Performance-Based Evaluation of RC Beam-Column Joints

Reinforced concrete (RC) buildings that were built in the 1960s and do not meet current design criteria have limited ductility. In this article, beam-column joints of such nonductile buildings are investigated using several performance-based criteria. Four half-scale RC exterior joints were tested to investigate their behavior in a shear-critical failure mode. The joints were subjected to quasi-static cyclic loading, and their performance was examined for lateral load capacity, ductility, drift, axial load-bearing capacity of the column at high levels of drift, joint shear strength, shear angle, and residual strength. The influence of the axial load applied to the column on the joint shear capacity was investigated. The exterior joints tested in this study reached approximately twice the joint shear strength assigned to joints of this geometry by FEMA 273 (BSSC 1997).


Reinforced concrete frames can achieve ductile behavior provided that brittle failure of structural elements and instability can be prevented in severe earthquakes. The design and detailing of beam-column joints is important in achieving satisfactory performance of RC frames. The design should be able to (1) prevent brittle shear failure of the joint, (2) maintain integrity of the joint so that the ultimate strength of the connecting beams and columns can be developed, and (3) reduce joint stiffness degradation by minimizing cracking of the joint concrete and by preventing the loss of bond between the concrete and longitudinal beam and column reinforcement. Joints in existing structures built before the development of current design guidelines such as ACI 352R-91 (1991) do not conform to the current requirements (ACI SP-123 1991). The research described targets the performance of exterior joints in existing RC frame structures in order to establish their adequacy in terms of performance-based criteria. Performance-based earthquake engineering is a PEER focal area for which the objective is to build better and more economical structures (Krawinkler 1999). Specifically, the influence of the column axial load on the shear strength of the joint is studied, and comparisons of the shear angle with established FEMA 273 (BSSC 1997) guidelines are presented.

Specimens and Experimental Setup

As is typical of buildings built in the early 1960s, the beam-to-column connections targeted in this research lack confining reinforcement in the joints and have insufficient anchorage of reinforcement extending into the connections. The specimen dimensions and reinforcements for the joints studied are shown in figure 1.

Fig. 1. Specimen dimensions, reinforcement details, and setup

There is no transverse reinforcement within the joint core, and the beam longitudinal bars are not adequately anchored in the connection. In addition, the lap splice length is insufficient, the column depth-to-beam bar diameter ratio is less than 20, and the details of the transverse reinforcement do not meet the ACI 352 criteria. The average concrete strength of the specimens was 5960 psi (41 MPa). The yield strength of the rebar was 62 ksi (427 MPa) and 67 ksi (462 MPa) for the transverse and longitudinal reinforcement, respectively. The steel ratio for both tension and compression steel
= 2.47% barely satisfies the ACI code maximum and was chosen to force a shear mode of failure in the joint.

A schematic of the test setup is shown in figure 1. The column was mounted horizontally, and an axial load equal to 0.1f 'cAg for two of the specimens and 0.25f 'cAg for the other two specimens was applied using a small hydraulic cylinder. The compressive axial load was transferred to the column portion of the specimen through four threaded rods as shown in figure 1. The compressive axial load was set to an initial value and was then left to change at will, as the beam was subjected to load reversals. The lateral load was applied cyclically, in a quasi-static fashion, at the end of the beam through a loading collar.

Experimental Results

The first portion of the test was load-controlled wherein the lateral load was increased in 5 kip (22.2 kN) increments. At every load step three cycles were performed, each cycle containing a push and pull segment. After the first yielding of the reinforcement, the testing was carried out using displacement control. Three cycles were performed at each displacement step, and the displacement was increased as a fraction of the initial yield displacement. The test continued until the lateral load dropped below 50% of its peak value. A typical moment versus shear angle curve is shown in figure 2 for test 2 which had an axial load of 0.1f 'cAg.

Fig. 2. Hysteresis response of the joint and FEMA 273 modeling parameters

The first yielding occurred in a longitudinal column bar at a lateral load of 27 kips (120 kN) and a displacement of 0.3 in. (7.62 mm); the joint ultimately failed at a displacement of 1.94 in. (49.3 mm), at which point extensive shear cracking in the joint region was observed along with spalling of concrete on the back of the column at the joint. Large x-shaped cracks were formed, and a significant crack which originated from a diagonal crack in the joint extended up the column along a longitudinal column bar. The cracks reached 0.2 in. (5 mm) in width, with the largest cracks residing in the joint. In two of the tests with the higher axial load of 0.25f 'cAg, buckling of the longitudinal column bars was observed in the joint region. A typical shear failure in the joint region is shown in figure 3.

Fig. 3. Condition of the joint at ultimate drift and axial load in the column

It should be noted that the joints tested do not qualify as either Type I or Type II joints of ACI 352 (91) — the reinforcement is of Type I but the loading is of type II. FEMA 273 (1997) classifies this as an esterior joint without transverse beams and with no transverse steel. For that classification, FEMA 273 specifies = 6,
where psi, Vn = joint shear strength, and AJ = effective horizontal joint area. The joint strength coefficients obtained from the experiments are shown in table 1.

Table 1.

The average value for the specimens with a column axial load of 0.1f 'cAg is = 12.4. For the joints with 0.25f 'cAg axial load, the joint strength coefficient is = 13.4. These tests indicate that the joint strength coefficient changes with the variation of the column axial load and that the FEMA 273 guidelines are conservative in this case.

The variation of the compressive axial load in the column due to the cyclic lateral load applied at the end of the beam is shown by the plot of axial column load for test 2 in figure 3. This behavior is typical of all joint specimens that were tested. The original axial load was set at 0.1 f 'cAg which corresponds to 155 kips (689 kN), as shown in the figure. At the peak lateral load, the specimen had a 2.6% drop in axial load; at 80% of the peak lateral load it had a drop of 3.9%; and at the ultimate drift the axial load degradation was approximately 10%. Therefore, the axial load capacity of the column is reduced as the joint strength degrades. The column axial load degradation at the ultimate drift varied from 10% to 24%. In addition, for the two specimens with the 0.25f 'cAg axial load, the longitudinal column bars buckled in the joint region.

A comparison was made of the test results with the FEMA 273 (BSSC 1997) modeling parameters for reinforced concrete beam-column joints. Specifically, the shear angle and residual strength ratio as defined in figure 2 were compared to table 6-8 of FEMA 273. FEMA 273 modeling parameters are for an axial load ratio of 0.10. Table 2 shows that for an axial load ratio of 0.1 f 'cAg the Guidelines are conservative by a significant margin.

Table 2.



In addition, the joint specimens with the lower axial load ratio of 0.1 f 'cAg are more ductile than the joint specimens with axial load ratio of 0.25 f 'cAg.


The tests have shown that the joint strength coefficient () changes with the variation of the column axial load, and that FEMA 273 is conservative in this regard. The tests have also shown that the applied column axial load is reduced as the joint is progressively damaged; the column axial load degradation at the ultimate drift level varied from 10% to 24%. For the two specimens with the 0.25f 'cAgaxial load, the longitudinal column bars buckled in the joint region. The results indicate that the FEMA 273 (BSSC 1997) modeling parameters for seismic rehabilitation are conservative. Joint substructures with an axial load ratio of 0.1f 'cAg were more ductile than those with an axial load ratio of 0.25f 'c Ag.



Funding for this research was provided by the Pacific Earthquake Engineering Research Center under Grant No. SA1810JB. Support from PEER and in-kind contribution from Eagle Precast Co. are greatly appreciated. The findings and conclusions in this paper are preliminary and do not represent the views or policies of the sponsors.

Chandra Clyde, Graduate Student Researcher
Professor Chris P. Pantelides, and
Professor Lawrence D. Reaveley
Department of Civil and Environmental Engineering
University of Utah, Salt Lake City


ACI Committee 352. 1991. Recommendations for design of beam-column joints in monolithic reinforced concrete structures.Farmington Hills, Mich.: American Concrete Institute. Report 352R-91.

ACI SP-123. 1991. Design of beam-column joints for seismic resistance. Farmington Hills, Mich.: American Concrete Institute. Report SP-123.

BSSC 1997. NEHRP Guidelines for the seismic rehabilitation of buildings. Washington, D.C.: Federal Emergency Management Agency. Publication 273.

Hose Y. D., and F. Seible. 1999. Performance evaluation database for concrete bridge components, sub-assemblages and systems under simulated seismic loads. /PEER/Paper1/paper1.html.

Krawinkler H. (1999). Structural engineering issues in performance based earthquake engineering. PEER White Paper.