Redundancy & Robustness in the Design of Tall Building Structures: A Panel Discussion

Structures Congress 2005

New York Hilton & Towers

New York, NY

April 22, 2005

8:30-10:00 am

Panelists (left to right): Robert J. McNamara, Akira Wada, Joe Burns, Craig Gibbons and Bill Baker.

Abstract:

The design of structural systems to resist lateral loads and accommodate gravity loads has formed the cornerstone of structural engineering, with innovative configurations propelling structures to unprecedented heights. While the issues of redundancy in structural design, at least at the sub assemblage level are well understood, there has been little formal discussion in the community regarding the role of redundancy and robustness on a system level, i.e., the ability of structures to provide alternate load paths and redundancy in the event of loss of a portion of the system. Unfortunately, these issues have become more critical in the wake of 9/11, prompting the Tall Buildings Committee of SEI to propose this panel discussion. The committee believes that such a discussion will be very well received and would highlight design philosophies appropriate for future developments and reassessment of the high profile structures.

Summary of Comments:

[AK: Ahsan Kareem, WB: William Baker, CG: Craig Gibbons, RM: Robert McNamara, JB: Joe Burns, AW: Akira Wada, AUD: Audience]

AK: Introduction of topics related to robustness and redundancy. Quotation on redundancy by Fazlur Khan. Introduction of panel members (as shown to the left).

[Presenters, in most cases, then stepped through a brief slide show of projects related to the topic.]

WB: Would like to see more work in research on reliability of systems in academics. Presented a number of case studies related to the WTC complex. Case study of WTC-3 a structure that withstood the impacts from the collapse of the main towers. Also showed WTC-4, example of building not designed in any way for what it ended up carrying. There is much to be learned from how seemingly “ordinary” buildings as part of that complex performed under such extreme conditions. These structures were exposed to very extreme loadings and managed to prevent progressive collapse.

CG: Admittedly cannot design for every unforeseen events. Essence of robustness: can building prevent collapse disproportionate to the cause. Do the codes address this? No. Many codes assume a static condition. Area of focus is dynamic robustness: how components can absorb energy to arrest failure. Of course designs are checked to determine what happens if key elements are removed, provision for alternate load paths. If alternate load paths are not provided, then the design of these elements should be supercritical. Also buildings that have to be designed for natural extreme events are better prepared for the unforeseen events. This should be considered.

JB: UK had a big response to Ronan Point collapse (UK Building Regulations A3). Very specific regulations surfaced in the 1960s in response to this. These codes require tying elements, defined disproportionate collapse criteria and loadings on key elements. NYC code has alternate load paths, specific resistence criteria. Chicago code has a generic statement (philosophy) on this issue but with no specific criteria. These issues are not embedded in the educational system in the US as it may be in other countries. Federal government buildings are requiring bridging and extensive retrofits. In high-rise systems, what system offers better reliability? Braced core and concreter core systems: lateral systems are protected because they are in the interior. Braced and tube are very redundant with alternate load paths. However the lateral resistance system is exposed on exterior, opening to potential collapse. Combined systems can provide alternate paths while not relying on either a completely interior or exterior system.

AW: WTC teaches us that local failures can produce progressive collapse. Japanese approach evaluates resistance to loss of main structural members to prevent total collapse; also evaluated the effect of arrangement of vertical elements. Another group in Japan is looking at resistance to large scale fire; important issue in Japan due to fires after earthquakes. Looking carefully at fire resistance. Japanese use of rigid connections is standard practice. Showed nonlinear frame collapse analysis when large number of columns are removed from the bottom floor. Evaluated several system types to determine the column axial loading utilization ratio when various members were removed. Similar studies were conducted to determine effect of fire at different locations in a building. Proposes the axial load utilization ratio of column as an effective parameter in study of prevention of progressive collapses. This ratio is the actual axial load normalized by axial capacity.

RM: Spoke on the inherent resistance of conventionally-designed buildings for which specific blast or progressive collapse have not been included in the design. Issues of continuity and alternate load paths are fundamental. All designs are inherently redundant but these alternate load paths simply do not have the capacity to handle the redirected loads. Bracing systems are inherently vulnerable. Looked at conventionally designed 40 story building and what happened when columns were removed. Connections are key. East coast we design connections for wind. The wind demands dictate the connection strength. Though demands for redundancy are the same on two directions of a building, the connections may differ due to wind demands. Connections are a key element in redundancy. Did an incremental “push down” analysis in ETABS when column(s) was (were) removed.

[Floor was then opened to the audience.]

AUD: Are the loads we are taking on really expected? Are there any loads that can be anticipated? Could this change the design from "unexpected" to "expected" load design?

WB: Design for NY Stock Exchange brought in security firms to discuss what the possible attacks, but this is still all based on what we have seen in the past and may not anticipate what is coming in the future. Still, big buildings are really hard to take down. Elements are very large.

CG: Depends really on the use of the building and its location to determine what kinds of threats it would warrant. Many ordinary buildings do not need these considerations and most owners would not be interested in absorbing the costs for design against these.

JB: The government is going with barriers to keep vehicles [potential threats] 100’ from the building.

RM: The codes do not need to specify this. It needs to be done on an individual project basis. They still focus on keeping the threat away. There always is a trade off in risk vs. costs.

WB: If you look at it statistically, these events are pretty darn rare.

AK: We do not design for tornadoes, because the probability of impact is so small.

AUD: It is important to note from the WTC complex the little details that helped hold the buildings up. Also if there are systems that work better, then we should know this.

RM: Again, the MRF provides a lot of redundancy. It is really the redundancy in the capacity. The moment capacity gives you lots of reserve. If you design for full capacity, they will have the reserve strength that you need.

CG: Sometimes in beefing up connections, you reduce the ductility. As a result, you reduce the dynamic robustness of the connection. Capacities of these connections under dynamic situations needs to be examined.

AW: In Japan, because of the threat of very large earthquakes, engineers are always considering plastic deformations. Because there is diminished seismic threat, very simplified connections are used in the US practice. The use of very simplified connections may prohibit the alternate load paths. Japanese buildings are three times more expensive due to these detailing considerations, but this will make the buildings more robust and redundant.

AUD: The conversation has been entirely here about steel, what about concrete?

WB: The Murrah Building is an example. Monolithic pours implies more continuity... which helps. Members are large also...which helps. A lot of construction is now composite.

CG: Problem with concrete is that the dynamic forces are larger due to the increased mass. There are definitely benefits to be gained from composite construction.

RM: Buildings are designed for materials based on where they are being constructed regionally. Concrete has its own problems. Detailing and development of reinforcement are important in insuring those connections.

AUD: There had not been a collapse of a steel building due to fire prior to WTC. Is there an example of progressive collapse outside of military attack?

JB: There has been evidence of disproportionate collapse, failure of one connection leading to the failure of the entire structure. Example of Hartford.

AUD: Single column strategy may not be appropriate evaluation given the attack level. Consider the amount of TNT used in these attacks. Progressive collapse provisions would not capture the Oklahoma City scenario.

Also, when you remove a column, everyone removes one at the bottom but what about at the top, where members are smaller and the loss may be more profoundly experienced.

RM: Structures don’t have a capacity for upward loads. The Oklahoma City scenario had an upward load scenario. Everything at the top of the structure is light and vulnerable.

AUD: Has the insurance industry weighed in post-9/11?

JB: One company, before they insure a building, they have a large staff of engineers that evaluate the structure. This company declined to cover the WTC.

AUD: How do you balance ductility and the stronger connection?

CG: Dynamic loads are larger than the static loads you typically design for. Ductility or energy absorption is important in the connections. How ductile do you need to be? Energy absorption is usually going to be away from the connection, member itself will deform and absorb the energy. It may remove some demand on the connection. In 2 International [in Hong Kong], we increased live floor loads 7-8 times to see deformation of floor systems. This helped us to determine where the enhanced connections need to be.

WB: Most designers try to engineer a reasonable amount of ductility into the structure.

AUD: Basically its all in the detailing. In Oklahoma City, the problem was in the detailing...in continuity of the rebar. When you compare steel to concrete, the connections are ultimately there. I think after 9/11 concrete has some form of upper hand. If anything, for the fire rating. Concrete should be considered seriously. I don’t know why it is not being considered more.

RM: The controversy between concrete and steel will go on forever. There is no one right material. I don’t see one material better than another. Money rules.

JB: We need more research done on the robustness and redundancy of conventional systems. Tools to examine the dynamic performance of typical connections in absorbing energy and arresting progressive collapse need to be developed in research.

WB: Let’s do more research on why buildings do as well as they do. There are many more survivors.

AW: Showed example of a reinforced concrete beam element that arrested the progressive collapse of a building in Japan.

AK: Closed session at 10 am.