Behavior and Efficiency of Bottle-Shaped Struts. Paper by Michael D. Brown, Cameron L. Sankovich, Oguzhan Bayrak, and James O. Jirsa/AUTHORS' CLOSURE

(ProQuest-CSA LLC: ... denotes formulae omitted.)

The authors have presented an interesting paper. The discusser would like to offer the following comments:

Based on Fig. 4 through 6, the discusser believes that the bond characteristics of reinforcing bars would have a significant impact on the compressive strength of bottle-shaped struts. Therefore, to account for a bond characteristic of reinforcing bars Watanabe and Mugumura7 have proposed a modification to the Vecchio-Collins equation, which is expressed as follows

... (1)

or

... (2)

where C^sub Bond^ equals the bond coefficient (equal to 1.00 for good bond, 0.25 for poor bond).

The discusser is somewhat confused about β^sub s^ (efficiency factor) values stated in Table 1 and Eq. (A-3) of ACI 318-05 and the v value (equation below Table 2, efficiency factor).

According to Eq. (A-3) of ACI 318-05, f^sub cu^/f'^sub c^ = 0.85β^sub s^. While the authors' equation (below Table 2), v = Peak load/(f'^sub c^ × bearing area) = 0.85β, or v = 0.85βf'^sub c^. This means that ?/f'^sub c^ = 0.85β.

When v value is considered as f^sub cu^, the aforementioned equation becomes similar to Eq. (A-3) of ACI 318-05, it is also noted from Table 2 that the mean value of v is approximately equal to 1.0, then, if β could be considered as 1/f^sub cu^, then f^sub cu^/f'^sub c^ = 0.85.

This means a reduction of concrete compressive strength of 15%, which is inconsistent with the German, Japanese, and Canadian test results. It was also noted from Table 2 that test Specimens I and J have identical parameters, but the v values are significantly different. A similar condition was noted in Specimens M and N and Specimens W and X. A significant difference was also noted in Specimens A and B. In fact, Specimen B should have higher v value than Specimen A, but it has a lower value than Specimen A.

The discusser hopes, and would greatly appreciate, that the authors would enlighten their concept used in this paper.

Though the discusser can interpret the peak load from the equation below Table 2 and v value from Table 2, these value would be approximate. To analyze the failure load (peak load), it would be better to have an actual test value of failure loads so that analytical study/research can be verified. Therefore, the discusser would greatly appreciate if the authors could include their peak load/failure load values of their test specimens (that is, each panel).

REFERENCES

7. Watanabe, F., and Mugumura, H., "Ultimate Strength of R.C. Panels," Proceedings of the Sessions related to Structural Design, Analysis and Testing, A. H.-S. Ang, ed., ASCE Structures Congress, San Francisco, Calif., May 1-5, 1989, pp. 31-38.

AUTHORS' CLOSURE

The authors thank the discussers for their interest in the paper and critical evaluation of the work presented therein. Closures to the two discussions are presented.

In regard to the discussion by Sahoo, Singh, and Bhargava, each of the five points is discussed individually in the following:

1. The geometry of the isolated strut specimens was selected primarily for ease of construction and testing. The specimens were small enough to be moved and installed in the testing machine without the use of the overhead crane in the laboratory.

2. The nominal compressive strength of concrete was intended to be 3600 psi (24.8 MPa). The specimens were cast in multiple batches; thus, the compressive strength varied among the specimens. Even with the maximum compressive strength used in this study, the concrete would not be considered high strength. The authors do not expect that the stress-strain response of these concrete batches were significantly different.

3. The isolated strut specimen size was chosen to facilitate moving the specimens within the lab; the thickness of the panels was either 4 or 6 in. (100 or 150 mm). With specimens this thin, it was only possible to place a single curtain of reinforcement at mid-thickness. These specimens were meant to simulate the formation of bottle-shaped struts using a simple specimen. They were not meant to represent real structures with more complex boundary conditions.

4. The efficiency factors that were calculated from the failure loads are given in Table 2. The failure load can be obtained by multiplying the tabulated efficiency factor with the concrete compressive strength and the bearing areas given in Fig. 5. The values of failure load are tabulated in Table A.

5. The typical failure mode is shown in Fig. 7. In the upper portion of that figure, a possible strut-and-tie model for the test specimens is shown. The node-strut interface is highlighted in that drawing. The node itself is confined by friction between the concrete surface and the bearing plate. In the presence of confining pressure, the compressive strength of concrete increases. The end of the strut abutting the node gets none of the beneficial effects of confinement because it is outside the most highly confined zone. Hence, the location where the node and strut intersect has the minimum crosssectional area within the strut and has no additional compressive strength due to confinement.