during curing compare to compressive pulse
4.3.5. Electrical resistivity index (RI24)
velocity measurements. Addition of up to 75% gravel to smart
Vipulanandan et al. [13] used the resistivity index (RI24) as an
cement showed 267% change in initial electrical resistivity while
indicator of hydration development after 24 h which is the maxi-
it was 41% for the initial compressive pulse velocity (Fig. 16).
mum percentage change in resistivity after 24 h with respect to
Change in electrical resistivity of smart cement in 28 days of curing
minimum resistivity. This index direct influence on the strength
is 1204% while it was only 235% for compressive pulse velocity
of the hardened cement and concrete.
(Fig. 17).
4.3.5.1. Smart cement. The RI24 for the smart cement was 550%.
4.5. Compressive strength
4.3.5.2. Smart cement composites.
i. 25% Gravel: The RI24 of the 25% gravel smart composite Compressive strength of smart cement composites was tested
increased by 12–617% compared to the smart cement with after 1 and 28 days of curing (Fig. 18).
no gravel.
ii. 50% Gravel: The RI24 of the 50% gravel smart composite
decreased by 7–512% compared to the smart cement with 4.5.1. 1 Day curing
no gravel. 4.5.1.1. Smart cement. After 1 day of curing, the compressive
strength of the smart cement was 8.6 MPa.

Table 4
Electrical resistivity parameters of the smart cement composites slurries.

Smart Cement Composite q0 qmin tmin q24 q24 qmin

qmin
(by volume) ðX  mÞ ðX  mÞ ðminuteÞ ðX  mÞ %

No Gravel 1.02 0.79 90 5.14 550%

25% Gravel 1.14 0.96 75 6.88 617%
50% Gravel 1.73 1.41 75 8.63 512%
75% Gravel 3.74 3.46 60 20.01 478%
 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530 527

Table 5
Model parameters of p-q model for evaluating the electrical resistivity of the smart cement composites during 28 days of curing.

Smart Cement Composite q10 A0 B0 p10 A B t0 tmin qmin R2 RMSE

No Gravel 0.75 130 1.65 1.21 2000 0.24 100 90 0.79 0.99 0.18
25% Gravel 0.84 100 1.25 1.64 1800 0.14 110 75 0.96 0.99 0.16
50% Gravel 0.89 100 1.29 2.22 1500 0.18 100 75 1.41 0.99 0.34
75% Gravel 0.91 100 1.09 2.59 1000 0.15 130 60 3.46 0.99 0.44

Fig. 12. Development of electrical resistivity of smart cement composites during Fig. 15. Relationship between electrical resistivity index and compressive strength
28 days of curing. of smart cement composites up to 28 days of curing.

Fig. 13. Model parameters of Vipulanandan Curing Model.

Fig. 16. Changes in initial electrical resistivity and initial compressive pulse
velocity of the smart cement composite with varying gravel percentages.

Fig. 14. Electrical resistivity index (RI24) of smart cement composites.

Fig. 17. Changes in electrical resistivity and compressive pulse velocity of the smart
cement during 28 days of curing.

4.5.1.2. Smart cement composites.

ii. 50% Gravel: The compressive strength of the 50% gravel
i. 25% Gravel: The compressive strength of the 25% gravel
smart composite decreased by 11% to 7.7 MPa compare to
smart composite increased by 28% to 10.9 MPa compare to
the smart cement with no gravel.
the smart cement with no gravel.
528 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530
Fig. 18. Compressive strength of cement composites up to 28 days of curing.
iii. 75% Gravel: The compressive strength of the 75% gravel
smart composite decreased by 29% to 6.1 MPa compare to
the smart cement with no gravel.
4.5.2. 28 days curing
4.5.2.1. Smart cement. After 1 day of curing, the compressive
strength of the smart cement was 21.7 MPa.
4.5.2.2. Smart cement composites.
i. 25% Gravel: The compressive strength of the
smart composite increased by 33% to 28.9 MPa
the smart cement with no gravel.
ii. 50% Gravel: The compressive strength of the
smart composite decreased by 21% to 17.2 MPa
the smart cement with no gravel.
Fig. 19. Piezoresistivity of smart cement composites after 1 and 28 days of curing: (a) No gravel; (b) 25% gravel; c) 50% Gravel and d) 75% Gravel.
iii. 75% Gravel: The compressive strength of the 75% gravel
smart composite decreased by 43% to 12.4 MPa compare to
the smart cement with no gravel.
Changes in compressive strength of the smart cement aggregate
composite can be justified with porosity of the composite.
4.6. Piezoresistivity
Piezoresistive behavior of smart cement composites was evalu-
ated after 1 and 28 days of curing.
4.6.1. 1 day curing
4.6.1.1. Smart cement. After 1 day of curing, the piezoresistivity of
the smart cement at the peak compressive stress was 375%
(Fig. 19. Table 6). Parameters p2 and q2 for the model were 0.61
and 0.57 respectively.
4.6.1.2. Smart cement composites.
i. 25% Gravel: The piezoresistivity of the 25% gravel smart com-
posite reduced by 23–288% compare to the smart cement.
Parameters p2 and q2 for the model were 0.51 and 0.79
respectively.
ii. 50% Gravel: The piezoresistivity of the 50% gravel smart com-
posite reduced by 46–200% compare to the smart cement.
25% gravel Parameters p2 and q2 for the model were 0.42 and 0.91
compare to respectively.
iii. 75% Gravel: The piezoresistivity of the 75% gravel smart com-
50% gravel posite reduced by 57–163% compare to the smart cement.
compare to Parameters p2 and q2 for the model were 0.40 and 0.80
respectively.
 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530 529

Table 6
Model parameters of p-q model for evaluating the piezoresistivity behavior of the smart cement composites.

Smart Cement Composite p2 q2 R2 Compressive Ultimate RMSE (MPa)

Strength (MPa) Piezoresistivity (%)
1 Day Curing
No Gravel 0.61 0.57 0.99 8.6 375 0.3
25% Gravel 0.51 0.79 0.98 10.9 288 0.4
50% Gravel 0.42 0.91 0.99 7.7 200 0.2
75% Gravel 0.40 0.80 0.99 6.1 163 0.3
28 Days Curing
No Gravel 0.83 0.42 0.98 21.7 204 1.0
25% Gravel 0.20 0.60 0.98 28.9 147 1.1
50% Gravel 0.29 0.73 0.99 17.2 115 0.7
75% Gravel 0.81 0.40 0.99 12.4 101 0.4

4.6.2. 28 days curing respectively. Vipulanandan Piezoresistivity Model can be used

4.6.2.1. Smart cement. After 1 day of curing, the piezoresistivity of
the smart cement was 204%. Parameters p2 and q2 for the model
were 0.83 and 0.42 respectively.
4.6.2.2. Smart cement composites.
i. 25% Gravel: The piezoresistivity of the 25% gravel smart com-
posite reduced by 28–147% compare to the smart cement.
Parameters p2 and q2 for the model were 0.20 and 0.60
respectively.
ii. 50% Gravel: The piezoresistivity of the 50% gravel smart com-
posite reduced by 44–115% compare to the smart cement.
Parameters p2 and q2 for the model were 0.29 and 0.73
respectively.
iii. 75% Gravel: The piezoresistivity of the 75% gravel smart com-
posite reduced by 51–101% compare to the smart cement.
Parameters p2 and q2 for the model were 0.81 and 0.40
respectively.
to predict the piezoresistivity behavior of the smart cement
composites very well.
4. Electrical resistivity index (RI24) was a good indicator of the
hydration process of the smart cement composites after 24 h
and also well prediction of the compressive strength of the
smart cement composites.
5. The failure strain of concrete is 0.3%, hence piezoresisitive con-
crete has magnified the monitoring resistivity parameter by 336
times (33,600%) or more higher based on the aggregate content
and making the concrete a bulk sensor.
Conflict of interest
There is no conflict of interest.
Acknowledgments

5. Conclusions
The smart cement was used as the binder in the concrete to
make it a highly bulk sensing material. Based on experimental
and analytical study on the behavior of concrete with varying
amount of aggregates and smart cement binder, the following con-
clusions are advanced:
1. Aggregate content and curing time increased the pulse velocity
of smart cement aggregate composites such that the pulse
velocities after 28 days of curing for 0, 25%, 50% and 75% gravel
content smart cement aggregate composites were 3520 m/s,
3760 m/s, 4150 m/s and 4750 m/s respectively. New Vipulanan-
dan pulse velocity-curing model predicted the variation of the
aggregate content and curing time very well.
2. Addition of coarse aggregate and curing time increased the ini-
tial electrical resistivity of the smart cement composite as well
as long term electrical resistivity. The initial electrical resistivity
of smart cement was 1.02 Om which increased to 1.14 Om,
1.73 Om and 3.74 Om for 25%, 50% and 75% gravel respec-
tively. After 28 days of curing, the electrical resistivity of
smart cement was 14.14 Om which increased to 17.00 Om,
25.26 Om and 61.24 Om for 25%, 50% and 75% gravel respec-
tively. A new Vipulanandan composite resistivity model was
developed to predict the resistivity from the constituent prop-
erties. Also Vipulanandan Curing Model predicted the electrical
resistivity development of the smart cement composites during
short and long term curing.
3. The piezoresistivity of the smart cement aggregate composites
with 0, 25%, 50% and 75% gravel content after 28 days of curing
were 204%, 147%, 115% and 101% at a peak compressive stress
This study was supported by the Center for Innovative Grouting
Materials and Technology (CIGMAT) and Texas Hurricane Center
for Innovative Technology (THC-IT) at the University of Houston,
Houston, Texas with funding from various industries.
References
[1] T.C. Hou, Y.M. Su, Y.R. Chen, P.J. Chen, Effects of coarse aggregates on the
electrical resistivity of Portland cement concrete, Constr. Build. Mater. 133
(2017) 397–408.
[2] C. Vipulanandan, E. Demircan, ‘‘Designing and Characterizing LEED Concrete
for Drilled Shaft Applications,” in: Proceedings, Foundation Congress 2009,
Contemporary Topics in Deep Foundations, ASCE, GSP 185, 2009, pp. 55–62.
[3] C. Vipulanandan, A. Mohammed, Smart cement rheological and piezoresistive
behavior for oil well applications, J. Petrol. Sci. Eng. 135 (2015) 50–58.
[4] C. Vipulanandan, K. Ali, Smart cement piezoresistive behavior with and
without sodium meta-silicate under temperature and curing environments for
oil well applications, J. Civ. Eng. Mater. Am. Soc. Civ. Eng. (ASCE) (2016),
https://doi.org/10.1061/MT.1943-055330001667.
[5] C. Vipulanandan, A. Mohammed, A.S. Ganpatye, Smart Cement Performance
Enhancement with Nano Al2O3 for Real-Time Monitoring Applications Using
Vipulanandan Models, in: Offshore Technology Conference. OTC-28880-MS,
2018.
[6] S.M. Parvasi, C. Xu, Q. Kong, G. Song, Detection of multiple thin surface cracks
using vibrothermography with low-power piezoceramic-based ultrasonic
actuator—a numerical study with experimental verification, Smart Mater.
Struct. 25 (5) (2016) 055042.
[7] C.M. Sayers, R.L. Grenfell, Ultrasonic propagation through hydrating cements,
Ultrasonics 31 (3) (1993) 147–153.
[8] J.T. Petro, J. Kim, Detection of delamination in concrete using ultrasonic pulse
velocity test, Constr. Build. Mater. 26 (1) (2012) 574–582.
[9] L. Zeng, S.M. Parvasi, Q. Kong, L. Huo, M. Li, G. Song, Bond slip detection of
concrete-encased composite structure using shear wave based active sensing
approach, Smart Mater. Struct. 24 (12) (2015) 125026.
[10] C. Vipulanandan, M. Ali, B. Basirat, A. Reddy, N. Amin, A. Mohammed, H.
Farzam, Field Test for Real Time Monitoring of Piezoresistive Smart Cement to
Verify the Cementing Operations, in: Offshore Technology Conference. OTC-
27060-MS, 2016b.
530 C. Vipulanandan, N. Amani / Construction and Building Materials 175 (2018) 519–530
[11] A.A. Ramezanianpour, A. Pilvar, M. Mahdikhani, F. Moodi, Practical evaluation
of relationship between concrete resistivity, water penetration, rapid chloride
penetration and compressive strength, Constr. Build. Mater. 25 (5) (2011)
2472–2479.
[12] F. Azhari, N. Banthia, Cement-based sensors with carbon fibers and carbon
nanotubes for piezoresistive sensing, Cem. Concr. Compos. 34 (7) (2012) 866–
873.
[13] C. Vipulanandan, M. Heidari, Q. Qu, H. Farzam, J.M. Pappas, Behavior of
piezoresistive smart cement contaminated with oil based drilling mud, in:
Offshore Technology Conference. OTC 25200-MS, 2014.
[14] C. Vipulanandan, N. Amani, Behavior of Nano Calcium Carbonate Modified
Smart Cement Contaminated with Oil Based Drilling Mud, in: Offshore
Technology Conference. OTC-25845-MS, 2015b.
[15] C. Vipulanandan, V. Garas, Electrical resistivity, pulse velocity, and
compressive properties of carbon fiber-reinforced cement mortar, J. Mater.
Civ. Eng. 20 (2) (2008) 93–101.
[16] C. Vipulanandan, A. Reddy, Smart Foam Cement Characterization for Real Time
Monitoring of Ultra-Deepwater Oil Well Cementing Applications, in: AADE
Fluids Technical Conference and Exhibition. AADE-16-FTCE-84, 2016c.
[17] C. Vipulanandan, A. Mohammed, Smart cement modified with iron oxide
nanoparticles to enhance the piezoresistive behavior and compressive
strength for oil well applications, Smart Mater. Struct. 24 (12) (2015) 125020.
[18] F. Presuel-Moreno, Y.Y. Wu, Y. Liu, Effect of curing regime on concrete
resistivity and aging factor over time, Constr. Build. Mater. 48 (2013) 874–882.
[19] R.A. Medeiros-Junior, M.G. Lima, Electrical resistivity of unsaturated concrete
using different types of cement, Constr. Build. Mater. 107 (2016) 11–16.
[20] H.Y. Chu, J.K. Chen, The experimental study on the correlation of resistivity and
damage for conductive concrete, Cem. Concr. Compos. 67 (2016) 12–19.
[21] A. Princigallo, K. van Breugel, G. Levita, Influence of the aggregate on the
electrical conductivity of Portland cement concretes, Cem. Concr. Res. 33 (11)
(2003) 1755–1763.
[22] O. Sengul, Use of electrical resistivity as an indicator for durability, Constr.
Build. Mater. 73 (2014) 434–441.
[23] M. Saleem, M. Shameem, S.E. Hussain, M. Maslehuddin, Effect of moisture,
chloride and sulphate contamination on the electrical resistivity of Portland
cement concrete, Constr. Build. Mater. 10 (3) (1996) 209–214.
[24] C. Shi, Effect of mixing proportions of concrete on its electrical conductivity
and the rapid chloride permeability test (ASTM C1202 or AASHTO T277)
results, Cem. Concr. Res. 34 (3) (2004) 537–545.
[25] D. Gao, M. Sturm, Y.L. Mo, Electrical resistance of carbon-nanofiber concrete,
Smart Mater. Struct. 18 (9) (2009) 095039.
[26] C. Chang, G. Song, D. Gao, Y.L. Mo, Temperature and mixing effects on electrical
resistivity of carbon fiber enhanced concrete, Smart Mater. Struct. 22 (3)
(2013) 035021.
[27] J. Katzer, Median diameter as grading characteristic for fine aggregate cement
composite designing, Constr. Build. Mater. 35 (2012) 884–887.
[28] ASTM C136 2016, Standard Test Method for Sieve Analysis of Fine and Coarse
Aggregates, ASTM International, West Conshohocken, PA, 2016.
[29] C. Vipulanandan, P. Prashanth, Impedance spectroscopy characterization of a
piezoresistive structural polymer composite bulk sensor, J. Test. Eval. 41
(2013) 898–904.
[30] D.S. McLachlan, M. Blaszkiewicz, R.E. Newnham, Electrical resistivity of
composites, J. Am. Ceram. Soc. 73 (8) (1990) 2187–2203.
[31] C.M. Sayers, A. Dahlin, Propagation of ultrasound through hydrating cement
pastes at early times, Adv. Cem. Based Mater. 1 (1) (1993) 12–21.
[32] L. Xiao, Z. Li, Early-age hydration of fresh concrete monitored by non-contact
electrical resistivity measurement, Cem. Concr. Res. 38 (3) (2008) 312–319.
[33] Z. Liu, Y. Zhang, Q. Jiang, Continuous tracking of the relationship between
resistivity and pore structure of cement pastes, Constr. Build. Mater. 53 (2014)
26–31.
[34] J. Zhang, L. Qin, Z. Li, Hydration monitoring of cement-based materials with
resistivity and ultrasonic methods, Mater. Struct. 42 (1) (2009) 15–24.