Seismic Retrofit of Octagonal Columns with Pedestal and One-Way Hinge at Base

ACI Structural Journal, Sep/Oct 2005 by Johnson, Nathan, Saiidi, M Saiid, Itani, Ahmad, Ladkany, Samaan

EVALUATION OF RESULTS

Measured lateral load and deflection

The cumulative load-deflection hysteresis curves for all specimens are plotted in Fig. 13. The envelopes of the measured lateral load-deflection relationships are compared in Fig. 14. All of the curves were idealized as elasto-plastic relationships to compare yield point, stiffness, and column ductility.

The idealized yield force of OLVR-2 is within 10% of that from OLVA. Also, the yield force of OLVR-2 is substantially lower than that of OLVR-1, demonstrating that the column cut was effective. OLVR-1 is approximately 16% stiffer than the as-built column. The stiffness of OLVR-2 is approximately 18% less than that of the as-built column. The increase of stiffness for OLVR-1 was because of the larger and stiffer pedestal. The stiffness reduction of OLVR-2 is because of the cut section, and was responsible for the higher yield displacement that led to the somewhat lower value of ductility reached by OLVR-2 than that of the as-built. The measured as-built displacement ductility capacity was 6.9. However, the unacceptable pedestal cracks began at a ductility of only 1.5. The measured displacement ductility demands for OLVR-1 and OLVR-2 were 4.6 and 5.9, respectively. The cut retrofit of OLVR-2 showed a ductility improvement of 28% over OLVR-1. Had the OLVR-2 shear failure been postponed, and the column base been allowed to fail in flexure, a large ductility improvement would be seen.

Measured column curvature

A comparison of column moment-curvature for the two retrofitted columns is shown in Fig. 15. The curvature data were taken for the column section immediately above the pedestal over a 280 mm (11 in.) gauge length. Data for OLVA are not included in this comparison because the lateral displacements in OLVA were mostly due the pedestal damage and not flexural yielding of the column. It can be seen in the figure that OLVR-2, the column with the severed bars at the base, achieved much larger curvature values. The maximum curvature ductility at failure in OLVR-1 was 7.1 and in OLVR-2 was 11, which was higher by a factor of 1.5. Because of the severed bars, curvature in OLVR-2 was highly concentrated at the base of the column. The higher curvature ductility and reduced shear demand led to the increase displacement ductility capacity in OLVR-2.

To determine the effectiveness of the procedure to shift the plastic hinge from the base of the pedestal to the column-pedestal interface, the moment-curvature relationships at these locations were compared. Figure 16(a) and (b) show the results for OLVR-1 and OLVR-2, respectively. At failure the curvature ductility of the pedestal base was 1.4 in OLVR-1 and 1.2 in OLVR-2. These values were respectively 80 and 89% lower than the column base curvature ductility in OLVR-1 and OLVR-2. Although the extension of the pedestal and the additional reinforcement reduced the curvature ductility demand at the pedestal base, the retrofit did not eliminate yielding of the reinforcement. The limited yielding at the pedestal base did not seem to have any adverse effect on the overall performance of the specimens.