1. Introduction
The transverse deformation analysis of structures is usually limited to obtaining Poisson’s ratio, which is only applicable to the range of elastic concrete behavior. At the same time, for compression, this range can reach half of the concrete strength. Consequently, the elastic compressed concrete stage is an essential part of concrete behavior investigations [
1]. However, according to modern design codes, the elastic stage is limited to about 40% of its strength [
2,
3]. Further concrete behavior is characterized by elastic–plastic deformations, which can be graphically represented by a convex square parabola relative to the deformation’s axis. If instead of a square parabola, a sinusoidal approximation is used, the difference in potential energy is about 7% [
4].
The relationship between stresses (
) and strains (
) in compressed concrete is characterized by the following three parts [
1]:
- –
A linear part from 0 to , where is concrete strength;
- –
A convex parabolic part bound by ;
- –
A descending branch.
Deformations range from 0 to 0.5‰, from 0.5 to 2‰, and from 2 to 3.5‰ in the first, second, and third parts, respectively.
This relationship was proposed to describe the behavior of normal-strength concrete (NSC) up to class C60. High-strength concrete (HSC) represents classes C70–C90 [
2] and exhibits mainly elastic behavior with weakly expressed elastic–plastic deformations [
5]. For the more accurate identification of HSC non-elastic deformations, the use of transverse deformation graphs of the compressed specimens was proposed [
6].
At the same time, an additional important factor affecting compressed concrete behavior is the specimen’s loading rate [
7,
8,
9]. The loading rate used in the present research was selected according to standard requirements of modern codes. Following [
10], it is recommended to load a concrete specimen at a rate within (0.2 ... 1.0) MPa/s; according to [
11], the recommended rate is (0.25 ± 0.05) MPa/s. However, at such loading rates, longitudinal deformations, even for NSC, mainly exhibit elastic behavior [
7,
8,
9].
Twenty-two cylindrical NSC specimens were tested under uniaxial loading [
7]. The specimen diameters were between 10 and 80 cm, and their heights were from 20 to 320 cm (which corresponds to a real column length). The specimens were loaded at a rate of about 0.1 MPa/s. The obtained
vs.
relationships, when
was the compressed concrete stresses and
was the longitudinal deformations, were characterized by an almost linear performance up until specimen failure. Similar results were obtained for uniaxial loaded 15 × 30 cm cylindrical specimens made of steel fibered reinforced concrete with a strength of 27.6 MPa [
8]. In this case, the loading rate was 1.3 mm/min (which corresponds to about 0.1 MPa/s). Comparable results were obtained for twenty-six cylindrical NSC specimens loaded at a rate of (0.25 ± 0.05) MPa/s [
9].
It was shown that in certain cases, particularly in steel-fibered high-strength concrete (SFHSC), transverse deformations efficiently indicate elastic–plastic concrete behavior, which is less pronounced when longitudinal deformations are used [
6]. A corresponding theoretical model for the type of transverse cracks and their development as well as a consequent scheme for specimen failure was proposed. Following this concept, the transverse crack development process can be divided into three stages:
- –
Linear specimen behavior;
- –
Nonlinear specimen behavior;
- –
Ultimate limit state (ULS).
The first stage continues until the initiation of cracks (
), where
is the transverse deformation value, and
is the ultimate elastic transverse deformation. At this stage, the cylindrical specimen’s cross section is nondamaged and has an original circular shape. The second stage begins from the initiation of cracks, corresponding to
, and continues until a twofold ultimate elastic transverse deformations value (
) is reached. In this case, it is assumed that the specimen’s cross section has an inner circular part (unbroken core) and several outer wedge-shaped segments separated by transverse cracks. The last stage corresponds to transverse deformations that exceed
. Similar concrete and mortar cylindrical specimen failure patterns were also obtained under cyclic uniaxial compression [
12].
The present research is focused on the experimental verification of the proposed theoretical model based on the dependence of transverse deformations on longitudinal ones. With this aim, cylindrical NSC specimens were tested, and transverse deformations were analyzed. The specimens were tested at a loading rate of . Longitudinal and transverse deformations were measured and recorded simultaneously.
The obtained experimental results were further evaluated from the point of view of confirming the principles of the structural phenomenon [
13] (a brief description of this phenomenon is given in the following sub-section).
The Essence of the Structural Phenomenon
The structural phenomenon was proposed as a result of an extensive experimental and theoretical database analysis related to concrete structures [
13]. The essence of the phenomenon is that structural parameters in elastic state increase or decrease twice at failure. This behavior usually occurs for symmetrical structures under symmetrical loading. The phenomenon has been proven at the material level and is also valid for reinforced concrete structures subjected to static and dynamic loads.
The structural phenomenon is based on two theories: the quasi-isotropic state of a structure at the ultimate limit state (ULS) [
14] and the mini-max principle [
15]. Theoretical and experimental confirmation of the phenomenon is obtained for different structure types, such as beams, frames, spatial structures, and structural joints. The structural phenomenon allows additional equation(s) for calculating structural parameters to be obtained. This, in turn, can lead to the development of efficient methods for structural analysis that allow using less empirical coefficients in design procedures.
2. Research Aims
A method for analyzing the weakly expressed inelastic behavior of HSC using compressed NSC subjected to relatively high loading rates is investigated in this study. Since longitudinal inelastic deformations in compressed HSC specimens are practically not manifested, transverse deformations can be used for concrete inelastic behavior analysis [
6]. The value of these deformations is approximately five times less than that of the longitudinal ones (Poisson’s ratio is approximately equal to 0.2). However, it also does not provide a good solution for the HSC problem since the initiation of longitudinal cracks caused by transverse deformations practically coincides with specimen failure.
For this reason, another solution is considered in this paper. As mentioned above, NSC behaves elastically almost until failure at rather high loading rates, as does HSC. However, maximal longitudinal deformations in NSC are almost two times higher than in HSC. At the same time, concrete tensile strength practically does not grow when the concrete class increases to over C70. Therefore, transverse deformation development in NSC is much more significant than in HSC, which leads, in turn, to the longitudinal crack development that determines inelastic concrete behavior in the transverse direction of a specimen. Thus, NSC subjected to relatively high loading rates can be considered as a model of HSC behavior at normal loading rates based on transverse deformation analysis.
The main purpose of this research is to investigate transverse deformation development in a concrete specimen under axial compression throughout the entire concrete behavior range until failure. The influence of transverse deformations on the nature of crack development and the specimen failure pattern is also considered based on the structural phenomenon [
13]. For this purpose, transverse deformation dependence on longitudinal deformations is investigated throughout the entire loading process, from the beginning of loading until failure.
The objectives of this study are:
- –
Detailed analysis of the elastic–plastic behavior of cylindrical NSC specimens and revealing the advantages of using transverse deformations as one of the main features of compressed concrete behavior;
- –
Analysis of the nature of specimen failure based on transverse crack development;
- –
Verification of a theoretical model of the specimen’s upper surface failure pattern [
6] based on the structural phenomenon;
- –
Confirming that the unbroken cylindrical core area is twice as small as the initial one;
- –
Checking the possibility of interpreting HSC behavior using relatively high NSC specimen loading rates.
3. Experimental Program
3.1. Testing Stages
In the framework of the present study, the first three 15 × 15 × 15 cm cubic specimens were tested to obtain the concrete strength. Measuring equipment suitability was preliminarily verified by applying step loading and unloading to three 15 × 30 cm cylindrical specimens. For this reason, ten load steps were selected (each load step was 40 kN). After that, two test series were conducted to analyze elastic–plastic behavior of 15 × 30 cm cylindrical specimens under a constant loading rate of 0.2 MPa/s. The first and the second test series included four and three specimens, marked as Sp1.1, Sp1.2, Sp1.3, Sp1.4 and Sp2.1, Sp2.2, Sp2.3, respectively.
3.2. Material Properties
Concrete proportioning was based on the absolute volumes method [
16]. The following materials were used to prepare the concrete mixture:
- –
A locally produced composite 42.5N CEM II/B-LL Portland cement [
17] with portions of limestone between 21 and 35%;
- –
A locally available quarry sand fraction of 0 to 4.75 mm;
- –
A locally available quarry crushed dolomite limestone coarse aggregate fraction of 19/14 according to SI 3-1998 [
18].
Basic properties of the concrete components are presented in
Table 1. The final composition is given in
Table 2. The concrete mixture was prepared using a 0.045 m
3 rotating pan mixer when the calculated mixture volume was 0.043 m
3. Dried aggregates were used for all of the mixtures.
3.3. Concrete Mixture and Specimen Preparation
The concrete mixture was prepared according to ASTM C192/C192M [
19]. Cubic and cylindrical specimens were cast. Internal and external vibrations were used to compact the fresh concrete in cylindrical and cubical molds, respectively. All of the specimens were removed from the molds 24 h after casting and were cured for 27 days at constant a
temperature and at a relative humidity of
.
3.4. Motivation for Selecting the Cylindrical Specimens
The cylindrical specimens are obviously more suitable to achieve the research goals because there is no local stress concentration problem as is the case, for example, in the corners of the prismatic ones. Additionally, cylindrical specimens are characterized by uniform deformation development along the entire perimeter of the specimen section. Thus, the edge effect problem is avoided in the tests.
The ratio between the diameter and the height of the concrete cylindrical specimen was 1:2 (15:30 cm). This ratio allows the modeling of real full-scale compressed structural element behavior. Additionally, the chosen ratio guarantees the presence of a relatively large unrestrained area in the middle of the specimen’s height [
20]. In this case, the element buckling problem is practically prevented. At the same time, due to some inaccuracy in the transmission of external compressive forces and concrete heterogeneity, additional stresses associated with small (technological) eccentricities may appear. This problem was solved by selecting a proper test setup.
3.5. Test Setup
The measurement system was selected according to the following main behavioral features of a specimen that needed to be obtained:
- –
Loading force;
- –
Longitudinal deformations;
- –
Transverse deformations.
To conduct the axial compression test, the class 1 testing machine (with a maximum error of 1% FS) and a maximum load of 3000 kN, which was controlled by a PC-unit, was used. The applied load was measured by an external pressure sensor, which had been installed in the oil line of the testing machine.
Four linear displacement transducers (LDT) with a 50-mm stroke and a high-measure resolution (
) were used to measure the longitudinal deformations. The sensors were located in the specimen’s central zone in order to exclude the influence of the edge effect associated with the friction forces occurring at the contact area between the specimen and the testing machine plates. The base of the longitudinal deformation measurement was 15 cm. The upper view of the tested specimen and the measurement devices was given in
Figure 1.
Figure 2 shows the location of the devices on the specimen’s surface development.
As mentioned above, there was a technological eccentricity problem caused by inaccuracy in the applied force transmission. This problem was solved by using four sensors located on the tested specimen’s opposite sides (see
Figure 1). Such sensor distribution allowed more accurate data to be obtained that revealed the presence of eccentricity in stress distribution. The LDT sensors were attached by originally developed mounts, allowing independent deformation transfer to each sensor.
Two strain-gauges (SG) with a 60 mm basis, a measure resolution up to 0.1 micro-strains, a resistance of 120 Ohm, and a gauge factor of 2.07 were used to measure the transverse deformations. The SG length was chosen based on the requirement of exceeding the coarse aggregate maximum dimensions by at least three times [
21]. The general view of the test setup is presented in
Figure 3.
A thin layer of sifted fine grain sand was used to level the upper surface of the cylindrical specimens. It also allowed for a reduction in the friction effect between the testing machine plate and the specimens.
5. Comparison of the Results Obtained for NSC with Available Data for HSC
Based on the data obtained in the framework of the present study (see
Section 4), the behavior of NSC, which corresponds to class C30 at relatively high loading rates, can be compared with that of HSC, which is described in [
6]. The available data on HSC (class C90) includes the test results for three cylindrical specimens. Longitudinal deformations were measured using linear variable displacement transducers (LVDT) and transverse deformations measured by strain gauges. The dependences of these two types of deformations on stresses under uniaxial load are shown in
Figure 13 and
Figure 14, respectively. Corresponding data for the comparison of the elastic and nonelastic behavior of HSC and NSC are provided in
Table 7 and
Table 8.
The limit transverse deformation of the HSC specimens corresponding to concrete elastic behavior is 0.15 ‰. In this case, the external load value is approximately half of the concrete strength, which corresponds to the structural phenomenon [
13]. The average Poisson’s ratio value is equal to 0.18. Further load increases are accompanied by the linear development of longitudinal deformations (
Figure 13), while the transverse deformations begin to exhibit elastic–plastic behavior (
Figure 14). Therefore, it is logical to analyze the ratio between the longitudinal and transverse deformations for further interpretation of the HSC behavior. Following
Table 8, at ULS, this ratio is equal to 0.31, which is close to the results obtained in the frame of the present study (see
Table 5). It should be mentioned that the relative transverse deformation growth,
, in both cases (for HSC and NSC) is equal to about 1.7.
A comparison of the results obtained for the NSC and HSC (
Table 5 and
Table 8) shows sufficient convergence of the standard deviations and the coefficients of variation for Poisson’s coefficient and
. The standard deviations are 0.024 and 0.031 and the coefficients of variation are 13.7% and 10.1% for Poisson’s coefficient and
, respectively. Thus, the obtained data confirm that it is possible to use experimental results for NSC specimens obtained at rather a high loading rate for the interpretation of HSC behavior. This, in turn, allows for the disclosure of the presence of the non-elastic behavior stage in compressed HSC, which is important for developing proper and effective design approaches.
6. Conclusions
The importance of this research is attributable to the fact that the inelastic behavior of compressed high-strength concrete specimens is practically not pronounced in longitudinal deformations and therefore is not even analyzed. Consequently, high-strength concrete is considered to be a pure brittle material, which significantly limits its applications, especially under dynamic loads. For this reason, the experimental investigation of normal strength concrete at relatively high loading rates was conducted in order to prove the effectiveness of using transverse deformations as an indicator for the interpretation of high-strength concrete behavior.
The obtained experimental results and their comparison with high-strength concrete specimen behavior showed good convergence of such parameters such as Poisson’s ratio and the ratio of transverse to longitudinal deformations in the inelastic stage of concrete behavior. The proposed theoretical failure pattern for compressed cylindrical concrete specimens, which includes an unbroken circular core surrounded by separate vertical wedge-shaped elements, was also experimentally confirmed. Additionally, the ratio between the initial specimen’s cross-section area and the destroyed one was obtained as approximately equal to 2, which corresponds to the structural phenomenon.
Thus, the approach of using transverse deformations for analyzing concrete performance proves that from the viewpoint of transverse deformations, high-strength concrete behaves inelastically and not as a brittle material. In our opinion, further investigation of this phenomenon can form a basis for the more accurate and effective design of high-strength concrete structures.