Minutes of October 19 - 20, 2004, Meeting - Gaithersburg, Maryland

The slides from the presentations are embedded as links to PDF files within this document and, thus, are summarized in these minutes. Each presentation was followed by a discussion period. “Q” indicates a question, “A” the corresponding answer, and “C” a comment. All questions and comments, unless otherwise noted, were made by Advisory Committee members. All answers, unless otherwise noted, were by NIST personnel.

The minutes summarize the main points of each discussion; they are not intended to be a verbatim transcript of the meeting.

October 19, 2004

Back to agenda

Opening Remarks
Dr. Robert Hanson, Acting Committee Chairman

Dr. Robert Hanson, acting chairman for the fifth meeting of the National Construction Safety Team (NCST) Advisory Committee, called the meeting to order. Mr. Paul Fitzgerald and Dr. Kathleen Tierney were unable to attend this meeting due to prior commitments. Dr. Hanson reviewed the process and rules of the meeting, stating that public observers may only participate during the public comment session at the end of the first day. He also reminded the Committee members of the procedures for approving and disapproving recommendations and actions. Dr. Hanson then turned the meeting over to Dr. James Hill, Director of the Building and Fire Research Laboratory at the National Institute of Standards and Technology (NIST), for his opening remarks.

Opening Remarks
Dr. James Hill, Director, Building and Fire Research Laboratory

Dr. Hill welcomed the Committee and the public attendees to the meeting. Dr. Hill noted that Dr. Hratch Semerjian, the Acting Director of NIST, and Dr. Richard Kayser, the Acting Deputy Director, were unable to attend this meeting of the NCST Advisory Committee. However, Mr. Matthew Heyman, NIST Chief of Staff, was in attendance and was the Acting Director of NIST in their absence.

Dr. Hill stated his appreciation for the commitment and dedication of the NCST Advisory Committee members. He apologized for the change in the dates for this meeting and noted that the change to an open session for most of the meeting necessitated the change in schedule.

Dr. Hill reminded the Committee that the purpose of this meeting was to review the technical progress on the World Trade Center (WTC) and Rhode Island Station nightclub investigations, to review the issues that have been identified as a result of these investigations, and to get the Advisory Committee’s feedback on these issues. Dr. Hill also noted that NIST would like to get input from the Advisory Committee on the format of the next meeting, scheduled for November.

With respect to the Committee’s report to Congress, Dr. Hill noted that NIST has discussed not taking on further investigations without resources.

Dr. Hill stated that Hurricane Charley did a great deal of damage in Florida. NIST worked with teams from the Federal Emergency Management Agency (FEMA), the Institute for Business and Home Safety (IBHS), and Texas Tech University that were deployed following the hurricane and forwarded an assessment of damage and failures that would merit further investigation to the director of NIST. The Department of Commerce asked what NIST could do in response to Hurricane Charley. NIST responded by saying that it could conduct a study of specific building failures. He informed the Committee that many of the criteria in the NCST act for initiating an investigation were met for Hurricane Charley. The Department of Commerce negotiated with the Office of Management and Budget, and there is a possibility that funds may be made available through supplemental appropriation to FEMA, although a final decision has not been made.

C: The reason for creating the NCST was for timely response to building disasters, and whatever we can do as the Advisory Committee to put pressure on appropriate entities to fund NIST is important. To conduct investigations 6 to 9 months after an event is not appropriate. The money came through FEMA before, and it should come directly to NIST.
C: It’s an item we’ll include on the agenda for the session on what our report to Congress should include.

Back to agenda

Status of the World Trade Center Investigation and Discussion
Dr. S. Shyam Sunder, Acting Deputy Director, Building and Fire Research Laboratory

PRESENTATION (pdf file)

Dr. Sunder, the Lead Investigator, gave an update of the WTC Investigation. He said that good solid technical progress continues to be made on the investigation, drawing talent from NIST, outside experts, and contractors. He noted that the Investigation is ongoing and that current findings may be revised and additional findings presented in the final report. NIST is not making any recommendations at this time; all recommendations will be made in the final report.

Dr. Sunder said that most technical work is complete; NIST plans to release a draft of the final report for public comment in December 2004 or January 2005. NIST plans to release the final investigation report in April or May 2005. The WTC 7 report will be issued as a supplement to the main report, with a draft issued in May and the final in July 2005. The report hierarchy includes the main WTC Investigation Report, supported by eight Program Reports, which in turn are supported by about 30 technical topic reports.

Dr. Sunder described the investigation objectives and some specific questions that NIST is seeking to answer. He also described the context of the findings and noted that the findings being presented to the Committee were not yet final. Dr. Sunder then described the analysis method used to determine the probable collapse sequence. He described the leading hypotheses for collapses of WTC 1 and WTC 2, which identify the specific damage scenarios and load redistribution paths for each tower. Both hypotheses involve (1) damage to the building support columns, floor framing, and fireproofing due to the aircraft impacts; (2) subsequent fires that softened the steel core and perimeter support columns; (3) buckling of support columns; and (4) global collapse. The leading collapse hypotheses will continue to be refined until the analyses are complete.

Dr. Sunder described the aircraft impact analysis and how NIST determined the extent of the damage to support columns, floors, and fireproofing. He summarized the findings of this analysis by stating that (1) the robustness of the perimeter frame-tube system and large dimensional size of the WTC towers helped the buildings withstand the aircraft impact; (2) the WTC towers displayed significant reserve capacity immediately after aircraft impact; (3) the loads from the damaged and severed columns were carried by nearby undamaged columns; (4) the towers would have continued to remain standing indefinitely, but for another significant event—the subsequent fires.

Dr. Sunder then reviewed the fire modeling performed for the WTC Investigation and showed a number of simulations. He noted that the fires played a major role in further reducing the structural capacity of the buildings, initiating collapse. He also stated that the jet fuel that ignited the fires was mostly consumed within the first few minutes after impact and that the fires were mainly due to burning the building contents, and to a lesser extent, aircraft contents.

Dr. Sunder noted that the time delay between collapses of the WTC towers was primarily due to: (1) the asymmetric structural damage caused by aircraft impact to WTC 2 compared to WTC 1; (2) the time it took for heat to soften and buckle the core columns that had fireproofing dislodged by debris impact; (3) the structures’ ability to redistribute loads as the core columns shortened; (4) the time it took for the fires to migrate through the buildings; and (5) the time it took to heat, soften, and buckle the perimeter columns.

Dr. Sunder then described the fire resistance testing of the floor system performed at Underwriters Laboratories. In summarizing the tests, he stated that the test structures were able to withstand fire conditions for between 45 minutes and 2 hours. The 45 minute fire resistance for the standard 17 foot floor system with a specified 0.5 inch fireproofing did not meet the 2 hour requirement of the 1968 New York City Building Code. The 2 hour fire resistance for the standard 17 foot floor system with the as-applied 0.75 inch fireproofing met the 2 hour requirement of the 1968 New York City Building Code. The likely cause of the difference in results is not the fireproofing thickness on the trusses themselves, but the presence or lack of fireproofing on the undersides of the metal deck.

Dr. Sunder summarized the analyses performed on steel recovered from the collapsed WTC towers. He noted that the 236 pieces of steel collected by NIST were adequate for determining the quality and properties of steel for the Investigation. Approximately 87 percent of the tested steel specimens exceeded the required minimum yield strengths. The safety of the WTC towers was most likely not affected by the fraction of steel that, according to NIST testing, did not meet the required minimum yield strength.

He described the NIST review of wind tunnel testing of the WTC towers that included the original WTC design wind loads, the state-of-the-practice wind tunnel studies conducted by independent private sector laboratories in 2002, and the wind load estimates developed by NIST. Dr. Sunder noted that the original design wind loads on the towers exceeded those established by the New York City Building Code. On September 11, 2001, the towers were subjected to in-service live loads that are considered to be approximately 25 percent of the design live loads. The wind loads on that day were minimal, thus allowing significantly more reserve capacity for the exterior walls. The safety of the WTC towers was most likely not affected by the fraction of members for which the demand exceeded capacity under the original design wind loads.

Dr. Sunder described NIST’s current findings related to evacuation and emergency response. He described the interviews conducted with more than 1,000 surviving occupants and the results of NIST’s decedent analysis. He noted that approximately 87 percent of the WTC tower occupants were able to evacuate successfully, including more than 99 percent below the floors of impact. Occupants were often unprepared for the physical challenge of building evacuation. Despite perceptions of some occupants, counterflow was found to not be a significant factor in the total evacuation time when compared to other factors. Based on existing egress models and actual evacuation times on September 11, 2001, it is estimated that a full building evacuation of each WTC tower with 25,000 people, three times the number present on September 11, would have required about 4 hours.

He noted that emergency responders were implementing three different operational strategies and that none of the more than 100 first responders interviewed by NIST thought that the WTC towers would collapse. Dr. Sunder stated that situational awareness varied from adequate for those outside the buildings that could see what was going on, to poor for those inside the buildings that could not see what was happening and that had poor radio communications.

All three responding departments (New York City Fire and Police Departments and the Port Authority of New York and New Jersey Police Department) experienced difficulties with radio communications. Two basic issues were function of the radio equipment in high-rise environments and the volume of radio traffic. Roughly one-third to one-half of the radio messages transmitted during the radio traffic surge conditions was not complete or understandable.

Dr. Sunder noted that first responders did not have adequate information on, nor an overall perspective of, the conditions in the WTC towers and what was happening elsewhere at the WTC site, and that interagency information sharing was inadequate. A significant amount of evidence (first-person interviews, reports, and photographic data) shows that the different agencies were working together during the WTC disaster). Interagency operations were detrimentally affected with the loss of the Office of Emergency Management command center that was located inside WTC 7. First responder interviews suggest that interagency competition had minimal effect on operations at the WTC complex on the morning of September 11, 2001. The data also suggest that some of the problems experienced were due to personnel not understanding the operating practices of the other agencies

The smoke management systems in the WTC towers were not activated during the fires on September 11, 2001. The likelihood of these systems being functional was very low due to the damage inflicted by the aircraft. Disruption of the heating, ventilating, and air conditioning (HVAC) system by the aircraft impact created a major path for vertical smoke spread in the buildings.

Dr. Sunder discussed the applicability of building codes. Although not required to conform to the New York City Building Code, the Port Authority of New York and New Jersey (Port Authority) elected to adopt the provisions of the 1968 code more than 3 years before it went into effect. The 1968 code had less restrictive provisions compared with the 1938 code. The Port Authority was created as an interstate entity under a clause of the U.S. Constitution permitting compacts between states and is not governed by the building and fire codes of any local, state, or federal jurisdiction. He then cited several documents that provide the rationale the Port Authority used for choosing the 1968 New York City Building Code.

Dr. Sunder noted that NIST does not set building codes and standards, but provides technical support to the private sector and other government agencies in the development of U.S. building and fire practice, standards, and codes. He reviewed NIST’s approach to making recommendations, stating that NIST has not yet formulated recommendations. Dr. Sunder reviewed the issues that have been identified by NIST. These issues are grouped under the following major categories: (1) increased structural integrity, (2) enhanced fire resistance, (3) improved building evacuation, and (4) improved emergency response. In its final report, NIST will recommend appropriate improvements in the way buildings are designed, constructed, maintained, and used. NIST will hold a conference in June 2005 to reinforce the importance of its findings and recommendations from the Investigation and encourage their implementation in practice.

Q: Referring to the column shortening in WTC 1, is the elastic strain reported at room temperature?
A: No. The values reported are for elevated temperature. The history is traced, including degradation of properties.

Q: For test 1 of the fire resistance tests of the floor systems at Underwriters Laboratories, you show unrestrained rating of one hour. Was that an analytical conclusion or a tested result?
A: We show in each case an unrestrained rating when we actually did a restrained test. What we are showing there is not the result of an actual unrestrained test, but the temperature criteria in the standard for a restrained test.
C: Right, one of the major significances of the series of these tests is that test 2 was an unrestrained test and showed superior performance.

Q: I want to ask about the floor performance. The way I understood your description of the collapse scenario, the behavior of the floor systems was not a central issue. Can you connect the floor results with that?
A: The results reinforce each other. The results of the fire test versus the load test support the finding that the floors were not a driving force in the collapse.

Q: Do we know for the pieces of steel that did not meet the requirement [referring to yield strength] what temperature they reached?
A: Indications are that the metal we tested did not see any kind of high-temperature excursions, and there was no damage to the paint on those pieces. They did not get above 250 ºC.

Q: The web is usually the strongest part of a column. I am puzzled by a data point on your slide showing the web of the columns lower than specified. What was the value of the flange?
A: I will address that in my presentation tomorrow.

Q: Do the 27 emergency responders from WTC 1 interviewed refer specifically to those people themselves as opposed to people that they know, other people that didn’t survive?
A: Yes, that is correct.

C (NIST): We have brought fire and structural communities together and developed models to analyze fire-structural interaction. We will need to make the models more robust before the private sector can adopt them for routine practice.

Back to agenda

Project 2 – Baseline Structural Performance and Aircraft Impact Damage Analysis
Dr. Fahim Sadek, WTC Investigation, Project 2 Leader

PRESENTATION (pdf file)

Dr. Sadek gave an update of progress on the portion of the WTC Investigation related to baseline performance and aircraft impact damage. He described the development of the structural models of the WTC towers and results of the baseline performance analysis under design loading conditions (i.e., gravity and wind).

He described the NIST review of wind tunnel testing of the WTC towers that included the original WTC design wind loads, the state-of-the-practice wind tunnel studies as provided by CPP (Cermak Peterka Peterson, Inc.) and RWDI (Rowan Williams Davis and Irwin, Inc.) for insurance litigation, and the wind load estimates developed by NIST. Wind loads developed by NIST were based on a critical assessment of information obtained from CPP and RWDI reports and state-of-the-art considerations in wind engineering.

He noted that the original design wind loads on the towers exceeded those established by the New York City Building Code. On September 11, 2001, the towers were subjected to in-service live loads, which are considered to be approximately 25 percent of the design live loads. The wind loads were minimal on that day, thus allowing significantly more reserve capacity for the exterior walls. The safety of the WTC towers was most likely not affected by the fraction of members for which the demand exceeded capacity under the original design wind load conditions.

Dr. Sadek then described the aircraft impact analysis and how NIST determined the extent of the damage to support columns, floors, and fireproofing, and the distribution of aircraft debris and fuel. He compared the NIST analysis with that performed by MIT and Weidlinger, noting that the NIST results compared well with the MIT analysis, but not as well with the Weidlinger analysis, which over predicted aircraft damage to the towers. A stability analysis by Weidlinger showed that the building, with their estimated damage, would have collapsed immediately after aircraft impact.

He summarized the findings for each area of analysis. The original design wind loads on the towers exceeded those established by the New York City Building Code and were higher than those required by other selected building codes of the era. Estimated wind loads vary by as much as 40 percent between two wind tunnel/climatological studies conducted in 2002 by CPP and RWDI as part of insurance litigation.

Dr. Sadek stated that the Demand/Capacity ratios estimated from the original design case are, in general, close to those obtained from a lower bound state-of-the-practice case. For both loading cases, a small fraction of structural components had Demand/Capacity ratios larger than 1.0; the safety of the WTC towers was most likely not affected by this.

He noted relevant issues for future investigation including (1) availability of standards for conducting wind tunnel tests and for methods to estimate wind effects from test results for design purposes, (2) availability of protocols for selection of site-specific wind speed and directionality, and (3) adequacy of prescriptive wind load standards for very tall buildings.

Q: How is capacity defined? [referring to the Demand/Capacity ratios for the structural component estimates].
A: Capacity is defined by using the American Institute of Steel Construction (AISC) Allowable Stress Design procedure.

Q: What material properties are used?
A: We are using specified or nominal design values for steel properties.

Q: What changes were made to the baseline design based on peer review by Skidmore, Owings & Merrill?
A: NIST and Skidmore, Owings & Merrill reviewed the structural database and corrected discrepancies. For the models, the exterior wall modeling and connections of the floors to the exterior walls were of concern. These were corrected. LERA (Leslie Robertson & Associates) also responded to comments from Skidmore, Owings & Merrill on the report, although there were not many comments on report content.

Q: NIST combined data from two wind tunnel tests to prepare their state-of-the-art wind load estimate. What input data were needed for the state-of-the-art case?
A: Data from CPP and RWDI were hard to review. We had to develop questions for each in order to understand what was done. After getting data, we checked data from the National Oceanic and Atmospheric Administration (NOAA) and found discrepancies.
Q: What is needed for the state-of-the-art case? What data is needed? Can it be done without other data? What is unique?
A: There are a series of state-of-the-art considerations that result in the NIST estimate of the design wind loads. Hurricane wind profiles are assumed to be flat; that is not the case. RWDI and CPP approaches both assume relatively flat hurricane wind profiles. There are significant differences to account for wind directionality and load combinations. RWDI and CPP use significantly different approaches – one uses the up-crossing method, and the other the sector-by-sector approach. There is no specific guidance on how to integrate climatological data to estimate site specific directional wind speeds.
A: Shyam’s explanation is clear. There were some constraints on the state-of-the-art case; for example, we did not have an independent wind tunnel test. We relied on the reports and formal questions. CPP did not respond to our questions; RWDI did. Both assumed flat hurricane wind speeds. A letter from the University of Western Ontario justifying the use of flat profiles has not yet been received by NIST. We identified errors and inaccuracies in their results. State-of-the-art results are closer to the physical realities.
A: RWDI and CPP are developing a white paper to address these issues.

Q: I am interested in what we know about matching the results of the impact simulations to observations—for example, the landing gear gets hung up in the core in the building in the simulation. Do you have an explanation for that?
A: We were hoping to match that. This was a very chaotic event. There are a number of uncertainties regarding the interior layout, especially for WTC 2, and the location of debris masses in the building. We assume there was a large mass somewhere inside the building that deflected the landing gear. It may be possible to simulate with additional runs and by changing parameters; we could possibly get the landing gear to exit.
Q: What about the engine that passed through the building?
A: For WTC 2, we were able to get the engine to exit the corner in the simulation. In order to get the engine to land at its actual location, we also calculated the exit velocity to be about 120 miles per hour. For the realistic case, the speed by which the engine exited was about 80 miles per hour, and for the plus case, the speed was getting close to 200 miles per hour. This was also chaotic behavior, and the engine may have been spinning with propulsion, so the 80 miles per hour may make sense.

Q: How many more columns could be removed and the buildings still stand?
A: When we did our preliminary stability analysis for WTC 1, we removed up to eight columns from the core in the model, and the building remained standing.
Q: There are uncertainties in these calculations. You can’t see inside the buildings. Could more core columns have been broken in WTC 1 than WTC 2? Or could none have been severed? Are they definite impossibilities?
A: We tried to use parameters in the model in the best way that we know them. We spent a lot of time to model the geometry, material properties, and the failure criteria correctly. We understand that these are subjected to uncertainties and tried to bound the case. We are trying to complete additional studies to confirm our model results.

C: This project has done a lot of excellent work. You and your team should be congratulated. You need to be clear about uncertainties in your report. I am concerned that there will not be any technical topic reports for this project. I recommend that technical topic reports be included for Project 2.
A: I agree. My thinking was that it would be hard to write a concise report referring to topic reports. I felt that the logic would be better in one large report with all the details together.
C: Would you consider putting the detail into appendixes attached to the report?
C: John’s comments are worthy of consideration. The Committee will look hard at the way this report is being put together.

C: We can see the critical importance of the lack of physical evidence to back up what is going on. I want to make this point for future investigations.

Back to agenda

Project 5 – Simulation of the Fires in WTC 1 and 2
Dr. Kevin B. McGrattan, WTC Investigation, Project 5, Mathematician

PRESENTATION (pdf file)

Dr. McGrattan presented the results of fire simulations of WTC 1 and WTC 2 performed with the NIST Fire Dynamics Simulator (FDS). He explained the major trends in fire movement and temperature on the various floors, and how these trends were validated by visual observations. He discussed the sensitivity analysis performed to determine how the results of the calculations changed as a consequence of varying the major input parameters. It was concluded that the fires in WTC 1, which could generally be characterized as oxygen-limited or ventilation-controlled, were less sensitive to changes in the fuel and combustible loadings, compared to the fires in WTC 2, which could be characterized as fuel-limited or fuel-controlled.

Q: Were the soffit conditions the same in both the realistic and most severe cases?
A: Yes. We felt that it made more sense to leave the soffits in the model.

Q: Aluminum burns well. Heat transfer for aluminum is five to ten times higher than steel. Was aluminum combustion considered in the model?
A: No. There are a number of issues and limitations in the model. We increased the fuel load of wood and plastic in areas where the aircraft was, but did not explicitly account for aluminum.

Q: It is likely that the aircraft impact removed floors. Did you account for double height floors?
A: Yes. We got data from the impact analysis.

Q: Did you vary the mesh size of the model?
A: In the beginning, we tested finer meshes for the single floor calculations, but we concluded that there was not much difference in the results.

C: You could do additional sensitivity studies with wider variations. There are lots of unknowns, for example, the fuel loads.
A: For 2 years prior to June 2004, we did orthogonal factorial design. We then chose a handful of parameters for the final work done since June. We wanted the less and more severe cases to be in the ballpark with a realistic case. Everybody was to choose three cases for each tower.
A: Kevin’s work was not the rate limiting step. The output of his calculations has to be run through the ANSYS calculation, and the feedback is quite long.

Q: You had a good grip on partitions, fuel load, and ventilation for WTC 1. There is a big difference in WTC 2 in the number of windows open at time zero. Do you have a case where you’re not ventilation-controlled?
A: In general, the fire in WTC 2 was not ventilation-controlled. There were enough broken windows to provide air to the fires. The bigger issue was the location and condition of the furnishings after the airplane impact. The fuel (e.g., office furnishings and aircraft debris) was redistributed by the aircraft impact. It is difficult to model rubble. Virgin furnishings are well understood. When the fire is not oxygen-limited, it is difficult to model, and the fire in WTC 2 was not oxygen-limited.

Back to agenda

Project 5 – Simulating the Coupled Fire, Thermal, Structural Response of the WTC Towers
Dr. Kuldeep R. Prasad, WTC Investigation, Project 5, Engineer

PRESENTATION (pdf file)

Dr. Prasad described the development of the NIST Fire Structure Interface (FSI), a software tool for coupling the Fire Dynamics Simulator with the thermal and structural response of complex building assemblies. A detailed algorithm was presented, based on plane layer analysis for predicting the radiative flux boundary condition to sub-grid scale structural elements. Finite element models were constructed to predict the thermal response of the perimeter and core columns, concrete slabs, and truss assemblies. Structural damage and fireproofing damage caused by the aircraft impact on the towers was incorporated in the models.

The software tool was used to estimate the global thermal response of WTC 1 and WTC 2 under two scenarios (realistic and severe). Dr. Prasad explained that numerical simulations indicate the core column temperature is a strong function of the state of the fireproofing as well as the geometry and location of the column. Several core columns in WTC 1 reached a temperature above 700 °C. The maximum temperature of an undamaged perimeter column was between 250 °C and 350 °C. He pointed out that fireproofing thickness and its variability were identified as the single most important parameter that has a direct effect on steel temperature.

Q: When the temperature of a core column reached 700 ºC, did you eliminate it from the model?
A: No. The analysis predicts the temperature history. For the visualization, it was assumed that above 700 °C, the column loses its load carrying capacity, and it does not regain its load carrying capacity when it cools down.
Q: On the next to last slide for WTC 1, the fire is moving around the core columns, but they still look cool.
A: The fireproofing on these core columns is not damaged in this region. So, although there is fire activity in this region, the steel columns have not heated significantly. Heavy columns heat slowly. Variation in fireproofing thickness and the size of the columns play a significant role.
Q: So the significance of the fireproofing if you look at that slide with 5400 seconds is that some columns did not go above 150 °C?
A: Not necessarily. As the fires move away, the columns cool down. What you’re observing is that the core columns start to cool in the later stage.

Q: Do you make a blanket assumption that if the truss or column is in the path of the airplane debris, then all the fireproofing was knocked off?
A: We looked at the degree of damage. The partition walls have a nominal strength of 500 pounds per square inch, and the fireproofing has a cohesive and/or adhesive strength of about 2 pounds per square inch. If the partitions were destroyed or knocked out of the way, then it was assumed that the fireproofing would be knocked off the column.
Q: What about the exterior columns—were they wiped clean of spray-on fireproofing?
A: The same rules applied. We looked at the degree of damage on the interior face.
Q: What about the fireproofing on the exterior of an exterior column?
A: It remains intact.
Q: So in some cases you took off just one face of fireproofing?
A: Yes.

Q: WTC 2 had 0.5 inch of spray-on fireproofing. In WTC 1, it was upgraded. It suggests that at first glance the fireproofing on floors is a significant factor.
A: The WTC 2 floor trusses do heat more quickly. As Shyam described, the loss of fireproofing is responsible for floor sagging, and was a secondary effect.
A: In WTC 1 on the south side, the pulling in of the exterior columns is due to core column shortening. In WTC 2, sagging of the floors was significant.
A: The removal of fireproofing during impact was more important than differences in initial thickness of fireproofing. To just have floor failure induce global collapse, a large number of floors would have to fail to cause collapse.

Q: The analysis converts Fire Dynamics Simulator (FDS) to a two-zone description. Would you consider using a zone analysis rather than FDS? To what extent would a zone analysis be useful?
A: Locally, a two-zone analysis is valid. We can not assume constant temperatures over the entire floor.

Q: Is there any illustration of where the fireproofing was in-place after impact? NIST needs to illustrate that and the thicknesses. With the floors in a secondary role, it will be important to illustrate clearly in the final report. You mentioned 2.2 inch thickness—is that correct?
A: We presented in June that we assumed the fireproofing was 2.2 inches thick in WTC 1 as estimated from the average installed plus the variability of measured thicknesses. We will make every effort to graphically depict where the fireproofing was intact and where it was dislodged.

Back to agenda

Project 5 – Reconstruction of the Fires in the WTC Towers
Dr. Richard G. Gann, WTC Investigation, Project 5 Leader

PRESENTATION (pdf file)

Dr. Gann summarized the findings and issues related to reconstruction of the fires in the WTC towers. The objective of Project 5 was to reconstruct the time-evolving temperature, thermal radiation, and smoke fields in WTC 1, 2, and 7 for use in evaluating the behavior and fate of occupants and responders and the structural performance of the buildings.

Findings to date include the following: (1) loading of building combustibles was on the light side of typical, (2) the movement of the fires was complex and varied from floor to floor of each building, (3) the FDS and FSI models can be used with confidence to recreate the thermal effect of a given WTC fire event, (4) both FDS and the photographic evidence indicate that the dwell time of the fires was approximately 20 minutes, and (5) FSI simulations show high sensitivity of structural element temperatures to the presence or absence of spray-applied fire resistant materials.

Issues include (1) the need for predictive tools for design-based fire scenarios; and (2) the availability of standards and codes, methodology, analytical design tools, and practical design guidance to permit considering fire as a design condition for the structure as a whole system.

Q: Ceiling tiles were removed on the floors of impact—were they removed from other floors?
A: Ceiling tile damage was observed throughout the buildings.
A: From the interviews, it was pretty well reported that ceiling tiles were down throughout the building.

Back to agenda

Project 6 – Standard Fire Tests of WTC Tower Typical Floor System
Dr. John L. Gross, WTC Investigation, Project 6 Co-Leader

PRESENTATION (pdf file)

Dr. Gross summarized the standard fire test of typical WTC tower floor systems performed at Underwriters Laboratories. The purposes of these tests were to (1) establish baseline performance of the floor systems as they were originally built, (2) differentiate the factors that most influenced the collapse of the WTC towers, and (3) study the procedures and practices used to accept an innovative structural and fireproofing system.

Tested floor systems included 17 foot and 35 foot spans with 0.5 inch or 0.75 inch of fireproofing, in restrained or unrestrained conditions. Conditions of acceptance for the ASTM E 119-00 and E 119-61 tests were described. These conditions differ for restrained and unrestrained test assemblies. Photographs were provided showing the preparation of the test assemblies, the furnaces in Toronto, Canada, and Northbrook, Illinois, and post-test observations.

Results and observations were then reviewed. In summarizing the tests, Dr. Gross stated that the test structures were able to withstand fire conditions for between 45 minutes and 2 hours. The restrained floor system with 0.75 inch fireproofing as installed obtained a fire resistance rating of 1.5 hours, while the unrestrained floor system achieved a 2 hour rating. Past experience with the ASTM E 119 test method would lead investigators to expect that the unrestrained floor assembly would not perform as well as the restrained assembly, and therefore, it would receive a lower fire rating. The 45 minute fire resistance for the standard 17 foot floor system with a specified 0.5 inch fireproofing would not have satisfied the 2 hour requirement of the 1968 New York City Building Code. The 2 hour fire resistance for the standard 17 foot floor system with the as applied 0.75 inch fireproofing met the 2 hour requirement of the New York City Building Code. The likely cause of the difference in results is not the fireproofing thickness on the trusses themselves, but the presence or lack of fireproofing on the undersides of the metal deck. The results raise the question of whether or not fire rating based on the ASTM E 119 performance of a 17 foot span floor assembly is scalable to a larger floor system, such as found in the WTC towers where spans ranged from 35 feet to 60 feet.

Issues include adequacy of the ASTM E 119 standard to provide guidance on (1) criteria for determining structural limit states, including failure, and means to measure them; (2) scale of test assembly versus prototype application; (3) effect of end restraint conditions on test results; (4) structural connections; (5) combination of loading and exposure adequately represent expected conditions; (6) procedures to analyze and evaluate data from fire resistance tests and other building components and assemblies to qualify an untested building element; (7) repeatability and reproducibility of test results; and (8) relationships between prescriptive ratings and actual performance of the contraction in realistic building fires.

Q: There was spalling on the underside of the assemblies in the large-scale restrained test. Was there similar spalling on the unrestrained and the smaller scale test assemblies?
A: Yes, the unrestrained test showed similar spalling, but to a lesser degree as evidenced by the amount of bowing. Spalling was a little bit delayed in these tests. It was more severe for the restrained test condition. Explosive spalling occurred on the top of the slab in both of the Northbrook tests.

Q: Did you do destructive analysis? One observation we tried to make was that compression occurs at the knuckle. Will you be able to dissect that area to determine localized crushing for continuity?
A: There was cracking at the first knuckle; that’s what Dave Collins is referring to. We’ll do our best to check that.

C: There is a limit on the number of tests, so I suggest doing mathematical modeling. Also, I suggest modeling the difference between tests and the actual connections used in the building. Also, consider modeling ¼ inch and 0 inch thicknesses of fireproofing on the diagonals of the floor trusses.
A: The connections were intentionally strong for the restrained test. The unrestrained specimen was faithful to the actual connections in that bolts were provided and the truss could expand. We can go back and do some modeling. I don’t know if elongation (tension) occurred in the diagonals.

Back to agenda

Project 6 – Strength and Impact Response of SFRM
Dr. John L. Gross, WTC Investigation, Project 6 Co-Leader

PRESENTATION (pdf file)

Dr. Gross summarized information regarding the strength and impact response of spray-applied fire resistant materials (SFRM). The NIST study attempted to determine the force required to dislodge SFRM from structural steel members, using in-place density and bond strength, laboratory static strength properties, and impact tests.

Tests were performed on 8 inch by 16 inch by 0.25 inch plates and 1 inch by 20 inch bars with and without primer. Adhesive strength and cohesive strength were measured.

Results indicate poor adhesion to primed surfaces with good adhesion to bare steel. For unprimed specimens, cohesive strength is similar to adhesive strength. Adhesion to planar elements is controlled by adhesive or cohesive strength, whichever is smaller; for a bar it is controlled by cohesive strength parallel to the surface. Dynamic tests are ongoing.

Q: Regarding the impact tests, you haven’t finished the acceleration results?
A: That’s correct.

Q: How many passes are needed to get the 1.5 inch thickness of spray-on fireproofing?
A: Two to three passes are needed. The supplier came back the second day and touched up the fireproofing.
Q: How long would you have to wait between passes to apply in the field?
A: We waited a day between passes. The binder is Portland cement.

Q: Was there cohesive strength data for primed steel?
A: No. The cohesive property is a test of the fireproofing only. The presence of primer affects only the adhesive properties test.

Q: Has anyone researched the primer used in the WTC towers?
A: We used the same primer as used on the exterior panels. We could not get the same primer as used for the trusses.
C: In general, most engineers prefer painted steel, not rusted, and without primer. The fireproofing sticks better. Some are comfortable with steel being primed; the feeling, though, is without primer is better.

Q: Did you measure thermal conductivity and diffusivity?
A: The thermal conductivity and diffusivity were measured in Project 5.

Back to agenda

Project 6 – Structural Fire Response and Collapse Analysis
Dr. Therese P. McAllister, WTC Investigation, Project 6 Co-Leader

PRESENTATION (pdf file)

Dr. McAllister summarized information regarding the structural fire response and collapse analysis. The objective of the Project 6 study is to determine the structural response of the WTC towers to aircraft impact and internal fires and to identify the most probable structural collapse mechanisms.

NIST developed and used a comprehensive approach to determine the probable collapse sequences from aircraft impact to collapse initiation. This approach combined mathematical modeling, well-established statistical and probability based analysis methods, laboratory experiments, and analysis of photographic and videographic evidence. The approach allowed for evaluation and comparison of possible collapse sequences based on different states, fire paths, and structural load redistribution paths. The approach accounted for variations in models, input parameters, analyses, and observed results. A flow chart was then projected that showed the sequence and interdependencies of the analyses that were performed.

Dr. McAllister stated that extensive sensitivity analyses using Orthogonal Factorial Design (OFD) were conducted to determine the most influential factors for each analysis step used in determining the probable collapse sequence. This included developing factors for the realistic case, more severe case, and less severe case.

Dr. McAllister then described the leading hypotheses for collapse of WTC 1 and WTC 2, which identify the specific damage scenarios and load redistribution paths for each tower. Both hypotheses involve (1) damage to the building support columns, floor framing, and fireproofing due to the aircraft impacts; (2) subsequent fires that softened the steel core and perimeter support columns; (3) buckling of support columns; and (4) global collapse. The leading collapse hypotheses will continue to be refined until the analyses are complete.

Dr. McAllister reviewed the damage to the WTC towers produced by the aircraft impact and how stresses were redistributed by the floor system and to a lesser extent the hat truss to the remaining undamaged columns.

Structural response to fires subsequent to impact damage included core column shortening, perimeter column bowing and buckling, and floor sagging and pulling. This is consistent with photographic evidence showing the inward bowing of the perimeter columns on the south face of WTC 1 and east face of WTC 2 approximately 5 to 10 minutes prior to collapse.

Truss temperature exceeded 500 to 600 degrees Celsius only where fireproofing damage and fire exposure occurred. Damage to the fireproofing was extensive near the impact areas.

Dr. McAllister noted that the final analyses are currently under way. These analyses are attempting to come to closure on issues involving large plastic and creep deformations and kinking (non-linear buckling) induced by localized plastic deformation. A global analysis is also being performed to determine the structural response of the WTC towers to large fires without impact damage.

Issues identified by Dr McAllister include (1) availability of explicit standards, code provisions, methodology, analytical design tools, and practical design guidance for designing structures to resist progressive collapse in the event of abnormal loads; (2) availability of analytical methodologies for prediction of complex failure phenomena of structural systems under abnormal loads; (3) availability of standards, codes, methodology, analytical design tools, and practical design guidance to permit considering fire as a design condition for the structure as a whole system; (4) availability of regulations that would adopt code provisions using the structural frame approach to fire resistance ratings, which requires structural members to be fire protected to the same rating as columns; (5) applied passive fire protection conformance to conditions in actual or equivalent tests used to establish fire resistance rating of the building component or assembly; and (6) availability of regulatory requirements for retention of documents related to the design, construction, operation, maintenance, and modifications of buildings, including retention offsite.

Q: In the absence of impact, fire only, burnout would have been achieved and the building would not have collapsed? Am I interpreting that correctly?
A: Yes. For the fires we have analyzed to date for floor systems with ¾ inch fireproofing in place, even with gaps observed in photographs, the floors would have deformed, but would not have initiated collapse.
A: We have looked at credible fires in an undamaged tower. Remember, for this scenario, there would not be broken windows to supply oxygen to fuel the fire. This is a working hypothesis and analyses remain to be completed.

Q: Regarding the findings for global analysis with impact damage, I want to make sure I’m interpreting the information correctly for floor 96 in WTC 1. At 600 seconds, there’s 23 inches of deflection on the trusses. When the fires move away, the trusses restore to 6 inches of deflection?
A: Yes. The 23 inches is next to the impact area.

C (NIST): Referring to the slide on global analysis without impact damage. You have a statement that burnout was likely prior to collapse. This infers that collapse would occur. You may want to change your wording to say burnout without collapse.
A: Agree.

Q: Do you have a complete run for the entire buildup of the tower?
A: We have completed the realistic case for WTC 1. The realistic case for WTC 2 is running and may be completed later today. We’ve also done the component analyses.

Q: Can you envision another set of conditions that gives the same observed failure mechanism?
A: We had to remove four to five floors to get global instability.
A: We looked at this very carefully. We could not find a way to make the building come down.

Q: As the fires moved, the trusses cooled but the slab continued to heat? Why?
A: The slab is 4 inches thick. The center is cooler than the surface. The center of the slab continued to rise in temperature to reach equilibrium with the warmer upper and/or lower surfaces of the slab.
A: The concrete slab has a very high temperature on the surface and a steep temperature gradient resulting in low temperatures in the center of the concrete slab. As the fires move away, the high temperature on the surface equilibrates, which results in lowering of the surface temperature, but the temperature of the inner layers of the concrete slab can increase.

Q: In the column areas in the debris field, you assume that fireproofing was knocked off. Did you make the same assumption for the open web joists that were in the debris field?
A: Yes.

Q: How did you model connections of the floor beam and the core to the columns? Did you model them per the shop drawings—the axial capacity? I’m surprised that you get a lot of redistribution of vertical load from the inside to the outside. My understanding is that all the column connections of the beams and the core are basically shear tab connections with no moment capacity and that a 60 foot long joist isn’t going to be transporting much vertical load unless you’re talking about a global three-dimensional model.
A: What you’re saying is true for a truss, but when you look at the entire floor system and think about a plate with a hole in the middle and start pushing down on a plate, you get a different response than you would expect for a single truss.
A: The connections were largely pinned, and they were modeled as pinned; those that were fixed were modeled as fixed. The 5 inch slab floor system participated with the beams.

C: The term “shortening” implies compression shortening. Maybe you should use “bowing” instead.

Public Comment Period

Dr. Hanson, the acting Committee chair, stated the ground rules for presenting public comments. Each speaker had 5 minutes to address the Committee. Members of the public may submit their comments in writing at the meeting or at any time.

Dr. Hanson called the first speaker to the podium, Mr. Jake Pauls.

Jake Pauls, Jake Pauls Consulting Services
STATEMENT (pdf file)

Mr. Pauls addressed occupant behavior and egress issues. He presented a synopsis of his 16 pages of comments on NIST Special Publication 1000-5, Progress Report on the Federal Building and Fire Safety Investigation of the World Trade Center Disaster, June 2004 (June 2004 Progress Report), which were submitted by the Skyscraper Safety Campaign. He stated that one of the problems with the Progress Report was the difficulty in finding specific topics because there were no bookmarks provided with the PDF files on the Web.

He questioned the data on the initial population estimates. Mr. Pauls and others would like to better understand how the estimates were obtained. He discussed the predictions of evacuation times and said there was not enough detail on the models, such as the input assumptions and the outputs.

Mr. Pauls indicated that he has a major issue with the discussion of exit stairway remoteness in the June 2004 Progress Report. He stated that there is immense detail dealing with other issues, such as perimeter column damage, but not for figures that represent the floor area showing the stairs. In his professional opinion, the exit stairs were grossly out of code.

Mr. Pauls said that the information on the size of the floor areas, which is essential for determining occupant load, was missing, and that basic information about the egress system in the building was not provided.

Mr. Pauls did not agree with NIST’s statement that it has received all of the essential information because NIST staff had stated earlier at that day’s meeting that they did not have the detailed floor plans for WTC 2.

Mr. Pauls commented on the analysis of first-person accounts, some of which were accounts of human dramas taking place on the morning of September 11, 2001. He stated the need to carefully question the use of such accounts. He also questioned the reliability of the chain of recording and the characterization of evacuee behavior as pushing, shoving, and yelling.

Dr. Hanson called the next speaker, Dr. James Quintiere.

James Quintiere, University of Maryland, Department of Fire Protection Engineering
STATEMENT (pdf file)

STATEMENT (pdf file)

Professor Quintiere said that one of the critical things he had heard at the meeting is that removing insulation from the core columns is a very critical event in explaining what happened to the towers. He added that there had not yet been significant explanation to convince him this is what happened.

He then presented his comments on NIST Special Publication 1000-5, Progress Report on the Federal Building and Fire Safety Investigation of the World Trade Center Disaster, June 2004 (June 2004 Progress Report), which were submitted to NIST by the Skyscraper Safety Campaign. Although much information and analyses have been compiled, he indicated that the NIST information is far from complete as many results and details are not included in the June 2004 Progress Report. He commended the building and fire groups of NIST for conducting a significant, collective study for the first time. He stated, however, that the investigative conclusions need to have focus, clarity, and impact. Professor Quintiere’s questions and comments focused on the following areas: the temperature reached by the fire and by the steel, the accuracy and consistency of the computer modeling predictions for the fire environment, and the NIST working hypothesis for collapse, including impact computations, single truss analysis, E119 standard fire tests, and insulation history.

Dr. Hanson called the next speaker, Ms. Sally Regenhard.

Sally Regenhard, Skyscraper Safety Campaign
STATEMENT (pdf file)

Analysis from Professional Advisory Panel (pdf file)

Joint Subcommittee Overview of WTC Investigation (pdf file)

Subcommittee on Buildings & Fire Codes (pdf file)

Ms. Regenhard expressed that after September 11, 2001, she and 343 firefighter families and the families of the victims were outraged that something like the collapse of the WTC could happen. She led a delegation of parents and families to Congress to demand valid investigation. With the help of the House Science Committee and Congressman Sherwood Boelhert, they picked NIST to do this investigation.

She thanked NIST and all the people that she has worked with continuously for almost 2 years now. Ms. Regenhard said that she was very happy to advocate for this investigation and thanked the members of the federal advisory committee. Her organization, the Skyscraper Safety Campaign, is a voluntary organization, and she indicated that some of her professional advisors were at today’s meeting and had just testified. Ms. Regenhard said that she knows what it is to have people of the Committee’s caliber give their time essentially for free. But they are doing it because they are doing good and she knows that their work is going to save people’s lives in the future. She stated that it’s too late for the over 2,700 innocent victims, most of them who died needlessly in a death which certainly could have been avoided if people did their job and if buildings were constructed in the safe manner and with means of egress to get out of a dangerous building.

Ms. Regenhard submitted reports by the Skyscraper Safety Campaign’s professional advisors and subcommittees. The major issue the Skyscraper Safety is concerned about is the character and the intention of the WTC Investigation June 2004 Progress Report—research and fact finding versus forensic investigation. Although the report contains a great deal of genuine research and fact finding, the Skyscraper Safety Campaign feels that the collection of data cannot take precedent over the investigative intention of the congressional mandate. She stated that the public demands to know why, as well as how, the WTC collapsed, and there is a need to correct errors in order to avoid such catastrophes in the future.

She said that NIST should aggressively assert itself as the appropriate center for forensic engineering and investigation of construction, fire, or egress disaster, using the full power granted by the congressional legislation and modeling its role on that of the National Transportation Safety Board.

Ms. Regenhard expressed her disappointment that she was not hearing anything in that day’s meeting on building code requirements, violations, and interpretations. The issues regarding local building codes and the Port Authority’s ability to abrogate them, reinterpret them, and substitute its own proposal with or without full testing or outside evaluation needs to be fully addressed. She stated that critically important aspects of the construction and the subsequent loss of life are connected to design decisions made by the Port Authority, including structural issues concerning the use of light weight truss systems for floor support as part of the primary structure system.

Ms. Regenhard said that the other issues involved documentation, modeling, and testing. There are serious doubts about the sufficiency of existing fire testing, virtual or real, especially the use of scaled elements. Moreover, there seems to be an insufficient concern for human behavior scenarios that give needed insight into egress system’s operational effectiveness.

Dr. Hanson called the next speaker, Mr. Robert Polk.

Robert Polk, National Association of State Fire Marshals
STATEMENT (pdf file)

Dr. Hanson called the next speaker, Mr. Richard Kuchnicki.

Richard Kuchnicki, International Code Council
STATEMENT (pdf file)

Dr. Hanson thanked the speakers and concluded the public comment session.

Back to agenda

Project 8 – Emergency Response
Mr. J. Randall Lawson, WTC Investigation, Project 8 Leader

PRESENTATION (pdf file)

Mr. Lawson presented an update on emergency response. He summarized changes after the 1993 bombing, including the establishment of the Office of Emergency Management (OEM), the agreement of Port Authority with the New York City Fire Department (FDNY) regarding fire safety, the installation of an FDNY high-rise radio repeater in the towers, and the upgrading of the WTC elevator intercom system.

To gain information for the investigation, NIST conducted face-to-face interviews with 68 FDNY personnel, 25 New York City Police Department (NYPD) personnel, 15 Port Authority staff, and 3 others from security, fire safety, and communications. Mr. Lawson explained the role of each responding department on September 11 and FDNY’s initial size-up of the WTC conditions.

He pointed out that emergency responders were implementing three different operational strategies and that no first responders interviewed by NIST thought that the WTC towers would totally collapse. The outside command posts and inside command communicating with the outside had the strategy to evacuate and rescue all below the fires. The objective for the command officers for inside operations was to get enough personnel and equipment upstairs to cut a path through the fire to rescue occupants above the fire and evacuate and rescue all below the fires. The strategy for the company level command was to get up to the fire floors and extinguish the fires. He stated that situational awareness varied from adequate for those outside the buildings that could see what was going on, to poor for those inside the buildings that could not see what was happening and that had poor radio communications.

Mr. Lawson reviewed the location FDNY, NYPD, and Port Authority Police Department (PAPD) command posts, as well as the details of staging, assignments, and personnel accountability on that day. Although some freelancing (not officially assigned) was experienced, emergency responders generally followed their department protocol. Many FDNY units assigned to the WTC site did not know WTC 1 from WTC 2. Mr. Lawson reported that personnel accountability lists were typically maintained on magnetic boards or paper, and with the collapse of WTC 2, all accountability lists for each department were lost. No records were maintained beyond the WTC site.

Mr. Lawson described the access to the buildings after aircraft impact. Only two elevators out of 198 were operating inside the two towers. The stairways were filled with occupants trying to evacuate. FDNY personnel and other emergency responders had to deal with the challenges of counterflow, the weight of personal protective equipment and gear, and climbing more than 10 to 12 flights of stairs. The estimated climbing rate, based on a 60 minute climbing period to the maximum height achieved, is 1.4 to 2 minutes per floor.

Standard evacuation procedure was down the stairs to exit at the bottom of the WTC towers, not roof evacuation. Mr. Lawson discussed the NYPD and FDNY roof rescue protocols, which were basically the same. The FDNY air support plan objectives include providing FDNY with the capability of placing fire personnel on the roof for the purpose of ventilating and search, comforting people who are trapped, evacuating persons in need of immediate medical attention, and evacuation as a last resort. The Port Authority reports that it never advised tenants to evacuate upward.

Mr. Lawson provided details on the emergency responder radio systems, the FDNY radio communications systems, and the readability analysis of radio communications. All three responding departments (FDNY, NYPD, and PAPD) experienced difficulties with radio communications. Two basic issues were normal function of the radio equipment in high-rise environments and the volume of radio traffic. Roughly one-third to one-half of the radio messages transmitted during the radio traffic surge conditions was not complete or understandable.

Mr. Lawson explained that first responders did not have adequate information on, nor an overall perspective of, the conditions in the WTC towers and what was happening elsewhere at the WTC site, and that interagency information sharing was inadequate. A significant amount of evidence (first-person interviews, reports, and photographic data) shows that the different agencies were working together during the WTC disaster). Interagency operations were detrimentally affected with the loss of the OEM command center that was located inside WTC 7. NIST interviews with first responders suggest that interagency competition had minimal effect on operations at the WTC complex on the morning of September 11, 2001. The data also suggest that some of the problems experienced were due to personnel not understanding the operating practices of the other agencies.

Command and control difficulties began with the dispatch of large numbers of firefighters to the WTC before adequate command posts and staff could be assembled to manage them. Mr. Lawson said that FDNY command and control was seriously affected by the lack of good communication and that their system of maintaining assignment records could not manage the number of units and personnel. In addition, he emphasized that there was no means to back up the unit assignment records for FDNY, NYPD, or PAPD.

Mr. Lawson identified issues that need to be addressed in changes to practices, standards, and codes and identified alternative practices and technologies that may address these issues. In the areas of access and firefighting, the issues are (1) physiological impact on firefighters with equipment climbing more than 10 to 12 floors during an emergency, (2) adequacy of capacity for egress and firefighter access during full evacuation of fully occupied tall buildings, and (3) distance between stairwells where standpipes are located. For emergency communications, the issues are (1) lack of rigorous preemergency inspection and testing of radio communications systems within high-rise buildings to identify performance gaps and inadequacies, (2) performance requirements for emergency communication systems in buildings, and (3) lack of communications network architecture (interoperability) and operational protocols for intra- and interagency communication at all levels of organizational hierarchy.

Command and control issues include (1) availability of detailed procedures for gathering, processing, and delivering situational information to all first responders; (2) availability of effective codes and protocols for establishment and uninterrupted operation of the incident command and control system and for preservation and dissemination of information; (3) rapid adoption and execution of a unified emergency response mission by all first responder ranks; and (4) the dispatch of large numbers of personnel and apparatus and the ability of management to maintain accountability in a timely manner associated with arrival and deployment of personnel, and the ability of the incident site to effectively accommodate large numbers of personnel and apparatus.

Q: Having three operational strategies for emergency response is like having none. How do you end up with three strategies? We need to be sure we have a good basis for this finding. Was the upper level strategy not being communicated? Why couldn’t this be communicated to the next level in the command chain?
A: There are clearly three concepts here. The level of the individual is considered as well as the years of experience and their perception. High-ranking officers were having trouble with communications. Also, once inside the building, they realized the fires were too big to fight. There was a blending over time. Situational awareness changed with location. There was a short window of time for response.
C: There is a common theme of information transfer or lack thereof.

Q: The accountability lists were lost. Is there any indication about the accuracy of locations of units and buildings?
A: The indication is that the lists were not complete.

Q: There is a report of firefighters on the 78th floor. What does that do to the climbing rate?
A: They took the elevators to the 40th floor and walked up the rest of the way.
C: The firefighters on the 78th floor had no gear. One was a marathon runner.

A member of the public, without invitation to comment, offered a personal statement about these items and this project.

Dr. Hanson thanked the presenters, attendees, and speakers. He adjourned the meeting at 6:15 p.m.

October 20, 2004

Dr. Hanson called the meeting to order at 8 a.m.

Back to agenda

Project 7 – Occupant Behavior, Egress, and Emergency Communications
Mr. Jason D. Averill, WTC Investigation, Project 7 Leader

PRESENTATION (pdf file)

Mr. Averill noted that first-person accounts were gathered from 225 interviews of occupants of WTC 1, 2, and 7; over 800 telephone interviews of occupants of WTC 1 and 2; and 6 focus groups. He presented an estimate of the number of decedents above and below the impact regions in WTC 1 and WTC 2, along with predominant factors associated with occupants who did not successfully evacuate. The initial population in WTC 1 and 2 at 8:46 a.m. on September 11, 2001, is estimated to have been approximately 17,400, plus or minus 1,200 persons. The majority of the deaths in the region below impact were accounted for by being trapped by debris on the starting floor, delayed evacuation initiation, or performing emergency response building responsibilities.

Next, observations of building damage in WTC 1 and WTC 2 from a variety of eyewitness accounts were presented. Mr. Averill stated that Stairwell A in WTC 2 was passable for a period of time after the aircraft impact.

Mr. Averill reviewed the public address system announcements made in the towers on September 11, 2001, and noted that opportunities to improve occupant situational awareness were often lost. Specific knowledge about locations of fires and impact damage was only occasionally communicated to occupants who requested the information and was without apparent coordination. For example, some operators advised sheltering in place while others advised evacuation.

The causal model results for evacuation initiation delay and normalized stairwell evacuation time were presented, along with more general findings regarding the overall evacuation process. Mr. Averill then noted that despite perception of some occupants, counterflow was found to not be a significant factor in the total evacuation time when compared to other factors.

He noted that after aircraft impact, most of the emergency communication system in WTC 1 was disabled. All three stairwells in WTC 1, and two of the three in WTC 2, were rendered impassable in the area of impact. The PANYNJ reports that it never advised tenants to evacuate upward.

Occupants were often unprepared for the physical challenge of full building evacuation. Mobility-challenged occupants were not universally identified nor prepared for evacuation. Occupants were often unprepared to encounter transfer hallways during the stairwell descent. Phased evacuation (defend-in-place) would not have been an appropriate strategy. He noted that approximately 87 percent of the WTC tower occupants were able to evacuate successfully, including more than 99 percent below the floors of impact. Use of elevators in WTC 2 saved roughly 3,000 lives.

Finally, Mr. Averill introduced the issues related to the design and performance of the egress system, emergency communications, occupant preparedness, and evacuation technologies. Some of these issues include (1) that building egress systems are not designed to accommodate full building evacuation; (2) stairwells can be physically proximate yet considered remote by a “walking path” measurement; (3) procedures for identifying and assisting mobility-challenged occupants may be insufficient; (4) the emergency broadcast system may be useful in large-scale emergencies; (5) emergency plans are filed to achieve regulatory compliance, but may not be adequately implemented in practice; (6) elevator door restrictor plate can entrap occupants in the event of an emergency; and (7) lack of adequate egress models and systematic methodology for accounting for human behavior during evacuation.

Q: Were observations of damage made at any time during egress?
A: Observations were made directly after impact. In WTC 1, people were still on the same floor. In WTC 2, many people had already begun their evacuation before the impact of the aircraft.

Q: Regarding the occupants who may have been from the impact zone in WTC 2, is it possible the descending persons were firefighters? The Palmer tape mentions firefighters in the twenties in WTC 2.
A: The interviewee did not identify them as uniformed personnel.
A: It is not stated explicitly, but it appears clear they were not firefighters. The interview indicated that firefighters were on the way up.
A: We don’t know where these persons were coming from. That’s why we are tempered in the way we state the finding.

Q: What is the issue of architects not being trained in egress system design based on?
A: Discussion of the issue of what it is that constitutes egress system design and what is it that qualifies someone to do design egress systems.

Q: With respect to the evacuation speed reported, how were speed and rate determined?
A: The speeds were determined from reported time that occupants spent in stairwells from the generalized interviews.
Q: It was self-reported?
A: Yes.

Q: And the rates?
A: For the stairwell door exit rates, we took the total population, averaging it over the 100 minutes and dividing it by the three stairwells, and knowing what the effective width is, which is different than the clear width. This gives you a number of about 37 people per minute going out each door on average over those 100 minutes.

Q: Are the speeds for total time in the stairwell or do they include rests that might be taken out of the stairwell?
A: They are based on the total time spent in the stairwell.

Q: Any idea why evacuation rates were so low?
A: People reported they were moving as quickly as they could. The flow in general was the limiting factor. In addition, they reported the environmental conditions in the stairwells of smoke, water, etc, and the physical challenge of going down 1,000 feet worth of stairs for some people was difficult.

Q: Did the towers meet the remoteness requirement in the New York City Building Code?
A: When the towers were originally designed, remoteness was defined as being as remote as practicable. It would not have met the half and third diagonal rule used in the current code.

Q: Communication issues are more of a general problem than some of the other issues. Was the counterflow issue one of perception? The emergency services going up are hitting a constant stream, whereas the people going down are only hitting counterflow every so often.
A: There is an issue of perception. People did encounter firefighters and stepped aside to let them pass. Statistics showed that the counterflow was less of an issue than other factors, which overwhelmed the counterflow issue.
C: It’s probably worth stating that the general result about counterflow being less of a problem compared to others has to be taken in context. It does not mean that specific individuals were not affected significantly by counterflow or that firefighters always were affected by counterflow. Those are all consistent statements.
Q: Could it also be that persons were affected by counterflow and didn’t even realize it?
A: We measured how it affected their total evacuation time. If it had a significant detrimental effect, their evacuation time would be longer than for those who didn’t encounter counterflow of significant degree. Finishing the hypothesis, there was a rate limiter at which the people were moving as quickly as they were allowed to move given the conditions inside the stairwell, but they weren’t moving as fast as they might otherwise in a clear stairwell. Therefore, if they have to step aside to allow a firefighter to go by, they can make up that time by moving faster and they are no worse off than they were before. From the flip side, if you look at a firefighter, the limiting factor appears to be the physical condition and how much they were carrying, and they were moving up essentially as quickly as they could. If you delay that person by any period of time, they can’t make it up by moving faster. So, those two conclusions—that counterflow had an effect on the firefighters, but not a significant effect on the occupants—are consistent if we understand the rate limiting factor.

C: I’d like to pick up on Dave Collin’s question and address two issues. One is the question of egress training for architects and the other is the design of a core. In smaller buildings, for example two-story buildings, usually architects are responsible for most of the design and interpretation of the code. In tall buildings, there is usually a code consultant, a fire consultant, and other types of consultants. I’m sure to pass the RA exam you need to understand the code and egress, at least to a degree, as an architect. We don’t need a new profession. Today no one looks at structural design; everyone looks at egress. The design of the core in a high-rise building is like war. The owner wants the core reduced to maximize rental space, the architect has his own design, the structural engineer wants space for the columns, the fire protection engineer wants stairs in the right location, the mechanical engineer wants mechanical and electrical systems in the best space. Where the shafts for mechanical, electrical, and plumbing systems are located is probably not an important consideration. The location of fire stairs is probably more important than anything else. In Kuala Lumpur, we added a stair in the bustle. No one wanted it, but we added it anyway because the path and distances were there. We looked at several international codes and synthesized what the best had to offer. A formalized hierarchy of importance for certain size projects needs to come out in this study.
A: We will look at the issue. It’s very important. We will also look at the architect training issue and modify it.

Q: Architects are currently required to be certified in life safety. I’m not sure this criticism is warranted. Also, in the comment about the stairwells being physically proximate, was the term “practicable” in the remoteness issue a contemporary criterion at the time these buildings were being built?
A: The issue of measurable diagonal came up after the 1968 building code.
Q: So the use of practicable remoteness was a standard of practice when these buildings were built?
A: That’s my understanding.
Q: Obviously a difficult term to interpret by anyone’s measure.
A: I’ll take a closer look and verify that.
A: The code language was changed from “as remote as possible” to “as remote as practicable.” The code people at the time said it was meaningless.
C: That is my understanding as well. You need to tie it back to what the practice was at the time. You use the “walking path” or what is typically dealt with as a means of some sort of separation that you have to walk around, typically a fire barrier, to measure the distance between stairwells and you say that that is not adequate under nonfire conditions, such as overpressure. What does that mean?
A: If stairwells are physically close, the 2 hour fire separation may not be adequate for both stairwells to survive, if overpressure is a design consideration.

Q: With regard to stairwell capacity, are we comparing apples and oranges? Doesn’t that include people not going through the door?
A: The data are for a total uncontrolled building evacuation. The average value is under a fire drill scenario.
C: If you have more people in the stairwell, the number would go up.

C (NIST): One thing special about this event is that the handicapped persons in the stairwell become much more of an obstacle than any fire department counterflow. We need to consider this. We must pay more attention to this issue, which will be an issue in every evacuation.
C: It is the whole issue of faster versus slower people and having to go around people who are handicapped.
C (NIST): As Jason’s material reports, a good deal of the handicapped people self-evacuated by holding both handrails and taking one step at a time, cutting their rate way down, forcing them to occasionally step aside to let others pass. It is still a major problem for research and development.

Q: Regarding the issue with rooftop evacuation, I wonder whether the opposite of the statement is true. Did the Port Authority tell people not to go up, especially since people were rescued from the rooftop in 1993? This is an important distinction.
A: No, we don’t have any evidence that the Port Authority told people not to go up. One Port Authority employee was instructed to go down. He went as far as he could. Then he was instructed by a Port Authority individual to go to the roof. We don’t have good information on this.
C: I would like to get your perspective on how we view this. Over half of the occupants were in the buildings for 3 years or 4 years or less. Should evacuation instructions include things that should not be done? In this particular case, it was roof evacuation. The roof access was locked, and the roof was not available. For other buildings, it might be something else. Should training include things you don’t have to do or should not do? That is an important question.
C: Yes.