To view the figures and tables associated with this article, please refer to the flipbook above.
Past earthquakes in California have shown that some concrete buildings, with certain design and detailing characteristics, perform poorly in large earthquakes and have long been known to present threats to life-safety. This undesirable performance is mostly attributable to shear-controlled strength of certain structural elements: pure frame structures lacking joint reinforcement and sufficient column ties (Fig. 1) and captive short columns or the influence of non-structural components on the structural system (Fig. 2).
As a result, the cities of Los Angeles, Santa Monica, and West Hollywood enacted mandatory retrofit ordinances for all concrete structures (except detached single family homes and duplexes) built with an application or using a building code edition enacted before January 1977. To quote the Los Angeles Ordinance (with dates added for historical purposes), the justification for the ordinance is:
In the Northridge Earthquake (1994), many concrete buildings constructed prior to the 1976 Los Angeles City Building Code provisions performed poorly and collapsed, causing loss of human life, personal injury and property damage. Similar loss of life, injury and property damage occurred in the Great Hanshin Earthquake (1995, often called the Kobe Earthquake), Mexico City Earthquake (2017 or possibly 1985), and Christchurch New Zealand Earthquake (2011) due to the outdated building codes in those communities.
Like most of the other retrofit ordinances discussed in this series of articles, the first phase involves an evaluation of the structure followed by a longer time frame for retrofit or demolition as the second phase. In the case of Los Angeles, the first phase is three years long and starts when the building department first notifies the building owner that their building is subject to the ordinance. This is called “the order.” The second phase is 25 years long dating from the order.
Buildings are subject to retrofit unless the seismic evaluation can demonstrate that:
The strength of the lateral-force resisting system meets or exceeds 75% of the base shear specified in the current seismic provisions of the Los Angeles Building Code (LABC) (basically ASCE 7 or the 475-year return period earthquake). Elements not part of the lateral-force resisting system shall be adequate for gravity load effects and seismic displacement due to the full (100%) of the design story drift specified in the current LABC. This is generally considered to be a collapse prevention performance goal.
Meet or exceed the requirements specified for “Basic Safety Objectives” (BSO) using ground motions and procedures established by the LA Building Department based on ASCE 41. For those not familiar with ASCE 41, the BSO has two thresholds, one for life-safety and the other for collapse prevention.
The breadth of this ordinance is very broad and rivals that of the retrofit ordinances for URM buildings and concrete tilt-ups during which past California earthquakes have shown as unacceptable hazards. It is unusual, however, to use the earthquake performance of concrete buildings constructed in other countries designed using non-U.S. building codes and constructed using non-U.S. practices to justify our retrofit ordinances.
San Francisco is currently considering enacting a concrete building retrofit ordinance. In mid-April 2024, after many years of debate, the city published a document titled Stakeholder Engagement for the Concrete Building Safety Program (CBSP) which discusses concrete building safety, various stakeholder perspectives and how a seismic retrofit program would be enacted, starting with developing an accurate inventory and completing a screening checklist prepared by the owner’s structural engineer. The CBSP if enacted in an ordinance will address concrete tilt-ups (discussed in Part 3 of this series and long overdue for San Francisco) and older concrete buildings.
Concrete buildings are not all the same when it comes to seismic performance. History has shown that “pure” frame buildings with widely spaced ties in the columns or those with deep spandrel beams and short captive columns are the worst performers. The most contentious as to whether they should be included or not were apparently the many mid-rise apartment buildings (Fig. 3) dating from the 1920s through 1940s with complete gravity frames (structural steel or concrete), wood or concrete floor diaphragms and “solid” non-bearing perimeter concrete walls (usually 6-inch thick over the full height with light reinforcing) pierced for windows. In some buildings, the perimeter steel frames are also detailed with “wind frame brackets” in the beam-column joints. Many engineers would argue that this second group of concrete buildings haven’t been shown in past earthquakes to perform unacceptably as a class of structure and therefore don’t warrant evaluation or inclusion in a retrofit ordinance. Post-elastic analysis, involving considerable engineering effort, is obviously required to truly understand how these structures will actually perform in major earthquakes mainly because US buildings of this class haven’t been tested in a really large earthquake.
Analysis of the shear walls in this class of structure usually shows that the wall piers are shear-controlled (very lightly reinforced) and lack sufficient shear strength using common acceptance criteria. But not all of these buildings are the same. Most engineers would argue that the ones with steel gravity frames are superior to concrete ones, concrete diaphragms are superior to wood diaphragms even though heavier, and the presence of embedded wind-frames is a superior feature since they create a secondary line of seismic resistance and contribute to structural stability should the concrete shear walls be overloaded. The CBSP lists a number of characteristics which exempts a building from the program, with buildings having steel gravity frames as one.
Steel Frame Buildings With Pre-Northridge Welded Moment Frame Connections
The 1994 Northridge earthquake was the first earthquake where a significant number of welded steel moment frame connections, with fully welded flanges and bolted webs, sustained considerable damage, mostly cracking in the beam flanges that sometimes propagated into the column. Metallurgists and fracture mechanics experts opined that these types of connections would almost inevitably failure due to the three-dimensional nature and stress flow in a WF beam to WF column arrangement. In any event, most of the profession was caught entirely off guard.
In a nutshell:
- It was common practice to leave the backing bars for the complete penetration welding of the beam flanges in place after the welding was complete. This created a discontinuity or “crack starter” at the root of the weld that could produce a crack when the beam was bent under high load.
- It was more common for cracking to occur in the bottom flange of the beam because the “crack starter” was at the extreme fiber subject to the greatest tension stress. This is not the case at the top beam flange since the flange welds were usually made in the downhand position. The crack starter is in a lower stressed location. In buildings with concrete fill on metal deck floors, the floor slabs resisted some of the bending in the beam-column joint due to composite action, which also reduced the demand on the top beam flange.
- The original research and testing of the welded connections employed smaller beam sections that happen for various reasons to perform better. The research was assumed to apply to larger beam and column sections when the prescriptive building code provisions were drafted. To the extreme dismay of some engineers, the continued use of traditional bolted-flange moment frame connections had to be justified.
- Steel frame system design was economized, from moment connections at all beam-column joints, to moment connections only on the perimeter of the building (Northern California engineers fairly or unfairly called these “LA frames”) and finally to only certain bays on the perimeter, sometimes only one bay on each side.
After the earthquake, the profession, university researchers and the steel industry conducted extensive investigative research that resulted in new code provisions, new types of moment connections and methods for retrofitting.
But the question remained, since to date none of these buildings has collapsed in an earthquake, as to whether retrofitting pre-Northridge steel moment frame structures is necessary. The retrofits are not particularly difficult (removing the back-up bar, back-gouging the weld, and adding a protective cover fillet weld) but the disruption to building tenants and the cost is high. So the answer is “it depends.”
The factors that need to be considered are:
- What criteria will be used to evaluate existing structures?
- What type of analysis can be used to exempt an existing structure?
- What part can or should redundancy play? How many connections are there in the building and how many can crack in some form without compromising the entire structure?
- How do we deal with the myriad number of member sizes in terms of predicting performance – smaller beams and columns, larger, deeper beams and columns, a mix of both?
- Since weld fractures are brittle by nature, and therefore random, how do we predict which ones will fracture?
- Should the column splices be evaluated too? Column splices often relied on partial penetrations welds. What are the implications of a column splice failure?
As you can see, any ordinance short of mandating that all welded connections be retrofitted is going to be complicated. It is no wonder that pre-Northridge steel moment frame structures have been left to last in line. ■
This is the final installment of a four-part series. Earlier installments were published in the February, March, and April issues of STRUCTURE.
About the Author
John A. Dal Pino is a Principal with Claremont Engineers, Inc. in Oakland, California. He serves as the Chair of the STRUCTURE Editorial Board.
(jdalpino@claremontengineers.com).