Two major events have caused us to question the risks for major high-rise buildings. Structural engineers John Meyer and Ronald Hamburger explain.
As a society, we have come to expect low-rise commercial and residential construction, particularly older examples, to experience large life and economic losses as a result of occasional extreme natural and human-induced events. However, until recently there has been almost a complete lack of significant damage to major high rise buildings, and as a result, society had come to believe that these buildings were not at significant risk from such perils.
In January 1994, the Northridge, CA, earthquake caused over $30bn of damage, including extensive, previously unanticipated and potentially life-threatening damage to modern commercial office structures in the Los Angeles region. On 11 September 2001, terrorists graphically demonstrated that even the most monumental of our structures are vulnerable to extreme, and sometimes complete, damage.
Both events have caused building owners, insurers, investors and tenants to question the safety of these buildings, as well as the economic risks associated with their construction and occupancy. In turn, this has prompted structural engineers to find better ways to design buildings to meet the performance expectations of the broader community.
One of the most significant impacts of the 1994 Northridge earthquake was unexpected brittle fracture of critical framing connections in a number of modern, steel high-rise buildings. All the affected buildings were of a type of construction known as steel moment frames that relies on the rigid interconnection of beams and columns to resist high wind and seismic forces. Previously, most engineers considered steel moment frames to be one of the most reliable earthquake-resisting systems, and as a result, they were widely used in major buildings, in both seismic and non-seismic zones. In fact, more than 75% of modern US high-rise construction employs this structural system.
The discovery of brittle fractures in these structures, once thought to be nearly invulnerable, created a crisis of confidence among design professionals as well as the emergency response, insurance and financial communities. The steel construction and structural engineering communities organised a large joint venture project known as the FEMA/SAC program to reduce seismic hazards in steel moment frame buildings to understand the causes of these failures and to develop practical design solutions.
The Federal Emergency Management Agency (FEMA) generously funded this effort and in July 2000 published its findings and recommendations in a series of engineering guidelines documents. These documents update the basic building code requirements for these structures and also provide a performance-based methodology enabling engineers to design for the specific levels of damage that are tolerable to the building owner and other stakeholders.
Since the publication of these documents, the building codes have adopted many of the recommendations, and most steel moment frame buildings in regions of high seismic risk in the US are now designed according to the methodologies contained in these documents. Most engineers believe adherence to these norms will preclude significant fracture problems in steel moment frames in future earthquakes.
While the steel moment frame problem was serious and costly, most damage costs in the Northridge earthquake were distributed throughout a wide variety of other types of building construction, including wood frame, masonry, concrete and non-moment frame steel structures and also included non-building structures such as bridges and utilities. Other code changes have been developed and adopted to address many of the issues this damage highlighted.
Concurrent with the development of these technical code changes, the attitude of the real estate and insurance industries to seismic risk also changed. The large losses from the Northridge earthquake caused great concern among insurers, lenders and owners of large portfolios of buildings in areas of high seismicity. Many such stakeholders have responded by retaining structural engineers to perform seismic risk evaluations, either of individual buildings or entire portfolios of buildings, to better understand their seismic risk.
The most commonly performed type of risk evaluation is a probable maximum loss (PML) estimate. The PML is an estimate of the likely cost to restore damaged buildings to pre-earthquake condition, excluding damage to contents. Frequently, the PML is expressed as a percentage of building replacement cost (not including land) and is generally computed for a specific seismic event or `scenario', such as a ground motion at the site with a 10% probability of being exceeded in 50 years.
Although the exact definition of PML varies by industry practice, it is typically expressed as that amount of damage that will not be exceeded in nine out of ten similarly constructed buildings subjected to the same scenario seismic event. Many lenders require, as a precondition of loan commitment, that the PML computed in this manner is less than 20%. The lenders selected this limit in the belief that buildings damaged to such an extent would likely be repaired and restored to service with only limited business interruption, and that their owners would be able to continue to service the loans, while owners of more heavily damaged structures could be a significant lending risk. Anticipating this requirement, some developers have commissioned new buildings designed to conform to such a PML rating. Other building owners have commissioned seismic upgrades of buildings to achieve that performance.
Based on the authors' experience as earthquake engineers, the practice of scrutinising buildings with a PML analysis during lending and sales transactions has been an important stimulus for seismic retrofit of the private sector building stock. This is true in spite of the fact that PML analyses are based on insufficient data on damage distribution from real earthquakes, and suffer from a lack of uniformity in the means of computation and even in the terminology used to discuss them.
The typical PML analysis should be considered a relative measure of probability for damage of a building or portfolio of buildings as compared to other buildings or groups of buildings. Most PML analyses are of little value for such things as financial cost/benefit studies and life cycle analyses. Usually these PML analyses do not account for total losses, including business interruption, decreased market value, lost market share, insurance deductibles and limits, and other costs associated with earthquake-induced damage.
The structural engineering community has responded to the heightened interest in designing buildings to achieve specific performance with the development of performance-based designs (PBD). With PBD, the stakeholders in the design of a building can choose the performance they want to achieve for various intensities of earthquake ground motions or scenarios. This is in direct contrast to the prescriptive design requirements of current codes, which seek primarily to protect lives, not to avoid losses to property or safeguard economic value. Considerable damage and business interruption can occur in a code-designed building if it is subjected to the code level ground motions. Unfortunately, many owners - and even some structural engineers - incorrectly believe the codes protect their investment and business interests.
In performance-based design methods, the engineer must recognise the potential building damage mechanisms, such as various types of cracking, and attempt to predict the level of ground motion that will cause such damage to a particular design. Engineers can then provide owners with choices for performance levels (expected damage and repair cost, business interruption, etc.) for various magnitudes of earthquakes, that potentially affect their site. The cost of designing to specific performance capability can then be weighed against the avoided costs associated with poor performance to achieve an optimal design solution. In the future, the PML and PBD technologies will be better developed and integrated to help owners and engineers make better decisions concerning appropriate risk objectives for various buildings.
As the September 11 tragedy has shown, buildings are susceptible to other risks. These risks can take the form of large aircraft turned into weapons, but also can be in the form of blasts, debris from an adjacent incident or even fire. Fire is not a new peril and, in fact, modern building codes were developed specifically to reduce the risk of fire-induced life loss and structural collapse. Prior to September 11 no modern, fire-protected steel frame structure had suffered fire-induced collapse, even though there had been numerous very large fires in such buildings.
Just as with earthquake resistance, the design community believed that modern high-rise buildings were invulnerable to fire-induced collapse. However, both of the twin World Trade Center towers collapsed not because of the aircraft impact damage, but fire. An adjacent 47-storey building, WTC-7 also collapsed as a result of fire on that day, while WTC-5, a nine-storey building experienced a large, fire-related partial collapse over several floor levels.
Design of fire protection for structures, today, is commonly done without the participation of the structural engineer. September 11 demonstrated that the prescriptive procedures for design of fire protection systems, routinely used in the past, are inadequate to provide the levels of protection appropriate for some buildings.
The twin towers failed in a mode of behaviour known as progressive collapse. Similar behaviour occurred in the 1995 bombing of the Alfred P Murrah building in Oklahoma City. Such complete collapse conditions can also result from a number of inadvertent events, including vehicle impact, blast and fire. Presently, US design codes commonly used for commercial buildings do not include provisions to protect buildings against blast or progressive collapse. Some European and Asian codes do, and the design of certain US government buildings also requires considerations for blast loads and progressive collapse. These risk reduction measures are available to private owners with concern for such risks.
Design and retrofit of buildings to resist progressive collapse is relatively costly and probably not justified for most buildings. Therefore, most private sector response to this risk has included such measures as erecting barriers to protect against bomb-laden vehicles, toughening walls and increasing security.
The kind of concerted response effort resulting in research and design recommendations that resulted from the Northridge moment-frame problem has not materialised to date.
That said, the engineering community and the building industry are much more aware of blast and progressive collapse issues than they were before September 11. Eventually, government agency-funded studies into blast resistant structures and progressive collapse will filter out to general building design usage and provide practical tools for routine application to commercial buildings.
Meanwhile, here are some observations shared by the authors and their colleagues:
Most terrorist attacks target individual buildings or groups of buildings. The risk is low that any particular building will be attacked, because there are many buildings and they are widely spread geographically. Of course, if terrorists use weapons of mass destruction, the geographical extent is considerably widened, but structural improvements in design are probably moot in this type of scenario.
On the other hand, earthquakes affect large populations of buildings. The probability of the events is high in high seismic zones. Even in areas of moderate seismicity, the probability of damage due to earthquakes may be significantly higher than damage due to terrorist attacks.
While engineers consider the details of how to design more blast resistant and fire resistant buildings, all the stakeholders must address the broader issue of how much money will be allocated to reduce the risks and where the money will be directed. The decisions will go beyond estimates of annualised risk of death and economic loss. Unfortunately, code writing is predominately performed by the engineering community. Other stakeholders need to be involved in the process as well.
By John Meyer and Ronald Hamburger
The authors are structural engineers at Simpson Gumpertz & Heger, San Francisco, CA. John Meyer, is vice president and principal and Ronald Hamburger is principal. Email: JDMeyer@sgh.com