World Trade Center 7 NIST Collapse Simulation

From 911Encyclopedia
Jump to: navigation, search

NIST Pages Index

World Trade Center 7 Index

World Trade Center 7 Demolition Acceleration

World Trade Center 7 Demolition Videos

World Trade Center 7 Design and Construction

Finite Element Analysis (FEA)

June 2008 NIST WTC7 Draft Report (unofficial release)

August 21 2008 NIST WTC7 Report (Draft for Public Comment)

November 20 2008 NIST WTC7 Report


http://www.seaonc.org/pdfs/SEAONC_September_2010.pdf


Video generated by NIST's collapse simulation:

WTC7 w/ no damage: http://nate.flach.s3.amazonaws.com/WTC7_No_Debris_Impact_Damage.wmv

WTC7 w/ damage: http://nate.flach.s3.amazonaws.com/WTC7_Debris_Impact_Damage.wmv

WTC7 collapse initiation: http://nate.flach.s3.amazonaws.com/Collapse_Initiation.wmv

12th floor fire: http://nate.flach.s3.amazonaws.com/12th-Floor_Fire.wmv


NIST says these videos match [World Trade Center 7 Demolition Videos reality] "reasonably well":

3.5.3 Accuracy Appraisal

Given the complexity of the modeled behavior, the global collapse analyses matched the observed behavior reasonably well. The close similarity of the timing and the nature of the events up to the initiation of global collapse is strong confirmation of the extent and nature of the structural failures in the interior of the building and the accuracy of the four-step simulation process. The overall simulation of the collapsing building with damage better matched the video observations of the global collapse. The global collapse analysis confirmed the leading collapse hypothesis, which was based on the available evidence. (NCSTAR1A_p78)


http://code.google.com/p/nist-wtc7/

http://wtc.nist.gov/comments08/jonathanColewtc7comments.pdf (see comment 22)

http://web.archive.org/web/20081203151950/http://wtc.nist.gov/comments08/justinKeoghwtc7comments.pdf

(original link) http://wtc.nist.gov/comments08/justinKeoghwtc7comments.pdf


These 3D renderings of the WTC7 LSDYNA were downloaded from NIST's public ftp site last year. The files are dated Sep 10 2007. Please [Contact contact] me if you can open them.

http://s3.amazonaws.com/nasathermalimages/public/images/d3plot

http://s3.amazonaws.com/nasathermalimages/public/images/d3plot01


Building Exterior - General

The outward appearance of the building was determined by the covering, often referred to as a curtain wall, which was attached to the exterior steel frame. The curtain walls included glass that covered the windows. The curtain walls were held in place by supports attached to the steel. The primary materials used to form the outer curtain walls were polished finish granite sections, aluminum window frames, and glass. The placement of the windows in conjunction with three colors of granite gave the WTC 7 its characteristic appearance. The window arrangements on the north and south faces were similar, but differed from those on the east and west faces. For a given face, the spatial arrangement of windows and granite was repeated on most of the floors. Variations were present on the lower floors, at the 22nd and 23rd mechanical floors, and near the top on the 46th and 47th floors. The 32 mm (1¼ in.) thick granite sections were attached to rectangular box structures, referred to as trusses on the building drawings, which were fabricated from 1.2 mm (0.048 in.) thick galvanized steel. These boxes had the same dimensions as the exterior granite plates and were 127 mm (5 in.) deep. The trusses were, in turn, attached by brackets to the steel framework. The centerlines for the exterior steel columns and beams were located 0.61 m (24 in.) inward from the outer part of the curtain wall. Window frames were also attached to the trusses.

(NCSTAR_1-9_Vol1_p142)


3.4.3 Fire Dynamics Simulator ( FDS )

The major fires on Floors 7 through 9 and 11 though 13 in WTC 7 were simulated using the Fire Dynamics Simulator (FDS), version 4, in a manner similar to the simulations conducted for WTC 1 and WTC 2 (NIST NCSTAR 1-5F). There were far fewer photographs and videos of WTC 7 than of the towers; and, thus, the details of the WTC 7 fires were not as precise as for the fires in the towers. However, the imagery was sufficient to guide the WTC 7 fire simulations. Unlike the computations for WTC 1 and WTC 2, the fire simulations for WTC 7 were conducted for each floor individually, as there were no obvious pathways for the flames and heat to pass from one floor to another, aside from the debris-damaged area in the southwest corner of the building (NIST NCSTAR 1-9, Chapter 9). The fires on Floors 7, 8, and 12 were simulated using input from the visual imagery and established fire physics. The fire on Floor 9 was similar to that on Floor 8, and the simulation was derived from it. For the same reason, the fires on Floors 11 and 13 were derived from the fire on Floor 12. While use was made of the appearance of flames and window breakage in photographs and videos in formulating the simulations, the Investigation Team realized that the absolute timing of the simulations might not align exactly with the timing of the fires on September 11, 2001. A typical single floor fire simulation took up to two days on a Linux cluster with 8 processors. (NCSTAR1A_p66)


3.4.4 Fire Structure Interface (FSI)

The Fire Structure Interface (FSI) was used to impose the gas temperatures from the FDS simulations on the structural components of WTC 7 to predict the evolving thermal state of the building (NIST NCSTAR 1-9, Chapter 10). The thermal analysis approach was similar to that used to simulate the fire induced thermal loads on WTC 1 and WTC 2 described in NCSTAR 1-5G. The FDS temperature data for use in the structural analysis were sampled at 30 min intervals. For each time step, a set of thermal data files was generated that specified the thermal state of the lower 16 stories of the building. Three different thermal response computations were used. Case A used the temperature data as obtained from the FDS simulation. Case B increased the Case A gas temperatures by 10 percent and Case C decreased the Case A gas temperatures by 10 percent. Given the limited visual evidence, the Investigation Team estimated, using engineering judgment, that a 10 percent change in temperatures was within the range of reasonable and realistic values for the fires in WTC 7 on September 11, 2001. The computational time for each Case was approximately one to two days on a single processor desktop computer. (NCSTAR1A_p66)


3.4.5 Structural Analysis of the Initial Failure Event using ANSYS

The structural response of the lower 16 stories of WTC 7 to the heating from the fires on Floors 7 through 9 and 11 through 13 was simulated using ANSYS, a finite element computational model that allowed including the temperature-varying properties of the structural materials. This analysis was used to determine the sequence of events that led to the collapse initiation (NIST NCSTAR 1-9, Chapter 11). In addition to the temperature-time histories from the FSI results, the structural model used temperature dependent mechanical properties of the steels, welds, and bolts used in the construction of WTC 7, including elastic, plastic, and creep properties. The description of the building structure was based on architectural and structural drawings of the original building and subsequent building alterations. Component structural analyses were conducted to identify critical behavior and failure mechanisms that could have contributed to the global structural response of the building. These component analyses included (1) buckling of a core column, (2) beam-to-girder connections under thermal loading, and (3) girder-to-column connections under thermal loading. Subsystem analyses were then performed that incorporated the behavior and failure mechanisms identified in the component studies. The subsystems analyses included (1) the northeast corner of a typical floor and (2) a full tenant floor, both under gravity and fire loads. Modifications were made to reduce the model size and complexity and to enhance computational performance, but without adversely affecting the accuracy of the results. Whenever modeling modifications were used, they were validated against the detailed component model results. Due to the nonlinearities in the analysis, as well as fire-induced damage, a 30 min analysis could take a few weeks to complete. Due to the range of time steps that were required to reach equilibrium (e.g., from 10-6 seconds to 10s of seconds), a complete ANSYS analysis for a given thermal case took approximately six months to complete on a 64 bit workstation with quad-core, 3.0 gigahertz (GHz) processor, and 64 gigabyte (GB) of Random Access Memory (RAM). The use of user-defined elements prevented the use of parallel processing on a Linux cluster. The three different thermal response cases (A, B, and C) were used in the ANSYS analysis. Based on the ANSYS model results, it became apparent that the calculated fire-induced damage to connections and beams were occurring at essentially the same locations and with similar failure mechanisms, but shifted in time. (Case C failures occurred at a later time than the same failures in Case A, and Case A failures occurred at a later time than Case B failures.) As a result, only the fire-induced damage produced by Case B temperatures was carried forward as the initial condition for the building collapse analysis, since the damage occurred in the least computational time (i.e., 6 months). Figure 3-9 shows an example of the extent of structural damage from the fires, in this case for the 13th floor. At both 3.5 h and 4.0 h, connections, floor beams, and girders were damaged or had failed at steel temperatures that were approximately 400 °C or less, primarily due to the effects of thermal expansion. After 4.0 h of heating, there was substantially more damage and failures in the WTC 7 structural system than at 3.5 h of heating. The ANSYS results were input to the LS-DYNA analysis when it appeared that an initial failure event might be imminent, and diagrams for the 16 floors modeled in ANSYS indicated some degree of uncertainty in selecting the time and damage state for the transition. However, it appeared likely the critical damage state occurred between 3.5 h and 4 h. Accordingly, as shown in the next section, LSDYNA analyses were performed for both of these damage states. (NCSTAR1A_p70)


3.4.6 Global Collapse Analysis using LS-DYNA

A global finite element model of the WTC 7 building was developed in LS-DYNA to study its structural response to an initial failure event due to fire and to determine the sequence of events that led to collapse propagation and, ultimately, global collapse. LS-DYNA was capable of explicitly modeling failures, falling debris, and debris impact on other structural components. It could also model dynamic processes, including nonlinear material properties, nonlinear geometric deformations, material failures, contact between the collapsing structural components, and element erosion based on a defined failure criterion. In addition, LS-DYNA had capabilities to include thermal expansion and softening of materials. (For more detail, see NIST NCSTAR 1-9, Chapter 12, and NIST NCSTAR 1-9A.) The description of the building structure was based on architectural and structural drawings of the original building and subsequent building alterations, as well as erection and shop fabrication drawings. Other input data required by the global LS-DYNA model were presented in the following sections of NIST NCSTAR 1-9:

• Extent of damage to the building by debris impact from the collapse of WTC 1 (Chapter 5).

• Photographic and videographic records with time stamps that documented the observed collapse sequence (Chapter 5).

• Temperature-dependent mechanical properties of steel (Appendix E) and concrete (NIST NCSTAR 1-6A) used in the construction of WTC 7.

• Fire-induced damage to floor beams, girders, and their connections from the 16 story ANSYS analysis (Chapter 11).

• Temperatures for structural components and connections, at the time the ANSYS analysis transferred data to the LS-DYNA analysis (Chapter 10).

Due to the nonlinearities in the analysis, as well as sequential local failures, a 25 s analysis took up to 8 weeks to complete. The analyses were run on a Linux cluster with a head node with two 64 bit, 2.4 GHz processors and 4 GB of RAM and compute node with two 64 bit, 2.6 GHz processors. Six of the compute nodes had 8 GB of RAM and the remaining two nodes had 16 GB RAM. Three simulations were performed with the global LS-DYNA model.

• The first was based on NIST's best estimate of both the debris impact damage form WTC 1 and the fire-induced damage as developed using the ANSYS modeling. This occurred at 4 h in the ANSYS computation.

• The second simulation differed only in the input of a lesser degree of fire-induced damage, which occurred at 3.5 h in the ANSYS computation. The purpose of this LS-DYNA simulation was to determine whether a lesser degree of fire-induced damage could lead to the collapse of WTC 7.

• The third simulation was the same as the first, except that no debris impact damage was included. The purpose of this analysis was to determine the contribution of debris impact to the WTC 7 global collapse sequence and whether WTC 7 would have collapsed solely due to the effects of the fires.

The LS-DYNA model was initiated as follows to minimize any spurious dynamic effects associated with the loading sequence. First, gravity was applied slowly to the 47 floor structure, taking 4.5 s of elapsed simulation time.. Then, the debris impact damage from the collapse of WTC 1 was applied to the structure instantaneously by removing from the model the damaged elements that were no longer capable of bearing their loads. The structure was then allowed to damp residual vibrations for 2 s. Over the next 2 s, the structural temperatures were ramped up to the levels from the ANSYS simulation . Fourth, the fire-induced damage obtained from the 16 story ANSYS analysis, including damage to floor beams, girders, and connections, was applied instantaneously. The damage was from the computation of Case B at 4 h after the initiation of the fires in FDS. The heated, damaged structure was then free to react. The time at which the east penthouse began to descend was defined as 0.0 s, i.e., the beginning of the collapse of WTC 7. The global analysis with fire-induced damage at 4.0 h most closely matched the observed collapse events, and the following discussion begins with the results from this analysis. (NCSTAR1A_p72)


Figure 3–10 through Figure 3-14 depict the state of the WTC 7 structure at various times as the structure collapsed. The first four figures are views of the lowest 18 floors of the WTC 7 building core from the south. In these graphics, the exterior columns and some of the tenant floor structure spanning between the core and the exterior have been removed for an unobstructed view of the core. The scale on the right side shows the absolute (i.e., without any indication of direction) lateral displacement of each structural element. Displacements greater than 0.15 m (6 in.) are also shown in red. (NCSTAR1A_p72)



NIST_NCSTAR_1A_for_public_comment.pdf_Figure_3-10_p74.png

Figure 3–10 shows the beginning of upper floor failures on the east side of the building at 0.5 s, following the buckling of Columns 79, 80, and 81. The east penthouse, which was supported by these three columns, had just begun to descend.


NIST_NCSTAR_1A_for_public_comment.pdf_Figure_3-11_p74.png

About 2 s later, as shown in Figure 3–11, the collapse of all the east sections of all the floors had occurred, the upper floors had moved southward, Truss 2 had been damaged, and the westward progression of the building failure was underway.


NIST_NCSTAR_1A_for_public_comment.pdf_Figure_3-12_p75.png NIST_NCSTAR_1A_for_public_comment.pdf_Figure_3-13_p75.png

Figure 3–12 and Figure 3–13, only 2 s apart, indicate the speed with which the westward column failures proceeded between Floors 7 and 14.


NIST_NCSTAR_1A_for_public_comment.pdf_Figure_3-14_p76.png

In Figure 3-14, the total collapse of the building is underway. The two views cover the lower half of the building. The purple area at the bottom is the Con Edison substation. With no fires on the west side of Floors 10 through 14, the intact floor framing pulled the exterior columns inward as the interior columns fell downward. Loads from the buckled interior columns were redistributed to the exterior columns, which, in turn, buckled the exterior columns between Floors 7 and 14 within approximately 2 s. At that point, the entire building above the buckled-column region moved downward as a single unit, resulting in the global collapse of WTC 7.


The second LS-DYNA analysis (lesser degree of fire-induced damage) did not lead to a collapse-initiating event, despite the extensive damage shown in Figure 3–9a. The third LS-DYNA analysis demonstrated that the fire-induced damage led to the collapse of WTC 7, even without any structural damage from the debris impact. However, the mechanism of the collapse differed from the first analysis. This is discussed further in the next section.


3.5 ACCURACY OF THE PROBABLE COLLAPSE SEQUENCE

Independent assessment of the validity of the key steps in the collapse of WTC 7 was a challenging task. Some of the photographic information had been used to direct the simulations. For example, the timing of the appearance of broken windows was an input to the fire growth modeling. However, there were significant observables that were usable as corroborating evidence, as shown in Table 3–1. The "Observation Times" were determined from examinations of photographs and videos shot on September 11, 2001. The times in the second and third columns are from the two global analyses with and without debris impact damage for Case B temperatures at 4.0 h. (NCSTAR1A_p76)



3.5.2 Aspects following the Collapse Initiation

Once simulation of the global collapse of WTC 7 was underway, there was a great increase in the uncertainty in the progression of the collapse sequence, due to the random nature of the interaction, break up, disintegration, and falling of the debris. The uncertainties deriving from these random processes increasingly influenced the deterministic physics-based collapse process, and the details of the progression of the horizontal failure and final global collapse were increasingly less precise. Thus, while the two predictions of the time of descent of the west penthouse also straddled the observed time, the mechanisms of building collapse were quite different. In the analysis without debris impact damage, the exterior columns buckled near mid-height of the building, approximately between Floors 17 and 29. In the analysis with debris impact damage, the exterior columns buckled between Floors 7 to 14, due to the influence of the exterior damage near the southwest corner. In both analyses, the eastern exterior wall deflected inward at the roof level as the structure became unsupported after the vertical collapse event. The western wall also deflected inward in the analysis without debris impact damage, as it was pulled inward as the last line of core columns failed. There was another observable feature that occurred after the global collapse was underway and no science-based simulation capability exists to capture it. After the exterior facade began to fall downward at 6.9 s, the north face developed a line or “kink” near the end of the core at Column 76. As shown in Figure 5-205, the northeast corner then began to displace to the north at about 8.8 s, and the kink was visible at 9.3 s. The kink and rotation of the northeast façade occurred 2 s to 3 s after the exterior façade had begun to move downward, as a result of the global collapse. The simulations do show the formation of the kink, but any subsequent movement of the building is beyond the reliability of the physics in the model. (NCSTAR1A_p78)


3.5.3 Accuracy Appraisal

Given the complexity of the modeled behavior, the global collapse analyses matched the observed behavior reasonably well. The close similarity of the timing and the nature of the events up to the initiation of global collapse is strong confirmation of the extent and nature of the structural failures in the interior of the building and the accuracy of the four-step simulation process. The overall simulation of the collapsing building with damage better matched the video observations of the global collapse. The global collapse analysis confirmed the leading collapse hypothesis, which was based on the available evidence. (NCSTAR1A_p78)



McGrattan, K. B., C. Bouldin, and G. Forney. 2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster: Computer Simulation of the Fires in the World Trade Center Towers. NIST NCSTAR 1-5F. National Institute of Standards and Technology. Gaithersburg, MD, September. 1-9A.pdf nist.gov_p35


HERE DOWN is 1-9A.pdf NCSTAR 1-9A


4.4 Global Analysis Results with Fire Induced Damage AT 3.5 H ................................................... 105

4.5 Global Analysis Results without Debris Impact Damage............................................................ 106



EXECUTIVE SUMMARY

E.1 INTRODUCTION

As part of the National Institute of Standards and Technology (NIST) investigation into the collapse of World Trade Center (WTC) 7, NIST worked with Applied Research Associates, Inc. (ARA) under contract to conduct global collapse analyses of the WTC 7 building. The purpose of this work was to analyze the global response of WTC 7 to initial failure events due to fire and to analyze the resulting component and subsystem failures to determine the events that led to the global collapse. The nonlinear dynamic collapse analyses were performed using LS-DYNA, which is capable of explicitly modeling sequential failures, falling debris, and debris impact on other structural components. LS-DYNA is well suited for this type of analysis since it can model the dynamic sequential failure processes, including nonlinear material properties, nonlinear geometric deformations, material failures, and contact between the collapsing structural components. In addition, LS-DYNA has capabilities to include thermal softening of materials and thermal expansion. Analyses of the LS-DYNA model of the 47 story WTC 7 were used to evaluate the global response of the building to initial failure events due to fire, including component and subsystem failures, to determine the events that led to the global collapse. Four global analyses were performed:

• Analysis of the building response to debris impact damage, temperature effects (Case B, 4.0 h temperatures), and fire induced damage based on the 16 story ANSYS analysis (Case B, 4.0h damage).

• Analysis with earlier fire induced damage (Case B, 3.5 h) to determine if a lesser degree of induced damage state was sufficient to initiate a global collapse of WTC 7.

• Analysis with temperature effects (Case B, 4.0 h temperatures), and fire induced damage based on the 16 story ANSYS analysis (Case B, 4.0 h damage) but without the debris impact damage. The purpose of this analysis was to determine the contribution of debris impact to the global collapse sequence and whether WTC 7 would have collapsed solely due to the effects of the fires.

• Analysis for a classic progressive collapse, i.e., disproportionate structural damage from a single failure, without debris impact or fire-induced damage, to determine if a progressive collapse would occur following the removal of a section of Column 79 between Floors 11 and 13.

The key aspects of the model and the results of the three global analyses are summarized.

E.3

GLOBAL MODEL DEVELOPMENT

The model of WTC 7 for the global collapse analyses is shown in Figure E-1. The building was modeled primarily with shell elements. Beam elements were used to model the diagonal framing elements in the structural frame as well as for the frame of the penthouse structures. The nonlinear discrete elements were used for details of the connection components. Brick elements were used for some rigid masses representing large equipment. A summary of the size of the global collapse model of WTC 7 is presented in Table E–1.

NCSTAR_1-9A_Figure_E-1.png

NCSTAR_1-9A_Table_E-1.png


A key aspect of modeling the structural integrity of WTC 7 for global collapse was the strength characteristics of the various structural connections in the building. The structural frame for WTC 7 contained a variety of structural connection types. These connection modeling efforts were a significant part of the WTC 7 model generation effort. The strengths and damage tolerance of the connections varied based on details such as whether the connection had welds or bolts. Similarly, the relative strength of different connections would vary based on the directionality of the load (e.g. the ability of connections to resist vertical and lateral loads).


Models of the various connections were developed based on the fabrication shop drawings and incorporated into the subassembly and global models. One example was the seated connection between exterior columns and interior framing beams. The connections included plates welded between the column flanges above and below the framing member, and the beam was bolted to the plates. The seated connection model developed for LS-DYNA is shown in Figure E-2. Bolts in the LS-DYNA model were represented with discrete elements. The properties of the discrete elements were tailored to match the bolt specifications in the actual construction. By including these details the relative strength in both vertical and lateral loading could be reproduced.

Wind bracing trusses were added to the lower core and exterior. Belt truss bracing at mid-height of the building was also included. The belt truss at Floors 22 to 24, which was part of the exterior framing, was also included as beam elements merged directly into the columns.


A simplified model of the penthouse structures was also generated using beam elements. The penthouse structures are shown in red in Figure E-1. Attachment of the penthouse structure to the global WTC 7 model was achieved by merging the beam ends into supporting columns or other framing in the roof structure. In addition, the global model accounted for the various load transfer mechanisms at the lower floors of WTC 7.


NCSTAR_1-9A_Figure_E-2.png


E.5

MODEL INITIALIZATION AND LOADING SEQUENCE

The LS-DYNA model was initiated as follows to minimize any undue dynamic effects associated with loading sequence. First, gravity was applied slowly to the 47 floor structure over 4.5 s of elapsed simulation time to damp residual vibrations and eliminate dynamic response. Then the debris impact damage from the collapse of WTC 1 was applied to the structure instantaneously by removing damaged elements from the model that were no longer capable of bearing their loads. The structure was then allowed to damp residual vibrations for 2 s. Next, the structural temperatures were applied at the same time as the damage data from the ANSYS simulation and allowed to damp residual vibrations for 2 s. Last, the fire-induced damage obtained from the 16 story ANSYS analysis, including damage to floor beams, girders, and connections, was applied instantaneously. The heated, damaged structure was then free to react. The time at which the east penthouse began to descend was defined as 0.0 s, i.e., the beginning of the collapse of WTC 7. A chart showing the initialization sequence with the load curve profiles used is shown in Figure E-3.


Live load (LL) was distributed evenly in the concrete slabs and steel structure. This was accomplished by multiplying the modeled material densities with a common scale factor to achieve the desired total load. A 25 percent live load was used, based on estimates from the design live loads in the building at the time of collapse.


NCSTAR_1-9A_Figure_E-3.png


After gravity initialization, debris impact damage from the collapse of WTC 1 was applied to the global model instantaneously to approximate the dynamic event. The damage applied was isolated to two zones on the southern side of the building. The first zone is on the south face of the tower and southwest corner extending over Floors 5 to 16 of WTC 7. The other zone had much less damage and was on the south face near the top of the tower on Floors 44 to the roof. The analysis demonstrated that the remaining structure was able to redistribute the loads from the damaged zone and that the tower developed a new equilibrium state.


Temperatures applied in the global model analysis were Case B temperatures at either 3.5 or 4.0 hours (NCSTAR NIST 1-9, Chapter 10). Elevated temperatures occurred between Floors 7 and 14. Temperature profiles were mapped onto the LS-DYNA model as nodal properties and followed the sinusoidal load curve shown in Figure E-3.


The final step in the initialization process was to apply fire-induced damage from the 16 story ANSYS analysis. The ANSYS analysis estimated the damage that occurred as the fires grew and spread on Floors 7, 8, and 9 and Floors 11, 12, and 13. The LS-DYNA analysis, by comparison, considered only a temperature profile at the time when thermally-induced damage was transferred from the ANSYS analysis. The damage was added to the LS-DYNA model as the final step before simulating the structural response to the temperatures and damage.



E.6

GLOBAL COLLAPSE ANALYSES

Global Collapse Results with Debris Impact Damage

The 47 story model calculation showed that the combination of debris-impact damage, fire-induced damage, and thermal loads resulted in the global collapse of WTC 7. The key events that occurred in the global analysis are summarized below.


• -16.0 to -7.5 (0 to 8.5) s: Model was initialized.1

o Gravity, WTC 1 debris impact damage, and temperatures were applied to the LS-DYNA model.

o The structure sustained damage developed from the load redistribution due to debris impact damage and Case B, 4.0 h temperatures.

o The structure was stable at end of initialization.


• -7.5 (8.5) s: Fire-induced floor damage from the ANSYS analysis was added.

o Sections of Floors 13 and 14 collapsed in the northeast region around Columns 79, 80, and 81.

o Floor areas below, also weakened by fires, collapsed from the falling debris until Columns 79 and 80 were unsupported between Floors 5 and 15. Column 81 was unsupported between Floors 7 and 15.


• -1.3 (14.7) s: The Initial Failure Event occurred and started the Vertical Progression.

o Column 79 buckled between Floors 5 and 14.

o Columns 80 and 81 buckled quickly in succession, both within 1 second of Column 79 buckling.

o Following buckling of Columns 79, 80, and 81, the remaining column section above the buckled lengths began to move downward, and the floors sections above were pulled downward, first by Column 79, then by Columns 80 and 81.

o A global southward sway developed and grew in the upper floors, emanating from the collapsed east floor area.


• 0.0-1.5+ (16.0-17.5+) s: Kink in east penthouse roofline appeared.

o 0 (16.0) s: The floor collapse progressed upward, and a kink appeared near the middle of the east penthouse roofline on the north side.

o 0.7+ (16.7+) s: The east face near the top of the building deflected westward as floors surrounding Columns 79 to 81 pulled the exterior wall inward .


•2.0 (18.0) s: The east penthouse fell below the WTC 7 roofline.

o The east penthouse fell completely below the WTC 7 roofline.


• 2.7-5.3+ (18.7-21.3+) s: The Horizontal Progression started.

o 2.7 (18.7) s: The Truss 2 eastern diagonal member buckled due to the impact of falling debris, compromising the support of Columns 77, 78, and 78A.

o 3.3 to 3.7 (19.3 to 19.7) s: Columns 77 and 78 buckled as a result of the Truss 2 failure. Column 76 buckled due to the load redistribution and impact of falling debris.

o 3.9 to 6.1 (19.9 to 22.1) s: The remaining interior columns buckled sequentially, as loads transferred from adjacent buckled columns and as they were impacted by falling debris.

o 4.3+ (20.3+) s: West core columns buckled between Floors 9 and 13.


• 5.5-7.4 (21.5-23.4) s: Exterior Collapse occurred.

o 5.5 (21.5) s: Exterior column buckling began at the southwest corner Column 14, adjacent to the debris impact zone, between Floors 10 and 12.

o 5.5-7.4 (21.5-23.4) s: Exterior columns buckled along the south and west faces between Floors 7 and 14. The buckling of exterior columns rapidly spread to the north, and then east, faces.

o 7.4 (23.4) s: All interior and exterior columns were buckled in the lower floors.


• 5.7 to 7.7 (21.7 to 23.7) s: West Penthouse and Screening Wall fell below the WTC 7 roofline.

o Collapse in the western core caused the west penthouse and screening wall to fall below the building roofline.


• 6.2 to 8.5 (22.2 to 24.5) s: Global Vertical Motion began.

o 6.2 to 6.3 (22.2 to 22.3) s: The east side of the north face roofline began moving downward.

o 6.5 to 7.5 (22.5 to 23.5) s: Global vertical motion spread across to the west side of the north face roofline.

o 7.5+ (23.5+) s: The entire building was falling and the upper floors continued accelerating downward.

o At 8.5 (24.5) s: The roof was falling with velocity of approximately 10 m/s to 15 m/s.


• 8.59 (24.59) s: The calculation was terminated.


1 The times are presented as follows: t1(t2), where t1 is the time following the first observation of the descent of the east penthouse (NIST NCSTAR 1-9, Section 5.7) and t2 is the elapsed time from the start of the simulation.


Figure E-4 shows the progression of collapse as viewed from the northwest. The figures present contours of vertical displacement of the structure with a range between -2.0 m and 0 m (-80 in. to 0 in.).


At -1.1 (14.9) s, floor segments had started collapsing around Columns 79, 80, and 81 between Floor 5 and Floor 14. The floor collapses were due to fire-induced structural damage obtained from the ANSYS analysis.


At 2.3 (18.3) s, Columns 79 to 81 had buckled and the floor sections above were pulled downward by the unsupported columns, resulting in a vertical progression of collapse up to the east penthouse on the roof.


At 7.3 (23.3) s, column buckling had spread horizontally across the entire core. As the core columns fell downward, buckling of the exterior columns developed as the floors attached to the west face pulled inward at the lower floors. There were no fires on Floors 9 to 14 on the west floor area, so the floors were not thermally weakened in this area. Note that the north exterior wall appears generally intact, while most of the structure behind the north face is collapsing. Much of the interior structure and the south wall was falling downward at that point. By 8.6 (24.6) s, buckling had occurred in all the exterior columns, and global downward vertical movement had started. The calculation was stopped after the building had fallen downward approximately 10 floors, as global collapse occurred.


Initial Failure Event and Vertical Progression of Collapse

The initial failure event was the buckling of Column 79. This event was quickly followed by the buckling of adjacent Columns 80 and 81.


The floor framing structure was thermally weakened at Floors 8 to 14, with the most substantial damage occurring in the east region of Floors 12, 13, and 14. During the LS-DYNA temperature application cycle, combined thermal expansion and thermally degraded material properties resulted in beam and girder connection damage throughout the heated floor structures. The connection damage and buckled beam data transferred from the 16 story ANSYS analysis were then applied. The LS-DYNA analysis calculated the dynamic response of the floor failures and resulting impact on the surrounding structure. After the fire-induced ANSYS damage was applied, floor sections surrounding Columns 79 to 81 on Floors 13 and 14 collapsed to the floors below.


The thermally weakened floors below could not withstand the impact from the collapsing floors, and resulted in progressive floor collapses, which removed lateral support to Columns 79 to 81 over several floors. Eventually, critical conditions developed for column buckling due to a large unsupported length. Once Columns 79, 80, and 81 buckled, the column sections above Floor 14 began to descend downward and pulled the floor structures downward with them, thereby creating a vertical progression of floor collapses. The floors pulled on adjacent Columns 76, 77, and 79 until their connections failed. The east penthouse, which was supported by Columns 79, 80, and 81, fell downward.



NCSTAR_1-9A_Figure_E-4a.png


NCSTAR_1-9A_Figure_E-4b.png


Figure E-5 is a cutaway view that illustrates the structural condition surrounding Columns 79, 80, and 81 when they buckled. Included in the figure are resultant lateral displacements and column axial stress histories for the three columns. Note that a rapid lateral displacement developed, indicating buckling. Likewise, the column stresses indicated a rapid loss of stress at the time, signifying column buckling. The buckling of Column 80 and 81 was preceded by a slight increase in compressive stress due to load redistribution and tensile forces from floor systems being pulled downward by Column 79, and then by Column 80.


Figure E-6 illustrates the girder-to-column connection status at the time of buckling for Columns 79 through 81. The figure indicates large unsupported or partially supported (in one direction) lengths for each column at the time of buckling for each column. In addition to axial compressive loads that could have caused buckling, the columns were in a dynamic environment where lateral loads from falling debris and failure of girder connections occurred frequently. These lateral perturbations aided the onset of column buckling.







Comparison with Observables:

Observed collapse events (NIST NCSTAR 1-9, Chapter 5) are compared to the results of the LS-DYNA analysis. Table 4-1 lists events with corresponding times from observed events and from the analysis.

NCSTAR_1-9A_Table_E-2.png 

Given the complexity of the modeled behavior, the simulation closely matched the observed behavior. 1-9A.pdf nist.gov_p48



Global Results with Fire-induced Damage at 3.5 h

The global analysis with fire-induced damage obtained from the 16 story ANSYS analysis at 3.5 h instead of 4.0 h (as before) indicated that the level of damage was not sufficient to initiate global collapse of the building. The structure remained stable due to the lack of an initiating event (i.e., buckling of an interior column). 1-9A.pdf nist.gov_p48


Global Results without Debris Impact Damage

The global analysis without debris impact damage showed that WTC 7 would have collapsed solely due to the effects of the fires. The initiation of collapse was virtually the same as for the global analysis with debris impact damage. The initial failure event, vertical progression of failure, and early stages of the horizontal progression of failure occurred in the same order and at essentially the same times. The progression of horizontal failure in both analyses (with and without impact damage) was due to loss of lateral support as floors failed and to debris impact from adjacent failing floors and columns. As the horizontal progression moved from east to west, some differences began to occur relative to the analysis with impact damage. Some floors failed at mid-height of the building around Columns 73 to 75, leading to the buckling of some interior columns at this location, rather than at lower floors. The buckling at lower floor elevations in the analysis with impact damage was influenced by the damage around the west core columns when the debris impact damage was applied. Additionally, the screenwall fell downward prior to the west penthouse, which was consistent with the observed sequence of events. In the analysis without impact damage, the horizontal progression continued to move westward after Columns 70 through 72 buckled. In the analysis with impact damage, the horizontal progression moved westward until Columns 70 to 72 buckled, after which the damaged west core columns also buckled. For the analysis without debris impact damage, the timing of the exterior column buckling and onset of global collapse occurred at a slightly later time than was calculated for the analysis with impact damage. 1-9A.pdf nist.gov_p48


Global Results for a Classic Progressive Collapse Analysis

The classic progressive collapse analysis showed that WTC 7 collapsed when a section of Column 79 between Floors 11 and 13 was removed. The global collapse occurred in the absence of debris impact damage and fire-induced damage. The collapse sequence demonstrated a vertical and horizontal progression of failure upon the removal of the Column 79 section, followed by buckling of exterior columns, which led to the collapse of the entire building. 1-9A.pdf nist.gov_p49


NCSTAR_1-9A_Figure_E-5.png

NCSTAR_1-9A_Figure_E-6.png

NCSTAR_1-9A_Figure_E-7.png




3.6.1 System Specifications

All simulations were run on ARA’s high speed Linux Beowulf computer cluster. The system consisted of a head node and eight compute nodes. The head node housed two 64 bit AMD Opteron 250 2.4 GHz processors. The head node had 4 GB of RAM and housed 1.5 terabytes of RAID 5 disk storage. Each compute node housed two 64 bit AMD Opteron dual-core 285 2.6 GHz processors. Six of the compute nodes had 8 GB of RAM, and the remaining two nodes had 16 GB RAM. The compute nodes communicated using Silverstorm Infiniband Interconnects.

1A.pdf NIST NCSTAR 1A_p116


3.6.2 Model Size and Typical Run Time

The global model was constructed with 3,593,049 nodes and 3,045,925 total elements. The type and number of the elements were: 3,006,910 shells, 3,190 beams, 2,461 solids, 33,364 discrete elements, and 3,050 rigid elements. The global simulation was first initialized under gravity loading over 4.5 s of simulation time. Then damage and temperature initialization states were applied over 4 additional seconds. Collapse propagation in the global model required approximately 16 additional seconds after initialization, for a total of close to 25 seconds of simulated time, which generally took up to 8 weeks using 12 CPUs across 3 nodes (4 CPUs per node).

1A.pdf NIST NCSTAR 1A_p116


3.6.3 LS-DYNA Version

The analyses were performed with the LS-DYNA finite element code (Version mpp971dR4 beta, revision 41161). Use of this double precision code results in longer run times than a single precision version would require. However, a double precision version was necessary. In structural models where there are physically large dimensions (spatial extents of the building) compared to the characteristic element length, and long time scales compared to the explicit time step size, rounding errors can occur and accumulate when using a single-precision analysis. To address this issue, the double precision version of LS-DYNA was used for the global analyses.

1A.pdf NIST NCSTAR 1A_p116





E.7.1 WTC 1 Base Case Impact Analysis

The combined aircraft and tower model for the base case WTC 1 global impact analysis is shown in Figure E–25. The base case impact analysis was performed for a 0.715 s duration following initial impact of the aircraft nose with the north exterior wall. The analysis was performed on a computer cluster using twelve 2.8 GHz Intel Xeon processors, each on a separate node of the cluster. The run time for this analysis was approximately two weeks. The calculations were terminated when the damage to the towers reached a steady state and the motion of the debris was reduced to a level that was not expected to produce any significant increase in the impact damage. The residual kinetic energy of the airframe components at the termination of a global impact simulation was typically less than one percent of the initial kinetic energy at impact. (source_p75)


Chapter 9 FIRE SIMULATIONS

9.1 OVERVIEW

9.1.1 Approach

The major fires in WTC 7 were simulated using the Fire Dynamics Simulator, version 4 (McGrattan 2004), the same software that NIST used to simulate the fires in the World Trade Center towers (NIST NCSTAR 1-5F). These simulations were part of a sequence of estimates that included the debris damage from WTC 1, the subsequent fires, the heating of the structural elements, the structural response, and the eventual collapse. Temperature predictions from the fire model were passed to the structural model via the Fire Structure Interface, again using the same process as in the reconstruction of the collapse of the

World Trade Center towers (NIST NCSTAR 1-5G). As described in Chapter 5, sustained and/or late fires were observed on Floors 7 through 9 and 11 through 13 of WTC 7. There was a single observation of a late, small fire on Floor 14. In addition, there might have been concealed fires in the mechanical spaces on Floors 5 and 6. No fire was observed on the 10th floor. The fires on the upper floors (19, 22, 29, and 30) died out well before the collapse of WTC 7. The fire simulations for WTC 7 were conducted for each floor individually. Section 5.6 showed that there were no externally visible observations of flames passing from one floor to another, aside from what might have remained hidden in the debris-damaged area in the southwest of the building. The only floorto-floor connection within the tenant spaces was a flight of convenience stairs connecting the 11th and 12th floors. It was located near the southwest corner of the building, near the multi-floor debris damage zone, and essentially acted as an enlargement of the debris damage zone. The stairs were thus not included in the FDS simulations. The analysis described in Section 9.1.2 found that any late developing floor breaches would not have had a major effect on the structural heating. It would have been difficult to observe fires on the 5th and 6th floors, due to the partitions just behind the various windows or louvers (Chapter 3). Thus, the Investigation Team simulated a variety of fire scenarios and examined the outcome of each to determine whether it would have produced a visible telltale or whether it would have been severe enough to weaken any structural elements on those floors. The actual fires on Floors 7 through 9 and 11 through 13 were likely initiated when WTC 1 collapsed (Section 5.6). However, there was limited photographic evidence or on-site personal observation upon which to base simulation of the early development of these fires. While use was made of the appearance of flames and window breakage in photographs and videos in formulating the simulations, the Investigation Team realized that the absolute timing of the simulations might not align exactly with the timing of the fires on September 11, 2001. Using a timeline reconstructed from the limited photographic evidence and eyewitness testimony, fire simulations were conducted for Floors 7 and 8. For the simulations, the fires on these two floors were assumed to have initiated at a time designated as noon. There was no evidence of visible fire growth on these floors during the morning; and, even at 2:00 p.m., the fire on the 7th floor was still in the southwest section of the building. The computations were continued for six hours, since the time in the structural simulations at which the building would become structurally unstable was not yet known. The fire on the 9th floor was not observed until about 4:00 p.m., approximately one hour after fire was first observed on the 8th floor. Since there was no information about the history of the 9th floor fire prior to 4:00 p.m., and since there was limited knowledge of the layout of combustible materials on this floor, this fire was approximated by displacing the 8th floor fire one hour later. The layouts of the 11th and 12th floors were essentially identical. The observation of fires on the 11th floor indicated that they were similar to those on the 12th floor, but lagged by about an hour in reaching equivalent positions along the building perimeter. The intense fires on the 13th floor appeared after about 2:30 p.m. on the east face and the east side of the north face. These fires also appeared similar to those on the 12th floor, but lagged behind them by about a half hour. Thus, these fires were simulated using the 12th floor fires, suitably offset in time. On the 14th floor, a small fire was observed at the far east of the north face at about 5:03 p.m. It was not

observed at 4:52 p.m., nor at 5:09 p.m. On September 11, 2001, the 14th floor was unoccupied and should have had a very small combustible fuel load. From this evidence, the Investigation Team deduced that it was unlikely that there had been a sustained fire of appreciable heat output. Thus, no fire simulation was performed for this floor. The calculations were very similar to those conducted for WTC 1 and WTC 2 (NIST NCSTAR 1-5F) and, therefore, only a brief description of the major assumptions is included. It should be remembered that there were far fewer photographs of WTC 7 than of the towers. Thus, while the severity of the simulated WTC 7 fires (as established from the visual evidence) is reasonable, the details of these fires are not as precise as for the fires in the towers. 1-9 Vol 2.pdf NCSTAR_1-9A_Volume_2_p23



http://www.youtube.com/watch?v=M-yuQeeYkq8


http://community2.myfoxdfw.com/_Texas-Engineer-on-911/BLOG/1614574/78592.html


http://911blogger.com/node/22152


http://www.prisonplanet.com/engineer-fires-torpedos-at-the-911-lie.html


http://world911truth.org/ae911truth-engineer-fires-torpedos-at-the-911-official-story/

Personal tools
Namespaces

Variants
Actions
Navigation
Toolbox