New PEER Report 2017/12: “Experimental Investigation of the Behavior of Vintage and Retrofit Concentrically Braced Steel Frames under Cyclic Loading"

August 9, 2018

PEER has just published Report No. 2017/12: “Experimental Investigation of the Behavior of Vintage and Retrofit Concentrically Braced Steel Frames under Cyclic Loading.” It was authored by Barbara G. Simpson, Stephen A. Mahin, and Jiun-Wei Lai.

Visit the PEER publications page to download a free color pdf of the document.

The parallel evolution of seismic design provisions and braced-frame research has led to inconsistencies between the design and construction of braced frames and the development of modern seismic design codes and now-typical detailing requirements. Since literature on concentrically braced frames (CBFs) spans over several decades, existing older or vintage concentrically braced frames–especially those designed prior 1988–may be prone to a number of deficiencies that are now limited in new CBFs due to contemporary seismic design requirements.

The number and range of these deficiencies and their likely interdependence, makes assessing the likely behavior of vintage braced frame systems problematic. Recent research has focused on improving the seismic behavior of modern braced frame systems, such as the Special Concentrically Braced Frame (SCBF). In contrast, relatively little research has focused on existing braced frames, even though vintage CBFs may be characterized by distinctly different behavior from modern SCBFs. Component tests of non-compact braces and connections and documented failures during past earthquakes have shown that vintage CBFs may be vulnerable to a number of complex damage states, including limited deformability and energy-dissipation capacity of the braces, potentially brittle connection failures, beam yielding in V- or chevron configurations, etc.

To improve this situation, experiments of complete sub-assemblages of vintage braced frame systems are needed to improve understanding of seismic response, assess the feasibility and efficacy of possible retrofit strategies, and calibrate computational models for future parametric studies. This report presents results of experiments and related analyses performed on vintage CBF specimens. Cyclic quasi-static tests were performed on three full-scale CBF specimens. A common two-story, one-bay configuration was adopted. The first specimen was representative of a pre-1988 CBF incorporating hollow HSS braces. The second specimen was similar, but the HSS braces were filled with concrete. The third specimen incorporated a mast (or strongback) retrofit and other features intended to mitigate the weak-story behavior observed in the first two specimens.

The first test structure utilized square HSS braces placed in a “chevron” configuration with one column oriented in strong-axis bending and the other in weak-axis bending. The first specimen was designed according to the 1985 Uniform Building Code; as such, it did not satisfy many requirements of current seismic design codes. These inadequacies were typical of vintage construction and included high brace width-to-thickness ratios, weak gusset connections lacking adequate yield-lines, weak beams designed without consideration of an unbalanced load that may arise due to brace buckling, and no capacity design considerations in proportioning members or connections. This specimen formed a weak story in the second floor, while the rest of the frame experienced only minor yielding and little permanent damage. Both second-story braces buckled–exhibiting considerable local buckling at the brace midpoint–and then fractured within a few additional cycles. Since the imposed story drifts were modest, the frame was subsequently repaired. The fractured second-story braces and gussets were replaced with the same sections.

The new braces in the “second” test specimen were filled with low-strength concrete in an effort to postpone brace local buckling and fracture observed during the first experiment. Net section reinforcement was also added at all the brace-to-gusset connections. Testing of the second specimen also resulted in a weak-story mechanism but in the bottom story. Delayed local buckling and subsequent fracture was observed in one of the bottom-story braces. After fracture of this brace, the frame tended to behave like an eccentrically braced frame (EBF) with a long link beam. This beam provided a relatively weak and flexible energy-dissipating mechanism. Many different local failure mechanisms were observed during subsequent loading cycles, including nearly-complete fracture at one column-to-baseplate interface, significant local buckling, and multiple connection weld and base metal failures.

The third specimen utilized a “strongback” (SB) retrofit aimed at alleviating the weak-story behavior seen in both the first and second experimental tests. The SB system employs a steel truss “backbone” that is designed to remain essentially elastic. This truss enforces similar drift demands in adjacent stories to delay or prevent weak-story behavior. The retrofit design was composed of two halves: an “inelastic” truss utilizing a buckling-restrained brace (BRB) that dissipated seismic input energy and an “elastic” vertical truss designed to control weak-story behavior. The specimen was successful in imposing nearly uniform drifts over the full height of the frame throughout the duration of the test. These preliminary experimental results show that the SB system can be an effective means in limiting weak-story mechanisms.

A number of numerical simulations were calibrated to the experimental results. These analytical models are capable of predicting the observed behavior. The models developed adequately simulated the observed brace global buckling, braces fatigue, column-to-baseplate fracture, and the overall global response of the test specimens.