1 Introduction
Shear walls are commonly used as lateral load resisting elements for reinforced concrete buildings. The objective of this study is to investigate the use of high-strength concrete (HSC) in boundary elements of slender reinforced concrete shear walls (aspect ratio > 2). Higher axial loads have been reported to significantly affect ductility, failure modes, energy-dissipation characteristics, and strength degradation of slender walls due to flexure-compression interaction. This in turn results in excessive spalling and buckling of longitudinal reinforcement and subsequent loss of axial-load capacity or out-of-plane buckling [
1–
5]. Over the past few decades, mixes of very high compressive strengths, referred to as HSC, have been developed. HSC may be able to alleviate the problems reported with respect to shear walls under very high axial loads. HSC has been used in high demand structures such as tall buildings, long-span box-girder bridges, and offshore structures with harsh environments [
6].
Many definitions of HSC have been reported in the literature. The lower bound of compressive strength for concrete, classified as HSC, is 55 MPa per ACI 363R-10 [
6]. Supplementary cementitious materials or mineral admixtures, such as fly ash, silica fume, and other pozzolans, are commonly used for the attainment of high compressive strengths in HSC, which also requires the use of water reducing admixtures [
6]. Additionally, HSCs may or may not have discontinuous fiber reinforcement. ACI 239R-18 [
7] classifies concrete mixes with strengths under 55 MPa as normal-strength concrete (NSC) and those above 152 MPa as ultra-high-performance concrete (UHPC). AASHTO LRFD Bridge Design Specifications [
8] warrants special circumstances for the use of concrete with compressive strength higher than 69 MPa. In this study, concrete with compressive strength of 149 MPa and without discontinuous fiber reinforcement was used in the boundary elements of shear walls. Examples of HSC with compressive strengths exceeding 138 MPa without using fibers elsewhere [
9–
13].
Use of multiple materials have been investigated for use in slender shear walls. The replacement of labor-intensive boundary-element reinforcement with structural steel led to strengths, stiffness, energy-dissipation characteristics, and failure modes similar to those of conventional walls [
14]. Furthermore, the use of concrete-filled steel tubes or carbon fiber-reinforced polymers with confined cores for boundary elements resulted in sections with enhanced strengths, stiffness, and ductility [
15]. The use of HSC in the upper half and steel fiber-reinforced high strength concrete in the lower half of slender walls improved flexural deformation capacity, toughness, and cracking behavior [
16]. The use of HSC in shear walls and other structural elements may also entail economic benefits [
17].
Dan et al. [
18] conducted experiments on composite steel concrete shear walls with encased profiles, involving different structural steel shapes placed longitudinally. Flexure was reported to dominate failure with the specimens exhibiting better ductility than their conventional equivalents. Alzeni and Bruneau [
19] evaluated the seismic performance of concrete-filled sandwich steel panel cantilever walls and reported ductile behavior with controlled strength degradation at high drifts. Ultimate failure mode was controlled by fracture of skin plates. Epackachi et al. [
20] tested four steel-plate composite walls with aspect ratio of one. Specimens exhibited flexural failure with crushing of concrete at wall toes and tensile yielding of faceplates. The initial stiffness of the wall was sensitive to its connection to the foundation.
Liao et al. [
21] conducted a study on shear walls with steel-reinforced concrete (structural steel encased in concrete) boundary elements for different aspect ratios and axial-load ratios (ALRs), and reported an increase in lateral-load carrying capacity and decrease in ductility and energy dissipation capacity with increase in axial load. Furthermore, an improvement in ductility of the proposed system, when compared to that of conventional reinforced-concrete boundary elements, was observed. More information on shear walls and strategies of mitigating damage thereof can be found elsewhere [
22–
24].
Although, the aforementioned studies investigated hybrid shear walls of various kinds and materials, there are no studies in the literature that studied strategic placement of HSC in boundary elements of slender reinforced concrete shear walls. The research performed in this study fills this knowledge gap, presents the impact of HSC boundary elements on shear wall strength and stiffness, and documents implications of using HSC on boundary element size and reinforcement requirements using analytical means. A conventional prototype wall, for which test data were available in the literature, was analyzed using the finite-element (FE) method. Potential reductions in the reinforcement area and size of boundary elements using HSC were investigated. Finally, the impact of cold joints resulting from dissimilar concrete pours in the web and boundary elements was examined. Analyses were run under quasi-static monotonic loading of the walls at the component level for different axial loads. Throughout the paper, walls with HSC boundary elements and NSC webs are termed as HSC walls, and benchmark walls with NSC boundary elements and NSC webs are termed as NSC walls.
2 Slender wall specimen and experiment
The benchmark NSC wall used in this study was wall RW1 tested by Thomsen [
4]. It was a slender cantilever wall with a height of 3657 mm, length of 1219 mm, and thickness of 101 mm, resulting in an aspect ratio of 3. It was constructed on a foundation with a depth of 686 mm and plan dimension of 1930 mm × 406 mm.
The concrete compressive strengths along the wall height ranged from 31.6 to 58.3 MPa from the first through the fourth story. Grade 60 rebars with yield strength of 414 MPa were used. The longitudinal reinforcement ratio for the boundary elements was calculated as 0.0330 using the cross-sectional area of the boundary elements. The longitudinal and horizontal reinforcement ratio for the web was 0.0033, with double the ratios in the top story. The reinforcement details of the wall can be found elsewhere [
25].
The wall was tested under displacement-controlled reversed cyclic loading. The vertical axial load was equal to 10% of wall’s axial capacity or 400 kN. The wall’s post-yield behavior was dominated by flexure with crack widening and yielding of longitudinal reinforcement contributing toward energy dissipation.
3 Finite-element modeling of RW1
A numerical model of the wall was made in LS-DYNA [
26], a general-purpose FE program. Nonlinear FE analysis of the wall was validated using test results reported by Thomsen [
25]. The components of the FE model are shown in Fig.1.
3.1 Wall web and boundary element concrete
The wall body was modeled using eight-node solid elements with single point integration or element formulation 1 (ELFORM 1). Winfrith concrete model was used to model concrete. It is a smeared-crack concrete model that approximates concrete compression behavior as elastic-perfectly plastic [
27]. Concrete tensile stress–strain behavior is approximated as linearly increasing and linearly decreasing for stresses up to and beyond tensile strength, respectively. Analysis results with an alternative concrete model for modeling wall elements can be found in Syed [
28]. The input parameters for Winfrith concrete model, used for NSC walls and HSC walls, are tabulated in Tab.1. The parameters for the web of the HSC wall for the four stories were kept same as those of the respective stories of the NSC wall.
The compressive strength, tensile strength, and elastic modulus of concrete, for both NSC and HSC, are significantly affected by curing conditions [
29,
30]. The elastic modulus is also very sensitive to the aggregate type and content [
30]. The properties of HSC show higher variability to proportions of mixtures and testing when compared to NSC per ACI 363R-10 [
6].
The compressive strength of NSC was per test data provided by Thomsen [
25] for each story of the wall specimen. The elastic modulus of NSC was calculated as the secant modulus at 45% of compressive strength (0.45
f’c) from the material-test data by Thomsen [
25]. HSC with a mean cube compressive strength of 148.9 MPa, taken from test data provided by Yang and Okumus [
31], was used in the boundary elements of the wall. The measured mean elastic modulus of the HSC was 57.2 GPa.
The uniaxial tensile strength was calibrated as 2.2% of the compressive strength for NSC using test results and this ratio was kept the same for HSC. The aggregate size for NSC was taken as 9.5 mm for all the cases as reported by Thomsen [
25], and was kept the same for HSC. The crack-width parameter is the crack width at which crack-normal tensile stress decreases to zero and is a function of the fracture energy. Since there was no experimental data on tensile properties of NSC or HSC, fracture energy and crack-width parameter were calculated from empirical equations of FIB Model Code 2010 for both NSC and HSC [
32]. The Poisson’s ratio ranges between 0.20 and 0.28 for concrete compressive strengths in the range of 55–80 MPa [
33] with the former being appropriate for HSC up to strengths of 124 MPa [
28]. A Poisson’s ratio of 0.2 was used for both NSC and HSC.
3.2 Reinforcing steel
The reinforcement was modeled as beam elements using the material model “Plastic kinematic or MAT 003”. Strain hardening of reinforcing steel was neglected in this study. The reinforcement was embedded in concrete using node-merger technique. The Young’s modulus, Poisson’s ratio, and yield stress of steel were taken as 2 × 105 MPa, 0.3 and 414 MPa, respectively.
3.3 Foundation and loading beam
The foundation and loading beam were modeled using eight-node solid elements with single point integration (ELFORM 1) using an elastic material model (MAT 001). The loading beam was assigned a very high modulus of elasticity for precluding any local deformations arising due to the application of the axial load and lateral push. The elastic modulus and Poisson’s ratio for the foundation were taken as 2.5 × 104 MPa and 0.2, respectively, and for the loading beam they were taken as 2.0 × 109 MPa and 0.3, respectively.
3.4 Loading and boundary conditions
Axial load and lateral push were applied on the loading beam. Axial load was applied using post-tensioned cables in the test setup. In the FE model, axial load of 400 kN was applied as pressure on finite elements nearest to the application point of post-tensioning. The common nodes at the loading beam-wall and foundation-wall interfaces were merged to simulate no-slip condition for both NSC and HSC walls. All degrees of freedom on the bottom surface of the foundation were fixed. Lateral load was applied as a displacement-controlled monotonic push of 89 mm that corresponds to a drift ratio of 2.2%. To maintain consistency with the validation specimen, the pushover analyses presented in this study were also terminated at this drift ratio.
3.5 Numerical analysis and validation
The backbone curve of the cyclic-test data was compared to the force−displacement curve obtained from the FE model as shown in Fig.2. As shown, the FE model captured the strength of the wall well. The FE model overestimated the strength by 7%, which can be expected since the test was performed under cyclic loading as opposed to monotonic loading of the FE model. The secant stiffness values at 0.1% drift, which corresponds to the end of first cycle of the test, were compared for the FE model and test specimen, as no clear data was available on the initial stiffness of the specimen. The ratio of secant stiffness obtained from FE to testing was 1.41. Matching the initial stiffness can be challenging, given that elastic modulus is very sensitive to the curing conditions [
30], concrete may crack before testing due to shrinkage, there may be strain penetration into the foundation, or there may be flexibility in the test setup. Therefore, the FE model was deemed acceptable in predicting strength and stiffness of the test specimen. Moreover, the backbone curve was obtained from cyclic-test data, which is expected to be softer due to damage accumulation from previous cycles [
34].
4 HSC walls
After validation of the FE model, HSC was strategically placed in the wall specimen at boundary elements for cost efficiency and improved performance. In this study, walls with HSC boundary elements and NSC webs are termed as HSC walls. Monotonic pushover analyses were performed for three ALRs. ALR is the ratio of axial load applied to the axial capacity of the wall. The axial capacity of the NSC wall was used for reporting ALRs. ALRs of 10%, 20%, and 25% were considered and results were compared with the respective cases of the NSC wall.
Two potential benefits of HSC walls were investigated. 1) Reduction in the boundary element reinforcement amount. This can aid in preventing reinforcement congestion in walls and accelerate construction. 2) Reduction in the size of boundary elements. Simulations were carried out to examine the prospective use of rectangular HSC walls as replacements of barbell-shaped NSC walls. This can find application in cases where architectural constraints limit the flexibility of designers to choose wall dimensions. Since HSC is envisioned to be particularly beneficial for walls that carry high axial loads, analyses were completed for varying axial loads. Finally, cold-joints, which can result from separate casting of two dissimilar concrete types in the web and boundary elements, were studied and the ensuing changes in strain distributions and global force−displacement characteristics were documented.
4.1 Effect of HSC on load–displacement behavior of wall under varying ALRs
The NSC wall, which was used to validate FE analyses (RW1), had an ALR of 10% (400 kN). To study the change in behavior of walls under higher loads, analyses were performed for ALRs of 20% and 25%, results of which are shown in Fig.3, for both NSC and HSC walls. Increasing the ALR for both NSC wall and HSC wall from 10% to 20% and 25% increased the peak strength, although the pre-cracking stiffness did not change. ALR of 25% corresponds to a load that is below the balanced-failure point on the P-M interaction diagram of both NSC and HSC walls, and hence any increase in the load up to the balance-failure point enhances the moment capacity, and hence the peak-strength.
Introduction of HSC in the boundary elements led to an increase in the initial stiffness of the walls because of HSC’s greater elastic modulus compared to NSC. Greater moment capacity of HSC wall was achieved because of decrease in the neutral-axis depth and increase in the moment arm due to HSC. Tab.2 summarizes peak strength and initial stiffness for NSC and HSC walls for varying ALRs.
The drift ratios at which cover concrete reached its ultimate strain (expected spalling strain,
εcu) and boundary-element reinforcement yielded were also obtained for the three ALRs. Compression strain of 0.0035, reported as spalling strain of cover concrete by Thomsen [
25], was taken as threshold compression strain for spalling for NSC walls. For HSC, the ultimate strain was taken as 0.0030 per fib Model Code 2010 [
32] for mean cube compressive strength of 148 MPa (C120). Drift ratios at which concrete spalled and reinforcement yielded are tabulated in Tab.3.
HSC wall experienced spalling of cover concrete at 42% to 56% higher drift ratios as compared to NSC wall for varying ALR despite its lower ultimate compressive strain capacity. Yielding of the boundary-element reinforcement was, however, attained at 7% to 10% smaller drift ratios for both walls. This is attributed to the smaller neutral axis of HSC walls as compared to NSC walls. It can also be observed that as the ALR increased, cover concrete spalled at lower drift ratios for NSC and HSC walls, but boundary-element reinforcement yielded at slightly higher drifts.
4.2 Effect of HSC on principal strains
The principal compressive and tensile strains were obtained for the NSC and HSC walls at drift ratios of 0.5%, 1.0%, and 2.0%. The principal compressive strain contours for 20% ALR are shown in Fig.4 for drift ratios of 0.5% and 2.0%. Strain contours for 1.0% drift ratio for all subsequent studies can be found in Ref. [
28]. Dotted lines show the edges of the boundary elements. Similar strain distributions were observed for other ALRs, which have been omitted for brevity.
HSC walls exhibited lower strains on the compression side than NSC walls (wall is pushed to the right). For higher drifts (2.0%), high strains were observed in the web for both NSC as well as HSC walls that originate near the boundary element-web interface. These strains had inclined orientations and were likely due to more prominent role of shear at such drift ratios. Increasing the axial load on the wall lowered these strains, likely because higher axial load aids in improving shear performance. Moreover, these high strain regions were not observed in the bending (vertical) strain contours, confirming higher shear straining in this region [
28]. The results were similar for other ALRs.
Principal tensile strain for 20% ALR are shown in Fig.5, where dotted lines are the edges of the boundary elements. Principal tensile strains were more severe for HSC wall than that for NSC wall. This is due to the fact that the use of HSC in the boundary elements decreases the neutral-axis depth leading to higher tensile strains. The change in principal tensile strains, marked by a slight streak running along the wall height, coincides with the border between the web and the boundary elements that have dissimilar concrete pours. An increase in the axial load on the walls led to localization of high-strained regions in the wall (not shown in figures).
5 Benefits of HSC walls
5.1 Reduction in boundary element longitudinal reinforcement area
The NSC wall used for validation (RW1) had four pairs of φ9.5-mm (#3) longitudinal rebars in each boundary element apart from four pairs of φ6.4-mm (#2) rebars in the web. Given that the introduction of HSC in the boundary element increases the moment capacity of the wall, it was proposed that this increase in the moment capacity could be used to reduce reinforcement area in boundary elements.
Simulations were performed for HSC walls with varying reinforcement areas in the boundary elements. The diameter of the boundary element rebars was changed in stages from 9.5 mm (#3) to 6.4 mm (#2), with one pair being replaced at a time. In principle, HSCs require higher transverse reinforcement to match the benefits derived by NSC from confinement [
35,
36]. Conversely, some fiber reinforced HSCs have been shown to require less transverse reinforcement as the fibers mitigate spalling and buckling of longitudinal bars [
37]. Investigating the shear reinforcement requirements was beyond the scope of this study, but transverse reinforcement should be considered in design according to the design requirements of specific HSCs. The longitudinal reinforcement ratios (
ρ1) of the boundary elements and layouts are shown in Fig.6 for the NSC and HSC walls.
5.1.1 Lateral load−displacement response
Lateral strengths of the resulting HSC walls with lower longitudinal reinforcement areas were obtained and compared with those of the benchmark NSC wall. Pushover curves for the different longitudinal reinforcement layouts at 20% ALR are shown in Fig.7. The results were similar for other ALRs.
The longitudinal reinforcement ratio was 0.033 for the NSC wall. For ALR of 10%, 20%, and 25%, the HSC wall exhibited the same peak strength as the NSC wall when longitudinal reinforcement area in the boundary elements was decreased by 28%, 41%, and 41%, respectively. Any further reduction in the reinforcement led to peak strengths, which were lower than those for the respective NSC cases. Longitudinal reinforcement layout for HSC walls that led to the same peak strengths as the respective NSC walls are listed in Tab.4.
The drift ratios at which cover-concrete spalling and boundary-element reinforcement yielding occurred are tabulated in Tab.5 for cases listed in Tab.4. The results show that for NSC and HSC walls with the same strength, HSC walls exhibited cover-concrete spalling at drift ratios that were 45% to 73% higher than those for NSC walls. However, HSC walls exhibited yielding of reinforcement at drift ratios that were 15% to 32% lower than those for NSC walls. This can be explained by the smaller depth of neutral axis and smaller reinforcement area of HSC walls when compared to those of NSC walls.
5.1.2 Principal strains
The principal strain contours for the “equivalent-strength” HSC walls with lower boundary-element reinforcement were obtained and compared with those for NSC walls. The comparison for 20% ALR is shown in Fig.8 and Fig.9 for principal compression and tensile strains, respectively. At this ALR, the longitudinal reinforcement ratio was 0.019 for HSC walls and 0.033 for NSC walls.
The HSC walls with lower longitudinal reinforcement areas exhibited lower principal compressive strains and higher principal tensile strains (although less dispersed at higher drift ratios) than NSC walls. High-strain regions in the web, which probably occur due to greater contribution of shear at higher drift ratios, were again observed. Based on the comparison of force−displacement behavior and principal strain contours, it can be concluded that reinforcement can be reduced by 28% to 41% in walls with HSC boundary elements.
5.2 Reduction in boundary element size
The effect of HSC boundary elements on shear wall shape was examined by comparing rectangular HSC walls to barbell-shaped NSC walls. A potential application for HSC walls includes structures in which barbell-shaped NSC walls are not suitable due to architectural constraints and rectangular HSC walls can be used as replacements without compromising strength and stiffness.
A barbell-shaped NSC wall, similar in dimensions to the rectangular walls examined in the previous sections with the exception of boundary elements, was considered. The barbell shaped NSC wall had a 101-mm thick web and 183-mm thick boundary elements as shown in Fig.10. The ratio of the boundary-element areas of the barbell-shaped NSC wall to rectangular HSC wall was 2.25. The layout and amount of reinforcement for the barbell-shaped wall were the same as those of the original rectangular wall RW1.
5.2.1 Load−displacement response
Force−displacement curves for 20% ALR are shown in Fig.11. The rectangular HSC wall exhibited stiffness and strength characteristics that were comparable to the equivalent barbell-shaped NSC wall. Similar behavior was observed for other ALRs, which is tabulated in Tab.6.
Overall, the global force−displacement behavior of the rectangular HSC wall was similar to that of the barbell-shaped NSC wall when the boundary element area of the NSC wall was 2.25 times that of the HSC wall. Both the higher compressive strength (HSC wall) and larger boundary elements (NSC wall) led to a decrease in the depth of neutral axis and increase in moment arm and moment capacity. This led to HSC and NSC walls exhibiting similar strengths and stiffness despite the difference in their boundary element dimensions.
The drift ratios at which cover-concrete spalling and boundary-element reinforcement yielding occurred are tabulated in Tab.7, and the values were 28% higher and 4% to 8% lower, respectively, for HSC rectangular walls when compared to those for NSC barbell-shaped walls. These changes in drift ratios are considered as small.
5.2.2 Principal strains
Principal compressive strains for drift ratios of 0.5%, 1.0%, and 2.0% were obtained for varying ALRs. The strain contours for 20% ALR (800 kN) for 0.5% and 2.0% drift ratios are shown in Fig.12 for rectangular HSC wall and barbell-shaped NSC wall. The rectangular HSC wall exhibited lower concrete strains than the barbell-shaped wall at all drift ratios. Similar results were observed for other ALR cases.
Principal tensile strains for 20% ALR are shown in Fig.13. They were slightly higher for the HSC wall when compared to those for barbell-shaped NSC wall at low drift ratio of 0.5% but comparable at high drift ratios. Based on load-displacement and principal strain results obtained from FE analysis, it can be reasonably concluded that HSC rectangular walls can replace NSC barbell-shaped walls when architectural constraints prevent the use of large-sized boundary elements.
6 Cold joints in HSC walls
For HSC walls, HSC was only placed in the boundary elements with NSC in the web. Two dissimilar concrete pours can result in a cold joint because of imperfect bonding, even though horizontal reinforcement is continuous across the NSC web and HSC boundary elements. A study was conducted to understand the interaction between the web and boundary elements along the interface. In this investigation, rectangular wall (RW1) described in Section 2 was analyzed with NSC web and HSC boundary elements.
Friction-based contact was defined along the NSC web and HSC boundary elements for upper-bound and lower-bound values of the coefficient of friction, μ (1.0 and 0.6, respectively), per Table 22.9.4.2 of ACI 318-19 [
38]. The “Contact-Automatic-Surface-to-Surface” option of LS-DYNA, which uses a penalty-based approach to prevent slave-node penetration through the master surface, was used [
26].
6.1 Effect of cold-joints on load−displacement response
Analyses were carried out for the three ALRs for the upper-bound and lower-bound values of the coefficient of friction, and results for 20% ALR are shown in Fig.14. Other ALRs yielded similar results and are shown in Tab.8. NSC wall strength and stiffness are also shown as reference. Inclusion of contact between boundary and web areas of the wall resulted in a reduction in strength and stiffness. Furthermore, the behavior of the HSC walls for the upper-bound and lower-bound values of the coefficient of friction was very similar.
The values of the initial stiffness and peak strengths for NSC wall and HSC rectangular wall, cast monolithically and with a cold joint, are tabulated in Tab.8. It can be observed that there is 33% change in the initial stiffness and 9% to 11% change in peak strength due to the cold joint. This is attributed to the lack of a full-composite action between web and boundary elements due to the occurrence of sliding along the interfaces.
Drift ratios at which the cover-concrete spalled and boundary-element reinforcement yielded are listed in Tab.9. Cover-concrete spalling and boundary-element reinforcement yielding were delayed when friction contact was considered between boundary elements and web. For example, for the HSC wall with cold joints and 10% ALR, spalling strain in cover concrete of the boundary elements was reached at a drift ratio of 1.58%. However, because of the loss of full composite action, the web corner, having NSC, reached its spalling strain earlier at a drift of 1.10%.
6.2 Effect of cold-joints on principal strains
The principal compressive strain contours for the HSC wall—monolithic and with cold joints—for 20% ALR at drift ratios of 0.5% and 2.0% are shown in Fig.15. In the figure, the contours shown are only for the upper-bound value of the coefficient of friction (µ = 1) because the upper-bound and lower-bound results were similar.
It can be observed from the contours for principal compressive strains that there is significant change in the way strains are dispersed across the wall. Wall with the cold joint experienced high principal compressive strains not only at the outer edges of the boundary elements, but also the inner edge of the web and its diagonally opposite corner. Visible struts were formed in the web, which explains the higher straining along the diagonal and corners of the web.
The principal tensile strains are shown in Fig.16. It can be observed that the webs experienced greater principal tensile strains when cold-joints were considered. These strains decreased as the ALR was increased, and were particularly localized near the horizontal reinforcement, which can be observed as conspicuous streaks of high-strain lines at nearly regular intervals in the web along the wall height.
The results indicated that the potential strength and stiffness enhancement due to HSC boundary elements can decrease if sufficient means of ensuring shear transfer between the different parts are not provided. Shear can be transferred at the interface by methods such as additional shear reinforcement, lateral post-tensioning, or shear keys. Ensuring proper shear transfer across the interface is even more important for the flexural design of HSC walls. If there is sufficient interface shear strength, the HSC walls can be designed using conventional fiber-section analyses. These section level analyses can be used to determine the required reinforcement and flexural capacity of HSC walls.
7 Summary and conclusions
In this study, the prospective use of HSC in slender RC walls was presented for improved performance. HSC was placed in only the boundary elements of shear walls (high flexural demand regions) for cost-efficiency. These walls were termed as HSC walls in this study.
A NSC slender rectangular wall (NSC wall) was modeled using the finite element (FE) method. The FE model was validated using test data, and was used to investigate the behavior of HSC walls. The HSC used in this study had a mean cube compressive strength of 148.9 MPa. All analyses were performed for a set of three ALRs––10% (400 kN), 20% (800 kN), and 25% (1000 kN) of the axial compressive strength of NSC wall. The main conclusions of the study are listed below.
1) The FE model was able to capture the global force-displacement behavior of the wall to a reasonably good extent. Input parameters that required calibration were the concrete tensile strength and modulus of elasticity.
2) The attainable increase in the peak strength due to the replacement of the NSC wall with the HSC wall was between 11% and 14%, and the increase in stiffness was 56% for the three ALRs. The increase in drift ratios for boundary-element concrete spalling due to HSC boundary elements was between 42% and 56%. The yielding drift ratios, however, were lowered by 7% to 10%. The HSC wall experienced higher tensile strains than the NSC wall. For HSC walls, the tradeoff between higher strength, higher stiffness, delayed spalling and earlier yielding with higher tensile strains should be discussed early in the design process among the potential stakeholders.
3) Whether the longitudinal-reinforcement amount in the boundary elements can be reduced by taking advantage of increased strength of HSC walls was also investigated. For the walls that were investigated, the longitudinal-reinforcement amount could be reduced by 28% to 41% for varying ALRs. Higher ALRs allowed reducing longitudinal reinforcement amounts in HSC boundary elements. The lateral drift ratio at which boundary-element concrete spalling occurred increased by 45% to 73% due to HSC. The drift ratio at which reinforcement yielded was reduced by 15% to 32% due to HSC. This was due to the reduced reinforcement of the HSC wall.
4) Suitability of using rectangular HSC walls as replacements of barbell-shaped NSC walls was assessed. Rectangular HSC walls exhibited similar strength as barbell-shaped NSC walls when boundary element area was 2.25 times that of the rectangular wall. The ratio of peak strengths and stiffness of barbell-shaped NSC walls to rectangular HSC walls was 0.98 and 0.90, respectively. Principal compressive strains were less severe for rectangular HSC walls when compared to those for the barbell-shaped NSC walls. Principal tensile strains of the rectangular HSC walls were slightly higher than and comparable to those of barbell-shaped NSC walls at lower and higher drift ratios, respectively. The HSC wall experienced spalling at drift ratio that was 28% higher than that for NSC barbell walls. Reinforcement yielding for HSC wall was observed at drift ratios 4% to 8% lower than those for NSC wall.
5) Cold joints, which occurred due to dissimilar concrete mixes in the web and boundary elements for HSC walls, were studied. Friction-based contact was defined at the joint interface, which led to a reduction in strength between 9% and 11%, and stiffness loss of 33% when compared to monolithic walls. Wall behavior was not sensitive to the friction coefficient at the cold joint interface. The weakening of composite action between web and boundary elements was observed, and this led to the formation of diagonal “struts” and “ties” within the wall web, with boundary elements acting as vertical frame elements and foundation and loading beam acting as horizontal elements.
6) It was observed that in order to make full use of the possible enhanced strength and stiffness of the HSC wall, it was essential to ensure that the entire wall system acted as a monolithic unit. This can be realized by providing additional horizontal reinforcement, shear keys or shear studs along web-boundary element interface, or by using lateral post-tensioning.
PDF
(18833KB)
