Professional Documents
Culture Documents
Steven W. Kirkpatrick
Robert
T. Bocchieri
Fahim Sadek
Robert A. MacNeill
Samuel Holmes
Brian D. Peterson
Robert W. dike
Claudia Navarro
Technology Administration
U.S.
Deparimeni
of
Commerce
NISTNCSTAR1-2B
Federal Building and Fire Safety Investigation of the
World Trade Center Disaster
Steven W. Kirkpatrick
Robert
T. Bocchieri
Fahim Sadek
Robert A. MacNeill
Samuel Holmes
Brian D. Peterson
Robert W. Cilke
Claudia Navarro
Technology Adminisirotion
U.S.
Deportmeni of Commerce
NISTNCSTAR1-2B
Federal Building and Fire Safety Investigation of the
World Trade Center Disaster
Steven W. Kirkpatrick
Robert T. Bocchieri
Fahim Sadek
Building
and
Robert A. MacNeill
Samuel Holmes
Brian D. Peterson
Robert W. Cilke
Claudia Navarro
Septennber 2005
i
%
U.S. Department of
Commerce
Technology Administration
Michelle O'Neill, Acting Under Secretary for Technology
National Institute of Standards and Tecfinology
William Jeffrey. Director
Disclaimer No.
Certain commercial entities, equipment, products, or materials are identified in this document in order to describe a
procedure or concept adequately or to trace the history of the procedures and practices used. Such identification is
not intended to imply recommendation, endorsement, or implication that the entities, products, materials, or
equipment are necessarily the best available for the purpose. Nor does such identification imply a finding of
negligence by the National Institute of Standards and Technology.
fault or
Disclaimer No. 2
The
policy of NIST is to use the International System of Units (metric units) in all publications.
however, units are presented in metric units or the inch-pound system, whichever is prevalent
Disclaimer No. 3
In this
in
document,
the discipline.
Pursuant to section 7 of the National Construction Safety Team Act, the NIST Director has determined that certain
evidence received by NIST in the course of this Investigation is "voluntarily provided safety-related information" that is
"not directly related to the building failure being investigated" and that "disclosure of that information would inhibit the
voluntary provision of that type of information" (15
In
USC
7306c).
provided to
in
been
Disclaimer No. 4
NIST takes no position as to whether the design or construction of a WTC building was compliant with any code
due to the destruction of the WTC buildings, NIST could not verify the actual (or as-built) construction, the
properties and condition of the materials used, or changes to the original construction made over the life of the
buildings. In addition, NIST could not verify the interpretations of codes used by applicable authorities in determining
since,
compliance when implementing building codes. Where an Investigation report states whether a system was
designed or installed as required by a code provision, NIST has documentary or anecdotal evidence indicating
whether the requirement was met, or NIST has independently conducted tests or analyses indicating whether the
requirement was met.
Use
No
in
Legal Proceedings
part of
any report
mentioned
in
resulting from a
Team
USC
National Institute of Standards and Technology National Construction Safety Team Act Report 1-2B
Natl. Inst. Stand. Technol. Natl. Constr. Sfty. Tm. Act Rpt. 1-2B, 496 pages (September 2005)
CODEN: NSPUE2
U.S.
WASHINGTON: 2005
Abstract
(WTC)
report
was
and
interior core
systems due to
to structural
initial
initiation analysis.
(2) the subassembly level, and (3) the global level to estimate the probable
damage
to the towers
due
to
Simplified analyses were also used to support the development of the global finite
aircraft impact.
element models. Analysis of uncertainties using the component and subassembly analyses were
conducted
most
models
As
tower
steels
and
damage
aircraft
damage
most
estimates for
WTC
and
Based on
in the global
WTC 2.
were varied
aircraft
test results
of the
in the exterior
walls and core columns of the towers, weldment metal, bolts, reinforced concrete, aircraft materials, and
nonstructural contents were considered.
and
The
in the global
WTC
drawings and the structural databases of the towers developed within the framework of the baseline
structural
towers
performance analysis. The tower models included the primary structural components of the
in the
impact zone, including exterior walls, floor systems, core columns, and connections.
of the
aircraft
elsewhere. The models also included the nonstructural building contents, such as partitions and
workstations, in the path of the aircraft debris.
aircraft
structural information,
aircraft engines,
the aircraft.
Boeing 767
detailed analysis
was
aircraft
aircraft.
carried out to estimate the fuel distribution in the aircraft wings at the
time of impact.
The
WTC towers and Boeing 767 aircraft are complex structural systems.
to include all
In the global
model
both the aircraft and towers. This approach, however, results in very large models. The component and
NISTNCSTAR
1-2B,
WTC
in the analysis.
Investigation
to
iii
Abstract
The primary
objectives of the
and
(2)
develop the simulation techniques required for the global analysis of the
to
aircraft
aircraft
WTC towers, including variations in mesh density and numerical tools for modeling fluid-structure
interaction for fuel impact
finely
coarsely
shell
impacting a
The
the
initial
strip
WTC
1,
addition, a large
to the core
that captured
with
An
segment of an
an aircraft engine
is
flight
impacted
1,
which impacted
WTC 2.
In
to start
to relatively
of the tower.
and progress
is
An example
fuel.
two videos
for
aircraft structures
roll
accuracy
The
initial
building contents, connection behavior, modeling uncertainties, etc. Sensitivity analyses were performed
at the
levels.
damage
and
to
of
analysis
was
of the
following the aircraft impacts using the global tower and aircraft models. This assessment included the
estimation of the structural
damage
that
nonstructural contents such as partitions, workstations, aircraft fuel, and other debris that influenced the
to the
the primary
method by
towers was estimated. The global analyses included, for each tower, a "base
They
also provided
more and
more and
damage
damage
less severe
less severe
scenarios provided a range of damage estimates of the towers due to aircraft impact. Comparisons
results
all
simulations due to the uncertainties in exact impact conditions, the imperfect knowledge of the interior
reasonably well.
aircraft
other studies
was
also
provided.
iv
NISTNCSTAR
1-2B,
WTC Investigation
Abstract
to provide
finite
element impact
analyses. These included: (1) the analysis of the overall aircraft impact forces and assessment of the
relative importance of the airframe strength
effects of the energy in the rotating engine
and weight
damage and
damage
(3) the
residual strength
core columns
compared
to
aircraft
NISTNCSTAR
1-2B,
WTC
World Trade
Investigation
failure, finite
Center.
Abstract
vi
left
blank.
NISTNCSTAR
1-2B,
WTC Investigation
Table of Contents
Abstract
iii
List of Figures
xiii
List of Tables
xxv
List of Acronyms
and Abbreviations
xxvii
xxix
Preface
Acknowledgments
Executive
Chapter
xxxix
Summary
xli
Introduction
Chapter 2
2.1
Introduction
2.2
2.3
2.2.1
2.2.2
2.2.3
Summary of Steel
2.2.4
Failure
Models
Constitutive Properties
Models
11
14
15
2.3.1
Mesh Refinement
2.3.2
Weld Zone
2.3.3
Effects
Constitutive
Modeling
16
19
23
25
2.4
2.5
2.6
34
2.7
References
35
Models
31
Chapter 3
37
3.1
Introduction
37
3.2
37
3.2.
3.2.2
NISTNCSTAR
1-2B.
WTC
Model Development
Investigation
39
44
vii
Table of Contents
3.3
3.2.3
3.2.4
Interior Contents
3.2.5
49
Model Development
52
53
60
References
Chapter 4
Aircraft
Model Development
61
4.1
Introduction
61
4.2
62
4.2.
4.3
Electronic Surface
62
4.2.2
Aircraft Inspection
4.2.3
Pratt
4.2.4
Aircraft
62
Models
66
Company
67
Model Development
69
4.3.1
72
4.3.2
Wing
76
4.3.3
Section
4.4
4.5
References
at
82
84
Impact
88
Chapters
Component
5.1
91
Analysis Methodology
92
Exterior
93
93
5.2.2
5.2.3
101
5.2.4
104
98
5.3
108
5.4
5.5
Aircraft
5.6
viii
91
Introduction
5.1.1
5.2
Level Analyses.
Wing
Section
Component Analyses.
10
1 1
5.5.1
116
5.5.2
124
5.5.3
126
Summary
130
NIST NCSTAR
1-2B,
WTC Investigation
Table of Contents
5.7
References
136
Chapter 6
137
6.1
Introduction
137
6.2
137
6.3
139
42
6.3.1
6.3.2
Wing
150
6.4
Summary
155
6.5
References
155
Chapter 7
157
7.1
Introduction
157
7.2
158
7.2.
Videos Used
in the
Analysis
58
7.2.2
159
7.2.3
163
7.3
166
7.4
172
7.5
Summary
173
7.6
References
174
Chapters
175
Uncertainty Analyses
8.1
Introduction
175
8.2
176
8.2.1
Engine
8.2.2
8.2.3
176
182
186
8.3
191
8.4
References
193
Chapters
195
Introduction
195
9.1
NISTNCSTAR
1-2B,
WTC
Investigation
ix
Table of Contents
9.2
195
9.3
Models
197
9.4
WTC
201
WTC
WTC
Damage
218
228
9.4.
9.4.2
Structural
208
9.5
235
246
256
257
276
290
292
312
327
327
328
330
331
331
332
334
335
9.1
338
340
9.1 1.2
WTC
Comparison with Observables on WTC 2.
9.1 1.3
Summary
363
9.1 1.1
349
363
9.13 References
368
Chapter 10
369
10.1 Introduction
369
369
370
NISTNCSTAR
1-2B.
WTC Investigation
Table of Contents
372
374
10.5.1
374
10.5.2
375
10.6 References
377
Chapter 11
Summary
379
379
11.1 Introduction
1
11.3
.4
1.5
Modeling
379
380
Model Development
381
382
383
1.7
1 1
.8
383
Uncertainty Analyses
384
385
385
1.9.2
387
388
1 1
.9.4
390
1 1
.9.5
390
391
Appendix A
Still
NISTNCSTAR
1-2B,
WTC
Investigation
in
Chapter 7
393
xi
Table of Contents
xii
left
blank.
NISTNCSTAR
1-2B,
WTC
Investigation
List
Figure P-1
The
OF Figures
and
fire safety
investigation of the
WTC
xxxi
disaster
Figure 2-2.
Figure 2-3.
Test data and true stress-strain conversion for the 75 ksi steel
Tabular true stress-strain constitutive model curve for the 75 ksi steel
Figure 2-1
Figure
Example
2^.
finite
Figure 2-5.
Calculated tensile test response with necking for the 75 ksi steel
Figure 2-6.
10
cuwes
for the
10
steel
Figure 2-7.
stress-strain
steel
steel
11
Figure 2-8.
12
Figure 2-9.
steel
Figure 2-10.
12
steel
Figure 2-11.
13
steel
steels
Figure 2-12.
Comparison of rate
Figure 2-13.
17
Figure 2-14.
18
Figure 2-15.
Mesh
test
18
Figure 2-1
19
Figure 2-17.
20
Figure 2-1 8.
Figure 2-1 9.
Drop
Figure 2-20.
Models developed
Figure 2-21
22
Figure 2-22.
23
Figure 2-23.
Measured
24
Figure 2-24.
Figure 2-25.
Intact
NISTNCSTAR
1-2B,
6.
effects
model and
test data
15
test
HAZ materials
fracture analysis
column weld
fracture analysis
WTC
Investigation
20
21
22
24
26
xiii
List of
Figures
unconfmed compression
27
Figure 2-26.
Figure 2-27.
test
test
27
data
Figure 2-28.
28
concrete behavior
Figure 2-29.
ksi
and 4
ksi
concrete
29
Figure 2-30.
30
Figure 2-3
30
Figure 2-32.
31
Figure 2-33.
aluminum
alloys
32
Figure 2-34.
aluminum
alloys
32
Figure 2-35.
True
stress-strain curves
aircraft
aluminum
alloys
33
Figure 2-36.
True
stress-strain curves
aircraft
aluminum
alloys
33
Figure 3-1
Model user
column generator
38
39
Figure 3-2.
User interface
Figure 3-3.
Model of the
Figure 3-4.
41
Figure 3-5.
42
Figure 3-6.
Model of the
WTC
WTC
40
Figure 3-7.
Model
Figure 3-8.
Model of the
Figure 3-9.
Example models
Figure 3-10.
Figure 3-1
detail
WTC
beam connections
44
45
WTC model
Impact face of the WTC 2 model
1
floors
91-101
46
floors
75-86
47
Model of spandrel
Figure 3-13.
Figure 3-14.
Model of a
Figure 3-15.
at exterior
Figure 3-1 6.
at
Figure 3-17.
Figure 3-18.
Figure 3-19.
xiv
44
94th-98th core
Figure 3-12.
Figure 3-20.
floor
43
slab)
48
truss floor
model
48
segment
floor of
49
wall
50
50
core perimeter
WTC
51
floor of
WTC
52
floor of
54
NIST NCSTAR
51
1-2B,
WTC Investigation
List of
Figure 3-28.
WTC tower
Interior structures and contents of the WTC
global impact model
Building contents in the WTC global impact model
Truss floor components in the WTC global impact model
Far field truss floor model in the WTC towers
Multi-floor global impact model of the WTC 2 tower
Interior structures and contents of the WTC 2 global impact model
Building contents in the WTC 2 global impact model
Figure 4-1
Figure 3-21.
Figure 3-22.
Figure 3-23.
Figure 3-24.
Figure 3-25.
Figure 3-26.
Figure 3-27.
Figures
55
56.
57
57
58
59
59
60
62
Figure 4-2.
Superimposed
Figure 4-3.
64
Figure 4-4.
65
Figure 4-5.
66
Figure 4-6.
Cutaway of a
Manual
Pratt
and Whitney
PW4000
63
67
Figure 4-7.
Economy
Figure 4-8.
Figure 4-9.
Finite element
Figure 4-1 0.
Figure 411.
Figure 4-12.
Empennage model
Figure 4-13.
73
Figure 4-14.
Underside of the 767 airframe model (skin removed) showing retracted landing gear
74
Figure 4-15.
Complete wing
74
Figure 4-1 6.
Wing
Figure 4-1
Model of fuselage
7.
fuel load at
for the
69
interior
Figure 4-19.
Wing
for
72
model
aircraft
Boeing 767
71
aircraft
73
model
75
model
70
time of impact
767-200ER
Figure 4-18.
Figure 4-20.
68
structural
section
and dimensions
wing
component
76
structures
level
75
77
Simplified wing section model and impact analysis used for the effective skin model
development
78
79
Figure 4-22.
80
Figure 4-23.
Figure 4-24.
NISTNCSTAR
1-2B,
Figure 4-2 1
WTC
Investigation
column
different
modeling options
80
81
xv
List of Figures
Figure 4-25.
Damage
Figure 4-26.
Pratt
Figure 4-27.
PW4000
82
Figure 4-28.
PW4000
83
Figure 4-29.
Pratt
Figure 4-30.
Flammable material
locations in a
Figure 4-3
in the
&
&
to an exterior
column from
81
Whitney
PW4000
82
83
Boeing 767
85
85
Figure 4-32.
Approximate
Figure 4-33.
Wing
Figure 4-34.
Figure 4-35.
Approximate
Figure 5-1
94
Figure 5-2.
95
Figure 5-3.
Obsei"ved impact
96
Figure 5-4.
86
87
Exterior
87
smooth and
damage
in the
recovered
88
level flight
magnitude
Figure 5-5.
Exterior
97
magnitude
Wide
Figure 5-6.
flange core
98
displacement magnitude
99
Displacement and kinetic energy comparison for wide flange core column wing
impact analysis
Figure 5-7.
Box
Figure 5-8.
100
magnitude
100
Displacement and kinetic energy comparison for box core column wing impact
Figure 5-9.
analysis
101
Figure 5-10.
Figure 5-1
102
Figure 5-1
2.
103
101
Figure 5-13.
Figure 5-14.
104
Figure 5-15.
105
Figure 5-16.
106
Figure 5-1
7.
107
Figure 5-18.
xvi
103
NISTNCSTAR
1-2B,
WTC
07
Investigation
01
List of
Figure 5-19.
Figure 5-20.
Figure 5-2 1
Figures
108
columns
109
1 1
Figure 5-22.
1 1
Figure 5-23.
112
Figure 5-24.
Figure 5-25.
Figure 5-26.
115
Figure 5-27.
Figure 5-28.
Empty wing
Figure 5-29.
Figure 5-30.
Figure 5-3 1
shell
section impact
118
at
mph
119
Figure 5-33.
Figure 5-34.
Wing
Figure 5-35.
Figure 5-36.
Aluminum model
two material
120
121
122
failure criteria
mesh fragmentation
123
components
124
Calculated impact of a coarse-mesh empty wing section with two exterior panels
at
442 mph
125
Impact damage to the exterior panel from a coarse-mesh empty wing section
mph
Figure 5-40.
Wing segment,
Figure 5-41
Wing segment
Figure 5-42.
at
442
Figure 5-39.
fiael
in the small
fuel,
126
wing segment
127
128
128
mesh wing
modeled using
ALE
131
elements
Figure
5^3.
Figure 5^14.
Figure
16
118
Figure 5-32.
Figure 5-38.
14
117
exterior panels
Calculated impact of a fine-mesh empty wing section with two exterior panels
442
Figure 5-37.
113
mesh wing
particles
132
133
5^5. SPH
analysis of structural
damage and
133
Figure 5-46.
Figure 5-47.
SPH
analysis of structural
damage and
view)
134
5^8. ALE
analysis of structural
damage and
view)
135
Figure
NISTNCSTAR1-2B.WTC
Investigation
134
xvii
List of
Figures
Figure 6-1.
Figure 6-2.
139
Figure 6-3.
Final
140
Figure
6^.
model
138
analysis
141
Figure 6-5.
Final subassembly
Figure 6-6.
143
Figure 6-7.
144
Figure 6-8.
Speed history
Figure 6-9.
146
Figure 6-10.
147
Figure 6-1
148
Figure 6-12.
Tmss
149
Figure 6-13.
Final subassembly
Figure 6-14.
Figure 6-15.
Impactor
Figure 6-16.
Exterior wall
Figure 6-1
7.
Figure 7-1
1.
floor
for the
damage from
model
145
momentum histories
damage
143
for the
150
analysis
151
152
153
154
158
Figure 7-2.
160
Figure 7-3.
161
162
Figure 7-5.
163
Figure 7-6.
164
Figure 7-7.
Figure
7^.
face of
WTC
aircraft
and fuselage
167
An
Figure 7-8.
168
Figure 7-9.
face of
WTC 2
169
Figure 7-10.
169
Figure 7-1
An
orientation and trajectory of UAL 175 that matches the impact pattern (vertical
approach angle = 6, lateral approach angle = 13)
171
Figure 7-12.
An
approach angle =
xviii
6, lateral
UAL
171
NISTNCSTAR
1-2B,
WTC Investigation
List of
Figure 7-13.
initial lateral
approach angle of 13
Figure 8-1
Figures
172
into
177
Figure 8-2.
Figure 8-3.
Main
Figure 84.
182
Figure 8-5.
Figure 8-6.
Main
Figure 8-7.
Figure 8-8.
188
Figure 8-9.
Main
Figure 8-10.
Figure 9-1.
Figtire 9-3.
Figure 9-4.
Figure 9-5.
Figure 9-6.
181
Figure 9-9.
Figure 9-10.
.
199
200
201
202
204
damage
Figure 9-13.
207
209
210
212
Figure 9-16.
WTC
Base case impact damage to the WTC
Base case impact damage to the WTC
Base case impact damage to the WTC
Figure 9-17.
213
214
columns
levels in core
to the
206
Classification of
Figure 9-15.
89
Figure 9-12.
Figure 9-14.
87
190
Figure 9-8.
84
185
Figure 9-7.
180
Figure 9-2.
Figure 9-1
column
to a core
215
216
217
218
Figure 9-19.
222
Figure 9-20.
223
Figure 9-21
224
Figure 9-18.
NISTNCSTAR
220
221
1-2B.
WTC
Investigation
xix
List
of Figures
Figure 9-22.
225
Figure 9-23.
226
Figure 9-24.
227
Figure 9-25.
229
250
Figure 9-42.
251
9^3.
252
Figure 9-44.
253
Figure 9-45.
254
Figure 9-46.
255
WTC
More severe WTC
Figure 9-26.
Figure 9-27.
Figure 9-28.
Figure 9-29.
Figure 9-30.
Figure 9-31.
Figure 9-32.
Figure 9-33.
Figure 9-34.
Figure 9-35.
Figure 9-36.
Figure 9-37.
Figure 9-38.
Figure 9-39.
Figure 9-40.
Figure
Figure
9^1
Figure
9^7. More
Figure
9^8.
Figure
9^9. More
severe
240
241
242
243
244
245
247
248
249
261
More
XX
239
Figure 9-54.
Figure 9-56.
238
Figure 9-55.
236
259
Figure 9-53.
235
Figure 9-52.
Figure 9-5 1
232
WTC
More severe impact response of the WTC
More severe impact response of the WTC
More severe impact damage to the WTC
Figure 9-50.
230
severe impact
damage
to the
for the
WTC
More severe impact damage to the WTC
More severe impact damage to the WTC
severe impact
damage
to the
263
exterior wall
1
core
264
core columns
265
more severe
WTC
impact
266
268
269
270
271
NISTNCSTAR1-2B, WTC
Investigation
List of
Figure 9-61.
WTC
Calculated more severe WTC
Calculated more severe WTC
Calculated more severe WTC
Calculated more severe WTC
Figure 9-62.
Figure 9-57.
Figure 9-58.
Figure 9-59.
Figure 9-60.
272
273
274
275
276
Figure 9-65.
WTC
Less severe impact response of the WTC
Less severe impact response of the WTC
Less severe impact damage to the WTC
Figure 9-66.
Figure 9-67.
Figure 9-63.
Figure 9-64.
Figure 9-68.
Figure 969.
Figure 9-70.
Figure 9-71
Figure 9-72.
Figure 9-73.
Figure 9-74.
Figure 9-75.
Figure 9-76.
Figure 9-77.
Figure 9-78.
Figure 9-79.
Figure 9-80.
Figure 9-81.
Figure 9-82.
Figure 9-83.
277
exterior wall
1
core
279
core columns
280
WTC
impact
283
284
285
303
286
287
288
289
290
more severe
1-2B,
Figure 9-88.
severe impact
WTC
damage
Investigation
295
297
298
299
300
302
303
NISTNCSTAR
Figure 9-87.
More
293
of the figure)
Figure 9-89.
Figure 9-86.
282
Figure 9-85.
281
from the
Figure 9-84.
Figures
to the
304
305
306
307
308
309
xxi
List of
Figures
310
Figure 9-113. Comparison of base case impact damage to the exterior wall
336
Figure 9-1 14. Comparison of base case impact damage to the core columns
337
Figure 9-1 15. Comparison of base case impact damage to floor truss (front view)
338
WTC towers
Figure 9-117. Representative floor plan (based on floors 94 to 95 of WTC
Figure 9-118. Base case aircraft debris distribution in WTC
Figure 9-119. More severe damage aircraft debris distribution in WTC
Figure 9-120. Less severe damage aircraft debris distribution in WTC
from the more severe damage global analysis
Figure 9-121. Damage to the south face of WTC
339
Figure 9-90.
1
311
312
313
315
316
317
3
319
320
321
322
323
324
325
326
327
328
329
331
1 1
332
333
334
1 )
at the
Figure 9-123. Landing gear found embedded in exterior panel knocked free from
Figure 9-124. Base case stairwell disruption
xxii
in
WTC
340
341
342
343
344
345
WTC
346
347
NISTNCSTAR
1-2B,
WTC Investigation
List of Figures
WTC
damage
(front view)
348
WTC 2
Figure 9-127. Documented damage to the 81st floor of the northeast comer of WTC 2
Figure 9-128. Base case analysis on the 81st floor of the northeast comer of WTC 2
Figure 9-129. Base case stairwell dismption on floor 78 in WTC 2
Figure 9-130. Base case damage aircraft debris distribution in WTC 2
Figure 9-131. Aircraft debris distribution in the more severe WTC 2 impact
Figure 9-132. Aircraft debris distribution in the less severe WTC 2 impact
Damage
to the Northeast
Comer of the
Exterior Wall of
Figure 9-133. Starboard engine fragment trajectory in the base case global analysis of
Figure 9-134. Speed of the
aft
349
350
351
352
354
355
356
WTC 2
358
359
analysis
WTC 2
Figure 10-1.
Figure
0-2.
357
362
362
365
366
371
effects
372
axial preload
373
Figure 10-3.
Figure 104.
Figure 10-5.
Core column
Figure 106.
P-1 cur\ es
NISTNCSTAR
1-2B.
WTC
and expanded
Investigation
fuel cloud
impact loading
375
376
377
xxiii
List
of Figures
XXIV
left
blank.
NISTNCSTAR
1-2B.
WTC Investigation
List of Tables
Table P-1.
Table P-2.
Table 2-1
of the
WTC disaster
WTC Investigation
- WTC tower
xxx
xxxiii
15
steels
Table 2-2.
26
Table 3-1
53
WTC towers
Table 3-2.
Summary of the
Table 4-1
67
Table 42.
69
Table 4-3.
Boeing 767-200ER
Table 4-4.
77
Table 4-5.
84
Table 5-1
97
aircraft
60
model parameters
Table 5-2.
Ill
Table 5-3.
Wing with
127
Table
5^.
Table 7-1
Fuel and
fuel
air
model parameters
129
UAL
59
Table 7-2.
Measured
Table 7-3.
Summary of measured
Table 7-4.
Summary
Table 7-5.
Table 7-6.
173
Table 7-7.
Table 8-1.
178
Table 8-2.
Fractional factorial
WTC towers
1 1
2'^"^
165
166
170
170
173
179
Table 8-3.
183
NISTNCSTAR
1-2B.
WTC
Investigation
xxv
List of
Tables
Fractional factorial 2'^"^ experimental design (with centeipoint) for the wing-panel
Table 8-4.
impact analyses
185
Table 8-5.
187
Table 8-6.
1 1
impact analyses
189
Table 8-7.
192
Table 9-1
Summary of the
Table 9-2.
Table 9-3.
Table 9-14.
WTC impact
impact
Fuel and aircraft debris distribution for the base case WTC
Summary of core column damage for the base case WTC 2 impact
Fuel and aircraft debris distribution for the base case WTC 2 impact
Input parameters for additional WTC
global impact analyses
Core column response for the more severe WTC impact
Core column response for the less severe WTC impact
Input parameters for additional WTC 2 global impact analyses
Core column response for the more severe WTC 2 impact
Core column response for the less severe WTC 2 impact
Comparison with observables from WTC
Comparison with observables from WTC 2
Table 9-15.
Comparison of damage
Table 10-1.
Table
9^.
Table 9-6.
Table 9-7.
213
228
256
257
267
278
Table 9-10.
1
241
Table 9-9.
Table 9-12.
Table 9-13.
XXVI
Table 9-8.
Table 9-1
198
Table 9-5.
291
301
;
318
363
to core
98
363
367
371
NISTNCSTAR
1-2B,
WTC Investigation
List of
Acronyms
A A
AA
American Airlines
15
Arbitrary-Lagrangian-Eulenan
American oocieiy
lor
Ur U
rtMA
Federal
14
A y
rlAZ.
CD
LbCT/"'
L
Mil
JVLUr
medium
Technology Corporation
density fiberboard
SPH
Smoothed
UAL
United Airlines
ULD
WTC
WTC
Particle
Hydrodynamics
(North Tower)
WTC 2
WTC 7
Abbreviations
plus or minus
ft
foot
square foot
cubic foot
gal
gallon
GHz
gigahertz
in.
inch
NISTNCSTAR
1-2B.
WTC
Investigation
xxvii
List of Acronyms
and Abbreviations
kip
a force equal to
ksi
,000 pounds
lb
pound
min
minute
mph
ms
millisecond
psf
psi
second
xxviii
NISTNCSTAR
1-2B,
WTC
Investigation
Preface
terrorist attack
the
(WTC) on September
Team went
to the site
away from
May
report in
2002. fulfilling
its
who
7, as
and began
its
assessment.
Team
issued
its
to identify areas of
On August
On
disaster.
October
1,
To
NIST WTC
Team Act.
Investigation
investigation of the
To
The
investigation of the
fire safety
Team Act
(Public
Law
WTC
107-231),
was
building and
Construction Safety
its
FEMA,
that
WTC disaster.
Improvements
Improved
Recommended
Improved public
tools
in the
way
officials;
aircraft
2.
WTC
and how WTC
all
injuries
and
and
7 collapsed;
fatalities
location,
emergency response;
3.
4.
1, 2,
and
7;
and
and
fire
NISTNCSTAR
1-2B,
WTC
Investigation
xxix
Preface
NIST
is
a nonregulatory
is
on
is to
fact finding.
NIST
life
United
in the
wake of any
life.
NIST
does not have the statutory authority to make findings of fault nor negligence by individuals or
organizations. Further, no part of any report resulting from a
NIST
from an investigation under the National Construction Safety Team Act may be used
for
damages
Law
arising out of
in
USC
281a, as
in
any
suit or action
amended by PubHc
107-231).
E.
whose
interdependent projects
is
available at http://wtc.nist.gov.
Table P-1, and the key interdependencies among the projects are illustrated
fire
Lew
project
in Fig.
Damage
Analysis; Project
summarized
P-1.
WTC
disaster.
protection, and
is
Project Purpose
eight
A detailed description of
1, 2,
and
7.
W. Gayle
recovered from
WTC
and
fate
2,
WTC
and 7 and
1.2. and
7.
emergency response,
G. Gann
P.
McAllister
Analyze the response of the WTC towers to fires with and without
aircraft damage, the response of WTC 7 in fires, the performance
of composite steel-trussed floor systems, and determine the most
probable structural collapse sequence for WTC 1, 2, and 7.
D. Averill
Document
of the
terrorist attacks
WTC
7.
Lawson
XXX
J.
Randall
on
WTC
NISTNCSTAR
1-2B,
WTC
Investigation
Preface
Team
Act.
initial
fire
safety
The
and
pubUc
solicitation.
These were:
Paul Fitzgerald, Executive Vice President (retired)
FM Global,
Chair
David
Collins, President,
Inc.
NISTNCSTAR
1-2B,
WTC Investigation
Inc.
xxxi
Preface
Inc.
'
.
Kathleen Tiemey, Director, Natural Hazards Research and AppHcations Information Center,
University of Colorado at Boulder
Forman Williams,
Diego
This National Constmction Safety
Investigation and
commentary on
NIST
has benefited from the work of many people in the preparation of these reports, including the National
Construction Safety
Public Outreach
During the course of this Investigation, NIST held public briefings and meetings
solicit input
Table P-2) to
(listed in
from the public, present preliminary findings, and obtain comments on the direction and
progress of the Investigation from the public and the Advisory Committee.
NIST maintained
a publicly accessible
Web
site
at http://wtc.nist.gov.
The
site
NIST's
WTC
Public-Private
Response Plan
WTC buildings has led to broad reexamination of how tall buildings are designed,
constructed, maintained, and used, especially with regard to major events such as fires, natural disasters,
and
Reflecting the enhanced interest in effecting necessary change, NIST, with support
terrorist attacks.
in place a
is
to
develop and
implement the standards, technology, and practices needed for cost-effective improvements
to the safety
and security of buildings and building occupants, including evacuation, emergency response procedures,
The
strategy to
meet
this goal is
for cost-effective
that
XXXI
improvements
facilitate the
implementation of
to national building
that
their occupants,
and
fire
NISTNCSTAR
1-2B,
WTC Investigation
Preface
WTC
Location
New York
Principal
NY
City,
pending
Gaithersburg,
December
Washington.
9,
2002
MD
DC
Investigation.
Agenda
comments on
the Draft
WTC Investigation.
Media
Media
April
New York
2003
8.
City.
NY
Joint public
mter\'iews.
Gaithersburg,
MD
NCST Advisory
WTC
May
7.
New York
2003
City.
Gaithersburg,
NY
MD
Pitv
NY
May 2003
for
and progress on
session.
Media
briefing on release of
NCST
comment
Npw YnrV
comment
K^pHiQ aliU
QTiH T^iiHlip
nnpTino
iVlcUia
puUllL UllCJIIi^
c\x\
UU
Progress Report.
WTC
session.
iTntintion
iirct pciaVJIl
nprcon uala
r\^t^
llllllallUU Ui lllol
collection projects.
Gaithersburg.
Npw
MD
York Titv
NY
NCST Advisory
lilCClillK
comments on
UU
llliuilij^a Willi
puuiiL
recommendations.
June
18.
New
2004
York
vjallllCI
City.
huUI^,
NY
lyiLJ
l_
comment
system
October 19-20, 2004
Gaithersburg,
MD
at
Gaithersburg,
MD
NCST Advisory
Committee meeting on
NCST Advisory
l^UIItiltjj, a
11 1 ILi 11
Ynrl- Titv
NY
VV
status
lUUl
Committee discussion on
WTC
UIl.
Npw
ICM Ul
lliC iCoJolallCC
oCl UJ UlC.lllllllluiy
session.
lilC
sCNoIVJIl lU
Investigation.
UlUUaUlC LUliaU^C
New York
City.
NY
briefing on release of
all
Gaithersburg.
September 13-15,
Gaithersburg,
MD
NCST
MD
WTC Technical
community
2005
community
in
and recommendations
comments.
to
make
technical
(DTAP)
to (a)
WTC towers.
proposed changes
and the
are to
make
first
events.
NISTNCSTAR
1-2B,
WTC Investigation
xxxiii
Preface
final report
Team Reports on
more
WTC
is
the
WTC
Investigation
being issued as
NIST NCSTAR
As
such,
it
of Investigation publications
is
NIST
NIST
Lew, H.
the
S.,
R.
The
titles
are:
NCSTAR
Gaithersburg,
1.
MD,
and Fire
Safety
September.
NIST NCSTAR
set
World Trade Center Disaster: Final Report on the Collapse of the World Trade
NIST
Investigation of the
one of a
is
full set
Center Towers.
A companion
of the
Investigation of the
and Fire
Safety
World Trade Center Disaster: Final Report on the Collapse of World Trade Center
A. Gaithersburg,
7.
MD.
W. Bukowski, and N.
J.
and Fire
Safety Investigation of
World Trade Center Disaster: Design, Construction, and Maintenance of Structural and Life Safety
NIST NCSTAR 1-1. National Institute of Standards and Technology. Gaithersburg, MD,
Systems.
September.
and Fire
Safety
Investigation of the
NIST NCSTAR
1-1
MD,
September.
Ghosh,
S. K.,
and X. Liang. 2005. Federal Building and Fire Safety Investigation of the World
NIST
NCSTAR
MD,
Systems.
MD,
NIST
Investigation of the
NCSTAR
and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Fire Protection and Life Safety Provisions Applied
Construction of World Trade Center
Occupancy. NIST
NCSTAR
I, 2,
and
to the
Design and
September.
Razza,
J.
C, and
and Fire
1, 2,
Grill, R. A.,
and
MD,
7.
in
World
NIST NCSTAR
1-lE. National
September.
Investigation of the
xxxiv
to Structural
September.
Grill, R. A.,
MD,
September.
World Trade Center Disaster: Comparison of the 1968 and Current (2003)
NIST NCSTAR
1-2B,
WTC
New
Investigation
Preface
NIST NCSTAR
MD,
Technology. Gaithersburg,
Grill, R. A.,
and
September.
and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World
to the Fire Protection and Life Safety Provisions of the New
York City Building Code by Local Laws Adopted While World Trade Center
Use.
NCSTAR
NIST
], 2,
and
Were
in
MD,
September.
Grill,
and Fire
and 2. NIST
Technology. Gaithersburg,
MD,
NCSTAR
to
September.
of the
MD,
7.
NIST
and Fire
to
Safet}' Investigation
NCSTAR
1-1
1.
National Institute of
September.
Grill, R. A.,
Systems
Grill,
Safet}-.
Safet}'
World
Safet}' Investigation
of the World
Trade Center Disaster: Design. Installation, and Operation of Fuel System for Emergency Power
MD,
7.
NIST NCSTAR
in
September.
Sadek, F. 2005. Federal Building and Fire Safet}' Investigation of the World Trade Center Disaster:
Baseline Structural Performance and Aircraft Impact Damage Analysis of the World Trade Center
Towers.
NIST NCSTAR
1-2.
MD,
September.
W. J., and R. B. Garlock. 2005. Federal Building and Fire Safet}' Investigation of the
World Trade Center Disaster: Reference Structural Models and Baseline Performance Analysis of
the World Trade Center Towers. NIST NCSTAR 1-2A. National Institute of Standards and
Faschan,
Technology. Gaithersburg,
MD,
September.
R.
W.
and Fire
Safet}' Investigation
Center Disaster: Analysis ofAircraft Impacts into the World Trade Center Towers,
NCSTAR
Gayle, F. W., R.
J.
J.
Fields,
W.
E.
Luecke,
S.
Disaster: Mechanical
Institute
W. Banovic,
and Fire
W.
A. Siewert, and F.
E., T.
Investigation of the
Specifications.
Gaithersburg,
NIST NCSTAR
1-2B,
W.
MD,
Steel.
MD,
September.
Luecke,
T. Foecke, C.
NIST
NIST NCSTAR
1-3.
National
September.
and Fire
Safety
NIST
MD,
September.
WTC
Investigation
XXXV
Preface
Banovic,
S.
W.
S.
Luecke,
MD,
NCSTAR
September.
W., and T. Foecke. 2005. Federal Building and Fire Safety Investigation of the World
NCSTAR
and Fire
W.
E.,
J.
D. McColskey, C. N.
W.
T. A. Siewert, and F.
McCowan,
W., C. N.
September.
Banovic,
MD,
McCowan, and W.
E.
MD,
September.
Investigation of the
NCSTAR
Safet\> Investigation
NIST NCSTAR
1-4.
W.
S.
Dols, and
W.
L. Grosshandler. 2005.
Federal
MD,
September.
Kuligowski, E. D., D. D. Evans, and R. D. Peacock. 2005. Federal Building and Fire Safety
Investigation of the World Trade Center Disaster: Post-Construction Fires Prior to September II,
NIST NCSTAR
2001.
MD,
September.
Hopkins, M..
Schoenrock, and E. Budnick. 2005. Federal Building and Fire Safety Investigation
J.
of the World Trade Center Disaster: Fire Suppression Systems. NIST NCSTAR 1-4B. National
Institute of Standards and Technology. Gaithersburg, MD, September.
Keough, R.
J.,
and R. A.
Grill.
M.
J.,
and
S.
M.
MD,
and Fire
NCSTAR
September.
Strege. 2005.
MD,
NCSTAR
Pitts,
in the
Ohlemiller,
the
J.
Pitts,
1-4D. National
September.
W. M.
World
W. M., K. M.
Butler,
MD,
September.
World Trade Center Disaster: Visual Evidence. Damage Estimates, and Timeline Analysis.
NIST NCSTAR
MD,
September.
xxxvi
NIST NCSTAR
1-2B,
WTC
Investigation
Preface
S.
J.
Ohlemiller,
M. Donnelly,
Building and Fire Safety Investigation of the World Trade Center Disaster: Experiments and
J.,
G.
W.
Safety- Investigation
NIST NCSTAR
Office Workstations.
Gaithersburg,
Mulholland, A. Maranghides,
MD,
J. J.
NCSTAR
Filliben,
September.
J.
M. Repp, A.
S.
Whittaker, A.
M. Reinhom, and
P.
A. Hough. 2005.
Federal Building and Fire Safety Investigation of the World Trade Center Disaster: Reaction of
Ceiling Tile Systems to Shocks. NIST NCSTAR 1-5D. National Institute of Standards and
Technology. Gaithersburg,
MD,
September.
J.
Building and Fire Safety- Investigation of the World Trade Center Disaster: Experiments and
B., C. Bouldin,
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.
Investigation of the
Gaithersburg,
MD,
September.
Baum. 2005. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Fire Structure Interface and Thermal Response of the World Trade Center
Towers.
MD,
Gross,
J.
NIST NCSTAR
September.
L.,
and T. McAllister. 2005. Federal Building and Fire Safety Investigation of the World Trade
Center Disaster: Structural Fire Response and Probable Collapse Sequence of the World Trade Center
Towers.
NIST NCSTAR
1-6.
MD,
September.
J.,
Gaithersburg,
Gross,
J.,
F.
MD,
September.
Hervey, M. Izydorek,
J.
Mammoser, and
J.
Fire Safety Investigation of the World Trade Center Disaster: Fire Resistance Tests of Floor Truss
Systems.
MD,
NIST NCSTAR
September.
Zarghamee, M.
S., S.
M. Mudlock, W.
NIST NCSTAR
1-2B.
I.
WTC
Bolourchi, D.
Naguib, R.
Investigation
W.
Eggers, O. O. Erbay, F.
P. Ojdrovic,
A. T. Sarawit, P.
W. Kan, Y.
Barrett,
J.
Kitane, A. A. Liepins,
L. Gross,
and
xxxvn
Preface
and Fire
and Subsystem
Safet}' Investigation
Structural Analysis.
Zarghamee, M.
S.,
Y. Kitane, O. O. Erbay, T.
MD,
and
P. McAllister,
J.
NIST NCSTAR
1-6C.
September.
L. Gross. 2005.
Federal
Building and Fire Safety Investigation of the World Trade Center Disaster: Global Structural
Analysis of the Response of the World Trade Center Towers to Impact
NCSTAR
MD,
McAllister, T., R.
W. Bukowski,
R. G. Gann,
J.
September.
W. M.
Pitts, K. R. Prasad, F. Sadek. 2006. Federal Building and Fire Safety Investigation of the World
Trade Center Disaster: Structural Fire Response and Probable Collapse Sequence of World Trade
Center?. (Provisional).
Gaithersburg,
NIST NCSTAR
MD.
W. Guo,
P.
Moloney, A. Montalva,
Oh, K. Rubenacker. 2006. Federal Building and Fire Safety Investigation of the World Trade
and Fire.
(Provisional).
Gaithersburg,
NIST NCSTAR
to
Debris
Damage
and Technology.
MD.
Kim, W. 2006. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: Analysis of September 11. 2001, Seismogram Data. (Provisional).
NIST
NCSTAR
1-6G.
MD.
Nelson, K. 2006. Federal Building and Fire Safety Investigation of the World Trade Center
Disaster: The
Con Ed Substation
in
7.
(Provisional).
Averill,
J.
NIST NCSTAR
1-6H.
MD.
and
H. E. Nelson. 2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster:
MD,
NCSTAR
1-7.
National Institute of
September.
Fahy, R., and G. Proulx. 2005. Federal Building and Fire Safety Investigation of the World Trade
Center Disaster: Analysis of Published Accounts of the World Trade Center Evacuation.
NCSTAR
Zmud,
J.
and Fire
Lav^son,
J.
R.,
MD,
XXXVIll
MD,
NIST
September.
NIST NCSTAR
1-7B. National
September.
Safet}' Investigation
MD,
NCSTAR
1-8.
of the World
National Institute of
September.
NIST NCSTAR
1-2B,
WTC
Investigation
Acknowledgments
Inc.
(ARA)
(NIST) World Trade Center (WTC) hivestigation by perfonning analyses of aircraft impacts
into the
WTC towers. A collaborative relationship was developed among the NIST Investigation Team that
The high
The
NIST and
team members
level
efforts at
Investigation.
development of the
development of the
like to
Fahim
acknowledge
to the
by ARA.
Several
including Dr. William Luecke, Dr. Richard Fields, Dr. Frank Gayle, Dr.
Banovic contributed
greatly appreciated
was
NIST
investigators,
Dr. Stephen
aircraft
contributed to the data collection effort on the Boeing 767 structure. Similarly,
finite
Youngwon
in the
development of the
model could not have been completed without the support of Dr. Peter
Shin of ARA. The authors also want to thank Dr. James Filliben of NIST for his guidance
in the total
to the
to the
Several other
NIST
investigators
made
important contributions to
this effort.
McAllister provided guidance throughout the Investigation and contributed to the characterization and
presentation of the
NIST. This
visual data
was used
NISTNCSTAR
1-2B,
In addition, Dr.
WTC Investigation
William
Pitts
to observables.
xxxix
Acknowledgments
xl
left
blank.
NISTNCSTAR
1-2B,
WTC Investigation
Executive Summary
INTRODUCTION
E.1
As
the
World Trade Center (WTC) towers, under contract from NIST, Applied Research Associates.
(ARA) conducted
was
anah ze the
to
WTC towers.
ftjel
of
Inc.
the aircraft
The
and
interior core
and workstations
to
damage
fire
tower
proofing and to
the mechanical and architectural systems inside the towers. Thus, this analysis established the initial
NIST
modehng and
In\ estigation.
The AVTC
aircraft
tower and
impact analysis
is ver>'
aircraft
believed to be the highest-fidelity simulations ever performed for this impact behavior using state-of-the
art analysis
methodologies. WTierex er possible, the models were validated against observ ables or
and
WTC Investigation.
at \ arious le\ els
of complexity including:
the
component
to the towers
level,
due
to
Analysis of uncertainties using the component, subassembly, and global analyses were
most
damage
E.2
An
( 1 )
damage
important requirement for high fidelity simulation of the aircraft impact damage was the development
in the
that
WTC towers,
of the towers. These materials exhibit significant nonlinear rate-dependent deformation and
behavior that need to be represented in the constitutive relationship. The following
the constitutive
WTC
Tower
The pnmary
in
failure
of the
Steel Constitutive
constitutive
model
1-2B.
that
This model
steel structures.
NISTNCSTAR
is
failure
a brief
summary of
LS-DYNA.
model
of steel
WTC
is
Models
was used
was
A tabular effective stress versus effective strain cur\'e was used in this
Investigation
xli
Summary
Executive
The
constitutive
model parameters
for each
grade of steel were based on engineering stress-strain data obtained from the mechanical and
metallurgical analysis of the
reproduce the
up
measured response
Mesh
typical approach
was
ASTM
The
necking onset. This curve was the input used to specify the mechanical behavior
tensile test (Figure E-1).
until the simulation
failure.
tower
summarized
Strain-rate effects
the
steels are
on the
and 100
steels
ksi
based on
tower
in
rate effect
this
steels in
the point of
in the simulation
of the
necking and
WTC
Figure E-2.
of tower
model could
Test Sample
Fine
The
finite
Grip
The
The
model
were included
in the constitutive
are
compared
to the
measured high
model
for
tower
in the constitutive
50
steels
with
modeling
ksi,
Figure E-3. In the figure, the dashed lines with open symbols are the
75
ksi,
fits
Cowper and Symonds model, while the corresponding solid symbols are the experimental data.
The comparison showed that the Cowper and Symonds model was capable of reproducing the rate effects
using the
xlii
NISTNCSTAR
1-2B.
WTC
Investigation
Executive
Summary
140
120
100
80 B-/
A,
36
42
50
55
50
-s..
20
'
'
'
'
'
'
'
0.05
0.00
'
'
ksi
ksi
ksi
ksi
'
66 ksi Model
70 ksi Model
75 ksi Model
80 ksi Model
-J?- 100 ksi Model
ksi Moflel
Model
Model
Model
Model
'
'
0.15
0.10
>
'
0.20
1-
'
0.25
0.30
Plastic Strain
Figure E-2. Tabular true stress-strain constitutive model curves for the tower steels.
140
I
I
120
o-
100
%^ o
0
80
-O-
O-
-f"
^-
*&
-&
60
40
20
-3
-2-1
Log
0
Strain Rate
(s"^)
NISTNCSTAR
xliii
Executive
Summary
Concrete Constitutive
IViodels
The LS-DYN A material Type 1 6 (pseudo-tensor concrete model) was selected for modeling the concrete
floor slabs due to its ability to accurately model the damage and softening behavior of concrete associated
with low confinement conditions. The model uses two pressure-dependent yield functions and a damage-
dependent function to migrate between curves. This allows for implementation of tensile failure and
damage
scaling,
which
are
more dominant
material behaviors at
model also accounts for the sensitivity of concrete to high strain-rates. Material constitutive parameters
for the pseudo-tensor
concrete specimen
3 ksi concrete
is
shown
in
was compared
to
ksi
and 3.8
Figure E-5.
Initial
configuration
2% compression
xliv
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
Summary
Strain (%)
The
Models
open
literature.
aluminum
Complete engineering
stress-strain curves
commonly used
in the construction
in the
were
then converted into true stress and true strain and used to develop tabular curves for constitutive models.
The
tabular constitutive
model
fits
are
shown
in
Figure E-6.
No
was considered.
Nonstructural Materials Constitutive Models
The primary influence of the nonstructural components on the impact behavior is their inertial
contribution. The effects of their strength are small. As a result, relatively simple approximations of their
constituti\ e behavior were used. Typically, a bilinear elastic-plastic constitutive model was applied for
these materials to allow for efficient modeling of deformation and subsequent erosion from the
calculations as their distortions
soft materials
was important
NISTNCSTAR
1-2B,
WTC
become
large.
The
Investigation
xlv
Executive
Summary
120
'
f-
40
Model Fit
7075-T7351 Extrusion
2024-T3 Clad Sheet
2024-T351X Extrusion
7075-T651X Extrusion
Constitutive
-m-
20
OlS
0.05
0.00
0.10
0.15
0.20
Plastic Strain
Figure E-6. True stress-strain curves developed for various aircraft aluminum alloys.
E.3
One of the
analyses was to minimize the model size while keeping sufficient fidelity in the impact zone to capture
the defoiTnations and
damage
distributions.
The
limitation
was
that the
combined
WTC tower model and the aircraft so that the tower model would be about
aircraft
aircraft
and tower
about 0.8 million nodes. The approach used to meet this objective was to develop models for the
at different levels
of refinement. Components
and
in the
failure,
inertial properties.
shell
is
floors 92
As
the table
bolted connections (beam and brick elements) and the floor truss diagonals (beam elements).
xlvi
while
The
WTC
NIST NCSTAR
1-2B.
WTC
Investigation
Executive
Summary
Table E-1.
TC
Number of Nodes
Hughes-Liu Beam Elemems
Behtschko-Tsay
Shell
Elemems
1,312,092
47.952
53.488
1.156.947
1.155,815
2,805
2,498
floors:
The
spliced
shell
and a coarser
in the
failure criteria.
The
shell
beams.
1.300.537
WTC towers.
models or the
Tower Model
Summary
is
shown
in
Figure E-7.
Exterior walls: The exterior columns and spandrels were modeled using shell elements with
tw^o
mesh
densities, a
refmed density
in the
far field
density elsewhere. For the bolted connections between exterior panels in the refined
areas, brick elements
bolts.
to
model the
butt plates
1
is
shown
mesh
Figure E-8.
Truss floor: In the direct impact area, the floor model included shell elements for the
combined
floor slab
and metal decking, and for the upper and lower chords of both the
trusses.
trusses.
explicitly.
elements
Beam
segments, simplified shell element representations were used for the floor slab and
Figure E-1
were used
The
is
shown
These included the partitions and workstations, which were modeled with
in the
The
live load
shown
Figure E-10.
1-2B.
WTC
Investigation
shell
NISTNCSTAR
Figure E-9.
partitions
in
in
interior contents
is
1.
xlvii
Executive
Summary
xlviii
WTC
NISTNCSTAR 1-2B.WTC
Investigation
Executiv9
NISTNCSTAR
1-2B,
WTC
Investigation
WTC
1,
Summary
xlix
Executive
Summary
tower.
E.4
The
WTC
finite
aircraft
(3)
a three-step process:
meshing of the
The
structure.
focus of this effort was on gathering sufficient structural data and including adequate detail in the aircraft
model so
that the
implementation
from
(1)
in the
documentary
impact analyses. Structural data were collected for the Boeing 767-200ER aircraft
aircraft structural
aircraft.
size
and parameters
Airlines 175
most of the
wing
(UAL
is
significant structural
structures
is
shown
in
Figure E-12.
A summary of the
WTC
components
and
WTC 2, respectively.
in the aircraft.
shell elements.
Models
and
3 in. to
in.
were between
in.
and 2
in.
for small
wing or fuselage
(AA
1)
aircraft
and United
empennage, and
engines were developed primarily using shell elements, but contained some
typical element dimensions
1 1
and
The
rib flanges,
skin.
NISTNCSTAR
1-2B,
WTC Investigation
Executive
Summary
Executive
Summary
AA
70.000
70,000
562,000
562,000
60,672
60,672
740,000
740,000
SPH
Total
Fuel Particles
Nodes
183,500
183,500 lb
lb
ULD/Cargo Weight
12,420 lb
21,660
lb
21,580
lb
10,420
lb
Fuel Weight
66,100
lb
62,000
lb
Special emphasis
to the
objective
175
No.
damage
UAL
11
was
283,600
277,580
lb
lb
tower components. The engine model was developed priinarily with shell elements. The
to
at
many
in.
and 2
in.
However,
Brick elements were used for some of the thicker hubs and the roots of the compressor blades. The
various components of the resulting engine model are
wing
as
shown
in
Figure
E-14 based on
lii
shown
in Figure
in the
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
Summary
E.5
impacted by models for a wing section and an engine. The primary objectives of the component
modeling were
to (1)
finely
aircraft
and
(2) develop the simulation techniques required for the global analysis of the
WTC towers.
shell
beam and
shell
for
to start with
aircraft structures
in the global
and
models.
connections, and modeling of aircraft fuel. Examples of the component impact analyses included:
Impact of a segment of an
aircraft
wing with an
Impact of a segment of an
aircraft
Detailed and simplified modeling of exterior panel bolted connection under impact loading
exterior column.
Impact of an
NISTNCSTAR
1-2B,
WTC
aircraft
Investigation
liii
Executive
Summary
Impact of an empty wing segment with exterior wall panels (Figure E-16).
Impact of fuel-filled wing segment with exterior wall panels (Figure E-17).
The following
results
exterior wall and failure of impacted exterior columns. If the engine did not impact a floor
slab, the majority
of the engine core would remain intact through the exterior wall penetration
with a reduction in speed between 10 percent and 20 percent. The residual velocity and mass
of the engine
after penetration
fail
a core
column
in a
direct impact condition. Interaction with additional interior building contents prior to impact
or a misaligned impact against the core column could change this result.
A nonnal
impact of the exterior wall by an empty wing segment from approximately mid-
span of the wing produced significant damage to the exterior columns but not complete
failure.
filled
with
fuel, resulted in
damage
extensive
the external panels of the tower, including complete failure of the exterior columns.
resulting debris propagating into the building maintained the majority of
its initial
to
The
momentum
prior to impact.
Three different numerical techniques were investigated for modeling impact effects and
dispersion of fuel: (1) standard Lagrangian finite element analysis with erosion, (2)
Particle
Hydrodynamics (SPH)
analysis.
Of these
analysis,
and
(3) Arbitrary-Lagrangian-Eulerian
SPH
Smoothed
(ALE)
its
computational efficiency.
(a) Initial
configuration
(b)
Impact response at 80
ms
liv
NISTNCSTAR
1-2B,
WTC Investigation
Executive
= 0.0 s
Summary
= 0.04 s
Figure E-16. Calculated Impact of a fIne-mesh empty wing section with two exterior
panels at 442 mph.
= 0.0 s
- 0.04 s
Figure E-17. Calculated impact of a coarse mesh wing section laden with fuel modeled
using SPH particles.
E.6
as a transition
level analyses
impact analyses. With the subassembly analyses, more complex structural behavior not captured in the
associated
model
size,
to investigate different
run times, numerical stability, and impact response. The final subassembly model
used structural components from the impact zone on the north face of
in the final
WTC
NISTNCSTAR
1-2B,
WTC
Investigation
The
structural
components
and
E-18.
Iv
Executive
Summary
Ext. Panel
Numbers
Core Column Numbers
11S
WTC
aircraft
is
shown
in
ftiel-filled
wing. The
in
Figure E-20, The subassembly model was used to investigate the effect of a number of modeling
parameters on the response and damage estimates. For the engine impact simulations, these parameters
included the strength of the building nonstructural contents and the concrete slab strength. For the wing
impact simulations, the effect of the ductility of the exterior column weldment on the impact response
was
investigated.
The deceleration
contents
The
results
profile
of the impacting engine indicated that the response of the nonstructural building
was dominated by
the
deceleration and
Varying the
It
by
their strength.
change
in
appears that the mass of the concrete slab had a greater effect on the engine
damage
ductility
rather than
in the
in the exterior
damage
pattern (Figure
to
panels, indicating that the weld ductility had a negligible effect on the impact response.
Ivi
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
Summary
(a)
Time = 0.00 s
(b)
Time = 0.25 s
._
NISTNCSTAR
1-2B,
WTC
Investigation
Ivii
Executive
Iviii
Summary
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
(a)
(b) 1
damage
for the
Summary
E.7
first
to
method used
a comparison of videos
motion
in a single
from
video scaled to the length of the aircraft in the video to calculate the impact speed.
damage on
was used
impact
The
aircraft
impact conditions matching the observed exterior wall damage are shown
WTC
and
WTC
2. respectively.
The
aircraft
damage
pattern
Matching the projected impact points of the wings, fuselage, engines, and
exterior wall of each tower to the observed
in
to
was an important
constraint in the
determination of impact conditions. The final set of impact condifions from the analyses are summarized
in
Table E-3
NISTNCSTAR
1-2B,
WTC
Investigation
lix
Executive
Summary
VV
\'J
<V
WW
V-y
V-y
vj' 'C'
\'J
'\7
\~J
VjS'
\'J
K-J
'C
I-/
V'
V-y
'v''
^"7
\S'
IS)
sj
^-"V
v".'
V-.-
'
(S)
(ici) ~
Wing
tip
ir
Engine
(i>(S)
(S)
lit
(a)
a^
rSi
V'
^
V
V^
Impact damage
VVaa
fsi
^v'
^v--
-7
V^
^^V^
(i) rs^
f2) (9^
-v-'
;[irrii:
liLiOllLLL
IpCIDjjJDDiCLffI
ILIIJjjIiOOICrr-
iDDCICCIIDjjJQDEllLLi
DTMELZlZZljljUmia^
3moDCcr[::3]iinniirrr;
iLiimMKLLiiiiiimKm
::Ij3]
noo
Ecca
(b)
Impact conditions
Figure E-22.
Ix
WTC
(vertical
mi
ILLLZZ]
ODBCtCB
OjJjDOO
[CCD
3Dai]
0333
[1
0)
NIST NCSTAR
1-2B,
WTC Investigation
Executive
ffe^,
V:p^
x^/'
vS'
ii i) 1,
yTnunroTfinrnrij t
y y
\5-'
vS'
m ^1
i
i-
ii
ut
rt
fS)
v;/
1
"^
'3''
^^-
^,7/
-c-'
(9)
in
(1?,
\^
fS\
.'i^s
u a Ti. ii'innrrtiinniiiuii
Damage
(a)
I'i)
'vp^
'c^'
fg]
Summary
ri^.
sS'
;
bwriTtiiwinrnjTuTO
accairately <jetemijnel
Impact damage
'C'
vr/
'nJ-
^f-
c/
Vf-
--r-'
"C'
V>
V^^
V3/ V*/
mi
mill]
SI]]
.,-^-.--^-1-1-1-1-
^i"9''p
h:iiin:i::iJ]i:!3i]t
llBlimnil
Engine
: t ji-ii
- dDi]]iij]::3]3:ii33iirj
--ani3]]]3]]]]i3]E]]^[|[
-on]!::ini:3iJ3:!i:ji|^i
|]]]]]]]]I]]Jj]|J]j]JI]jl
(>-
|]11]]]JJ1]31J]]]J1]]1J]I
(b)
Impact conditions
Figure E-23.
NISTNCSTAR
1-2B,
WTC
WTC
Investigation
(vertical
2 impact
angle =
6, lateral
angle = 13)
conditions.
Ixi
Summary
Executive
Summary
Table E-3.
AA
Impact Speed (mph)
Vertical
3 below
10.6
Approach Angle
Angle
and
2 clockwise
from
-3 clockwise
from
lateral
approach angle
approach angle
25 2
38 2
aircraft
15
Structure North"
wing downward)
is
ueiow nonzoniai
2:
approach angle
UNCERTAINTY ANALYSES
E.8
The
(left
approach angle
2)
(heading downward)
Relative to Trajectory
Structure North
horizontal
Structure North'
Relative to Trajectory
(WTC
175
542 24
4 clockwise from
180.3
Roll
UAL
1)
(heading downward)
(Velocity vector)
a.
(WTC
443 30
Approach Angle
(Velocity vector)
Lateral
11
were
WTC towers parameters on the level of damage to the towers after impact and to detennine
damage
Aircraft impact parameters: aircraft speed, horizontal and vertical angles of incidence,
orientation,
Material properties: high strain rate material constitutive behavior and failure criteria for the
Aircraft
mass and
stiffness properties,
Tower parameters:
relative to
aircraft.
and mass
structural strength
distribution, connection
may
impact.
An
is
random
is
known
as
modeling
errors, are
methodologies on the calculated response. All of these variables did not necessarily have a significant
effect
WTC towers.
Because of the complexity of the problem and the limited number of parameters
the global analyses,
it
was necessary
to down-select a refined
first
conducted
at
that could
be varied in
all
of the
Ixii
list
most
influential parameters
NISTNCSTAR
1-2B,
WTC Investigation
Executive
Summary
parameters to a more manageable number for the global impact analyses. The sensitivity analyses
included engine impacts against core columns, wing section impacts against exterior panels, and engine-
Based on
following parameters were selected for variation in the global impact analyses:
Impact speed.
Tower
vertical
analyses due to their importance in the component and subassembly impact response.
in the
magnitude of the
this
in the
The
on the
global
vertical
truss floor
this debris.
amount of
Variations in the strength and ductility (failure strain) of materials had a similar effect on the
absorbed energy.
An
was
The
was
typically
known more
where the
value of the failure strain needed to be assigned based on the model resolution and failure criteria used.
As
a result, only the uncertainties in the material failure strain were used as a material uncertainty
The
in the
uncertainty in the weights associated with building contents (corresponding to service live loads)
found
to
was
were expected
to play a
As
were
more
and
was included
as an
E.9
The primary
was
to
WTC towers
immediately following the aircraft impacts. This assessment included the estimation of the structural
damage
that
degraded their strength and the condition and position of nonstructural contents such as
NISTNCSTAR
1-2B.
WTC
Investigation
Ixiii
Executive
Summary
in the towers.
was
method by which
the primary
the
damage
fires
to the towers
estimated.
WTC
E.9.1
The combined
aircraft
WTC
shown
is
in
Figure E-24. The base case impact analysis was performed for a 0.715 second duration following
impact of the
aircraft
nose with the north exterior wall. The analysis was performed on a computer
GHz
Intel
Xeon
was approximately
tu'o
state
components
at the
in the
monitored on average every two days. The calculations were teiTninated when the damage
reached a steady
initial
to a level that
to the
towers
to
was
percent of the
initial
case
WTC
is
shown
in
tower interior contents were removed, and the tower structures were shown as transparent so that the
tower interior could be seen.
impact response
in the
shown
E-26. The
in Figure
aircraft
impact response
fragmentation of the airframe structures. The entire aircraft fully penetrated the tower
0.25
s.
The
fuselage structures
at
approximately
were severely damaged both from the penetration through the exterior
downward
is
The
trajectory of the aircraft structures caused the airframe to collapse against the floor,
and the
subsequent debris motion was redirected inward along a more horizontal trajectory parallel to the
floor.
The downward
trajectory of the aircraft structures transferred sufficient vertical load such that the truss
floor structures
Ixiv
in the
impact zone.
NISTNCSTAR
1-2B.
WTC
Investigation
Executive
(a)
Summary
Top view
ICO
ICCBDDi
[001
8d:
d:
:::ccLDDcfi]i:
ICCtDDED ]]
:::i:t^o4^i
iccIddd]
iiitffaaDjji
IlCCODDD]]!
11.
ICllDOBIIII
zcctooDim
iccConn:: ii
:cc|Don| ]]
TCc|oDDl| 11
[CO
(b)
Figure E-24.
NISTNCSTAR
1-2B.
WTC
Investigation
[CO
Side view
WTC
Ixv
Executive
Summary
(a)
Kn
Ixvi
uni
iiiimamii
bh
unnnrra
mmm
oai
EHi
ffini
memuMi
iiimaiiiiiBebi
Em
(b)
WTC
ma
mB
iiaiiaBBiii
11
.,,ta
Figure E-25.
Time=0.00 s
'
mi
bhhi
lomii
uni
mi
laiiiiBaiii
mm3b
sa
ehi
iBEainiii
in
raiEniina
mi
Tlme=0.50 s
NISTNCSTAR
1-2B,
WTC Investigation
Executive
(a)
Summary
Time=0.00 s
(b)
Figure E-26.
The wing
structures
WTC
Time=0.50 s
aircraft fuel
cloud began to
spread out after impact but remained relatively dense until the leading edge of the fuel reached the tower
The
core.
their
aircraft fuel
and debris cloud eventually penetrated most of the distance through the core before
Load
plotted
was
momentum
for
all
momentum
is
shown
in
momentum
of the aircraft structures and contents (including fuel) and was normalized by the
first 0.1 s
interior structures.
Between
initial
and 0.25
a higher
Figure E-27. the aircraft completely penetrated the building and retained approximately 30 percent of its
initial
momentum. Beyond
NISTNCSTAR
1-2B,
WTC
little
load
s.
Investigation
Ixvii
Executive
Summary
0.0
0.1
0.2
0.3
0.4
Time
Figure E-27. Normalized aircraft
The
aircraft
0.6
0.5
0,7
(s)
momentum
for the
WTC
in
Figure E-28. At the end of the impact analysis, the aircraft was broken into thousands of debris
shows
still
in
Figure E-28(b).
A closer inspection
of the debris
shown
field
in
Figure E-28(c). Both engines had significant impact damage with one of the engines broken into two
large pieces.
Ixviii
was
still
one-third of the distance from the core to the south exterior wall.
Each had
a speed of
50 mph.
NISTNCSTAR
1-2B.
WTC Investigation
Executive
NISTNCSTAR
1-2B,
WTC
Investigation
Summary
Ixix
Executive
Summary
Exterior Wall
Damage
was
significant for
two reasons:
was
(1
which
was important
for
direct visual
evidence of the impact damage was available. Therefore, the comparison of the calculated and observed
exterior wall
partial validation
in the global
impact analyses.
Damage
base case
WTC
shown
is
in
Figure E-29. This damage was compared with a schematic of observed damage developed from
inspections of the video and photographic data collected on the tower after impact.
observed damage
in the
in
the position
of the impact damage agreed provided partial validation of the modeled geometiy of the
and shape
aircraft,
including the aircraft orientation, trajectory, and flight distortions of the wings.
observed damage. The exterior wall completely failed in the regions of the fuselage, engine, and fuelfilled
exterior
various locations
in the
was observed
wing and
local severity
at the
damage served
the
aircraft
of the tower.
Core Damage
The magnitude of damage
to the core
was
this
damage
stability
and collapse analyses. The overall model for the core structure and calculated impact damage
the core
is
shown
in
Figure E-30. The figure shows that the core had significant damage
The columns
beams were
in line
also severely
in the
to
region
damaged
A summary
of the column damage is shown graphically in Figure E-31. The damage classification levels
shown in the figure are light damage, moderate damage, heavy damage, and failed (severed). The light
damage level was defined as showing evidence of impact (low level plastic strains), but without
significant structural deformations.
in a flange), but
(e.g.,
bending
classification
was
level
lateral
had
of the
centerline.
and could can-y no residual load. The calculated damage to the core of
ixx
no
WTC
consisted of three
damaged columns.
NISTNCSTAR
1-2B,
WTC
Investigation
(a)
?
QDDCDDiDDDD:i::i::l
IDDCDDlIIDDfflDIl
(b)
Calculated
NISTNCSTAR
1-2B.
WTC
Investigation
damage
damage
(t=0.715 s)
to the
WTC
exterior wall.
Ixxi
Executive
Summary
(a)
Calculated impact
Ixxii
damage
damage
for the
base case
WTC
NIST NCSTAR
impact.
1-2B,
WTC
Investigation
Executive
(2
Severed
damage
for the
#Heavy
base case
WTC
Summary
C Moderate
1
impact.
Any damage
to the loss
of
the fires to spread from floor-to-floor. Therefore, the calculation of the floor system
damage was an
An
overall
damage
in the
is
shown
Figure E-32. The figure shows that the trusses experienced significant
impact zone. The calculated impact response produced severe damage to the truss
downward
The
truss floor
The
truss structures
NISTNCSTAR
1-2B,
WTC
Investigation
Ixxiii
Executive
Summary
Column
Column
135
109
Coliann
141
Column
151
Floor 96
Column
Floor 95
107
Column
ColiHTin
115
(a) Initial
Column
135
109
11S
(b)
Calculated
on
damage
damage
floors
(t=0.715 s)
to the
WTC
94 through 97 of WTC
is
shown
in Figure
E-33.
Fringes of damage were set such that the concrete slab failed in the regions colored red (2 percent plastic
strain
unconfined
compression). At these strain levels the concrete was severely damaged, indicating that
it
was probably
metal decking material, the elements were eroded (seen as holes ruptured in the
completely destroyed
the truss floor and
Ixxiv
E-33 shows
in the
impact zone. The zone of heavy damage to the floor slab extended beyond
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
Impact
(a)
Floor 94 slab
damage
(c)
Floor 96 slab
damage
Summary
Impact
(b)
Floor 95 slab
(d)
damage
WTC
to the
damage
Floor 97 slab
damage
determine the
to
initial
influenced the initiation and propagation of the fires in the towers. These
initial
distribution of the jet fuel in the towers, the distribution of tower contents
and
(SPH) algorithm
in
fires,
fires.
The jet
NISTNCSTAR
1-2B.
WTC
fuel
base case
WTC
smoothed
provided
particle
hydrodynamics
Investigation
the
conditions that
in a
shown
Ixxv
Executive
Summary
previously in Figure E-25 and Figure E-26, respectively. At the tennination of the global impact
analysis (0.715
the
initial
fiiel
s),
the residual
momentum,
was an ested
momentum
WTC
1).
A quantitative
in the
was nearly
percent of
aircraft debris
and
is
of the fuel and aircraft debris was deposited on floors 93 through 97, with the greatest concentration on
The calculated
floor 94.
debris cloud included 17,400 lb of debris and 6,700 lb of aircraft fuel outside of
end of the impact analysis, either rebounding from the impact face (north wall) or passing
through the tower (south wall). This amount might have been larger in the calculation since the exterior
walls were not modeled with
windows
that
would contain
towers. In addition, the impact behavior of the aircraft fuel cloud did not include the ability to stick
SPH
particles tended to
to,
or
structures.
WTC
Table E-4. Fuel and aircraft de bris distribution for the base case
Tower Location
Total Outside
WTC
WTC
WTC
WTC
WTC
WTC
WTC
WTC
Total
Aircraft Debris
6,700 lb
17.400 lb
260
8101b
Floor 92
Floor 93
6,100
Floor 94
Floor 95
lb
16,100 lb
96,000
lb
12.200 lb
28.000
lb
Floor 96
11,700
19.400
lb
Floor 97
9.500
lb
6.000 lb
Floor 98
2.200
lb
6.000 lb
Floor 99
770
WTC
lb
lb
90
lb
66.100
into the
of the
is
shown
in Figure
WTC
The fuselage
duration following
at 0.2 s after
structures
debris motion
structures
impact of the
WTC 2 global
was very
similar to that
were severely damaged both from the penetration through the exterior columns
that sliced the fuselage structures in half
The downward
trajectory of the aircraft structures caused the airframe to collapse against the floor,
Ixxvi
initial
impact and was dominated by the penetration and fragmentation of the airframe structures.
downward
lb
nose with the south exterior wall. The side view and top view of the base case
impact response
lb
196,000
lb
aircraft
lb
22.600
Weight
impact.
Aircraft Fuel
E.9.2
The
Tower
was
more
The
had transferred
to collapse in the
NISTNCSTAR
1-2B,
WTC Investigation
Executive
Summary
Figure E-34.
NISTNCSTAR
1-2B,
WTC
WTC
Investigation
EED
EU
mil
(a)
Time=0.00 s
(b)
Time=0.50 s
HE
MS
ED
EH
Ixxvii
Executive
Summary
Figure E-35.
The
aircraft
aircraft fuel
wing
structures
WTC 2
(a)
Time=0.00 s
(b)
Time=0.50 s
and fuel tank were fragmented by the impact with the tower
exterior.
The
cloud started to spread out immediately after impact, but the leading edge of the fuel
until
s after
At
0.3
had penetrated
approximately two-thirds the distance through the core and was spreading out. Beyond
subsequent motion of the aircraft fragments and fuel debris cloud was noticeably slowed. The spread of
the fuel and debris cloud
in the
open
core as a result of the open volume above the workstations in the truss floor zone.
momentum
shown
Load
0.1 s
of impact, the forward fuselage penetrated the exterior wall and impacted the interior structures.
Between
0.1 s
and 0.2
transfer rate
is
was observed
in
first
(extending outward in the wing impact regions) and a higher percentage of the aircraft mass impacted the
interior structures.
30 percent of its
Ixxviii
At 0.2
initial
s,
momentum. Beyond
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
WTC
impact,
speed on
WTC
shown
approximately 0.4
in Figure
s.
impact
to the higher
2.
Time
The
Summary
(s)
momentum
for the
WTC
WTC 2 as shown in
Figure E-37. At the end of the impact analysis, the aircraft was broken into thousands of debris
showed
that larger
fragments
still
Damage
field
occurred for specific components such as the engines and landing gear
WTC
Exterior Wall
aircraft
break up shown
in
Figure E-28.
Damage
base case
is
shown
in
Figure E-38. The exterior wall completely failed in the regions where the fuselage, engine, and fuelfilled
exterior
the structure.
Damage
wing and
extended
wing
tips,
but the
Failure of the exterior columns occurred both at the bolted connections between
various locations in
to the
at
the column, depending on the local severity of the impact load and the proximity of
The calculated and observed geometry and magnitude of impact damage were
the posifion and shape of the impact
aircraft
Agreement of both
the
1-2B,
and
flight distortions
NISTNCSTAR
damage
WTC
aircraft
in
geometry of the
of the wings.
and
Investigation
Ixxix
Executive
Summary
(a) Aircraft
(b) Aircraft
structure (time=0.00 s)
Ixxx
in
WTC
NISTNCSTAR
2 impact.
1-2B,
WTC
Investigation
(a)
Calculated
damage
(t=0.62 s)
damage
to the
(b)
NISTNCSTAR
1-2B.
WTC
Investigation
WTC
2 exterior wall.
Executive
Summary
Core Damage
In the base case
WTC 2
comer of the
core).
to the core
fuselage failed on the impact side of the tower and several of the core
some
column
at the
is
was calculated
The columns
beams were
splices located
was
in line
also severely
total
damaged
shown
in the
heavy
summary
Summary
Figure E-39.
of core
column damage
for the
base case
WTC 2
impact.
An
overall
case impact
damage
to the trusses, is
shown
in Figure
experienced significant damage in the impact zone, with the greatest damage on floor 8 1
damage
to the floor 8
truss structure
in the
This severe
extended from the exterior wall to the core. The truss floor system on floors 79 and 81 had sufficient
damage from
truss floor
damage
to
1,
Ixxxii
NISTNCSTAR
1-2B,
WTC Investigation
Executive
angle directed
WTC
to the
WTC
in
The
WTC 2.
vertical
Summary
The secondary
factor
was
that the
in
damage
WTC 2.
Floor 82
Floor 80
Floor 79
Floor 78
(a) Initial
Floor 82
Floor 78
(b)
damage
Calculated
damage
WTC
to the
The calculated
and
fuel cloud
E-34 and Figure E-35. At the end of the analysis, the residual momentum of
the jet fuel in the impact direction was less than 1 percent of the initial momentum, indicating that the fuel
cloud was nearly at rest at about 0.62 s. The bulk of the aircraft debris and fuel was arrested prior to
exiting the tower structures. However, a significant amount of aircraft debris was calculated to exit the
analysis
is
shown
in
Figure
north and east sides of the tower (sides 300 and 200 of
trajectory,
it
was estimated
that the
was calculated
comer of the
NISTNCSTAR
1-2B.
WTC
m the analysis by
it
WTC
at
to 2
ft,
which
is
Investigation
any
and
that
would
Ixxxiii
Summary
Executive
WTC 2 was relatively free of office materials the engine fragment would have been free to move
relatively unrestricted
would
likely
of speed. As a
loss
little
engine
in the simulation.
A quantitative characterization of the fuel and aircraft debris distribution is given in Table E-5.
of the fuel and
aircraft debris
aircraft debris
was deposited
on floor 80, and the largest concentration of aircraft fuel on floors 79, 81, and 82. The
tower
at the
ftiel
outside of the
end of the impact analysis, either rebounding from the impact face or passing through the
tower.
The bulk
in floors
the tower
windows
that
is
is realistic
could contain the fuel cloud and small debris inside the tower.
In addition, treatment of the aircraft fuel cloud did not include the ability to stick to, or wet, interior
aircraft fuel
SPH
particles
would tend
to
Table E-5. Fuel and aircraft de bris distribution for the base case
Tower Location
Total Outside
WTC 2
WTC 2
WTC 2
WTC 2
WTC 2
WTC 2
WTC 2
WTC 2
10,600
Floor 77
2 impact.
55,800 lb
lb
400
1,300 lb
lb
Floor 78
6,200
Floor 79
11,400 lb
16,200 lb
Floor 80
6,000
83,800 lb
Floor 81
14,400
Floor 82
10,600 lb
Floor 83
1,500
Floor 84
200
Total Weight
4,800 lb
lb
lb
27,300
lb
lb
3,600 lb
4,300
lb
500
lb
62,000
E.9.3
WTC
Aircraft Debris
Aircraft Fuel
Tower
structures.
lb
lb
197,600
lb
lb
Analyses
Additional impact analyses were performed for each tower to provide a range of damage estimates due to
the uncertainties in the calculated impact response.
Two
less severe
impact scenario. The variations in impact conditions for the different severity global impact analyses
WTC
472 mph
in the less
in the
severe and
less severe
WTC
in
mph and
lower bounds obtained from the analysis of aircraft impact conditions. The vertical trajectory of the
aircraft
was
also varied
Ixxxiv
lateral trajectory
more severe
in a greater
was
the impact
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
and normal
effect
to the
nonh
WTC
face of
1.
A small variation
in the lateral
Impact \
Flight
Parameters
Aircraft
Parameters
To>ver
Parameters
a.
calculated
analyses
is
pitch
Trajectory
yaw
mpn
4 /z
/
was
mpn
414 mpn
^ f.
J.O
.0
0.0
0.0
pitch
8.6
5.6
11.6
Orientation
yaw
0.0
0.0
0.0
^^ eight
00 percent
105 percent
95 percent
Failure Strain
00 percent
125 percent
75 percent
Failure Strain
100 percent
80 percent
120 percent
Live Load N\
25 percent
20 percent
25 percent
eight''
in Figure
E^l,
of WTC
trend
Less Severe
0.0
may have
little
More Severe
u.o
in
Orientation
damage
shown
44j
Trajectory
The
WTC
Base case
elocity
Summary
in
The
calculated
similar
severed columns in one case and were removed at the connections in another. In general, the
for a larger
less severe
in the
more
severe impact. The increase in the opening with reduced severity impacts can be explained by the
increased tower panel material toughness, resulting in a transition from severing of columns to the failure
Toward
the
wing
tips,
where the columns and spandrels were not completely severed, the more severe
impact damage analysis calculated higher damage to the exterior wall panels. These columns had the
largest
plastic strains
damage
The
calculated
damage
The
Figure
to the
figure
WTC
shows
damage to the
of damage near the wing tips.
in the
As would be
exterior wall,
is
and the
shown
in
A strong
correlation in core
was
a clear correlation in
severity
were selected
to
was expected.
produce an increase
in core
damage.
A higher
impact speed and a shallower impact angle were selected to direct more energy toward the core. The
aircraft
NISTNCSTAR
1-2B,
WTC
Investigation
Ixxxv
Executive
Summary
reduced, and the tower materials were given a lower toughness. All of these variations contributed
severity.
(a)
(b)
(c)
(d)
WTC
damage
for the
the
downward impact
trajectory angle
Ixxxvi
was
amount of damage
impact analyses.
to the truss floor
However,
WTC
in the
more energy
NISTNCSTAR
in
many
analysis.
effects
1-2B.
and
of varying
WTC Investigation
Executive
#Severed
(a)
(b)
WTC
Moderate
damage
Heavy
UModerate
1-2B.
Heavy
Severed
NISTNCSTAR
iModeiaie
* Severed
(c)
Heavy
Summary
Investigation
damage
for the
WTC
impact analyses.
Ixxxvii
Summary
Executive
WTC
2 Global Impact
Analyses
less severe
WTC 2
corresponding parameters in the base case analysis in Table E-7. The impact speeds were 521
570 mph
in the less
and lower bounds obtained from the analysis of aircraft impact conditions. The
aircraft
was
mph and
severe and more severe impact scenarios, respectively. These speeds were the upper
also varied
from
vertical trajectory
of the
Aircraft
Parameters
Tower
Parameters
a.
Trajectory
546
WTC
floor
Damage
570 mph
521
mph
pitch
5.0
8.0
yaw
13.0
13.0
13.0
pitch
5.0
4.0
7.0
Orientation
yaw
10.0
10.0
10.0
Weight
00 percent
105 percent
95 percent
Failure Strain
00 percent
5 percent
75 percent
Contents Strength
100 percent
80 percent
100 percent
Failure Strain
100 percent
90 percent
120 percent
25 percent
20 percent
25 percent
analyses.
damage
mph
Less Severe
Orientation
damage
severity impacts.
More Severe
6.0
Calculated
the
Trajectory
Base case
Impact Velocity
Flight
Parameters
WTC
Only
Damage
slight
if
and
changes
to the exterior
truss floor
of
WTC 2
showed
did not vary significantly due to the competing effect of the parameters varied.
WTC 2
core columns
is
shown
in
Figure
perfonned. The figure shows that core damage was concentrated in the core region closest to the impact
point.
There
is
Ixxxviii
in the
in the
more severe
in the less
A total of three
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
508
501
01008
1001
Heavy
Severed
(a)
Summary
CModeiate
508
501
1008
1001
Severed
(b)
Heavy
Moderate
damage
508
1008
Severed
(c)
1-2B.
WTC
Moderate
NISTNCSTAR
Heavy
Investigation
damage
for the
WTC
2 impact analyses.
Ixxxix
Summary
Executive
E.9.4
The observables
available to help validate the global impact analyses included the following:
Damage
and floors
in the
WTC
for
who were
stairwells).
Not
all
impact conditions, the imperfect knowledge of the interior tower contents, the chaotic behavior of the
aircraft
break up and subsequent debris motion, and the limitations of the models. In general, however,
were performed by
(WAI). The
WAI
MIT
damage
FLEX
finite
to that
of LS-DYNA. In the
model. The
WAI
was included
aircraft fuel
damage
aircraft
is
in the
to the
finite
damage
calculated by
model by increasing
WAI
WTC
WTC towers
and
mass of the
the
structures
Damage
predictions from
profiles in the
WAI
less influence
differences in the
damage
models. The
WAI
modeled
aircraft
impact
in this
damage
included complete failure of the exterior columns over the entire length of
profiles in the
One major
to,
to the
aircraft
their
two
columns adjacent
to the
failure.
likely resulted
studies
was
from
a variety
in the fidelity
of
of the aircraft
military transport
compared
Figure E-44. The figure also shows a schematic of the damage observed in
in
The damage
is
the wings and top of the vertical stabilizer, and displayed local rupture of the
between the
WAI
and tower model differences (material properties, geometry, joint modeling, number of
elements,
xc
The
to both towers.
in the
etc.)
impact damage
Inc.
wing box.
The
These studies
staff at
FLEX
well.
E.9.5
Two
etc.).
NISTNCSTAR
1-2B,
WTC
Investigation
Executive
(a)
Summary
immimjOzciiiLCLCKLij:''^^
tlJlSnOOCDlCriCHECi E
mJlOOlQOEllJtHlLdq
IIJQIQD
iDaoiocsi
IQOQILLa.ilQ:
IDCESCCl
lODIHQLi
lOQOiDDiiBEloirE Joninnpi^f
(b)
...1
(c)
WAI
calculated
S'i S S
eft
damage
damage
to the
WTC
exterior wall.
Table E-8 compares the estimated core column damage from the various studies. For
WTC
MIT
estimated 4-12 core columns were failed (4 columns failed over a three-story length or 12 columns failed
a single floor
NISTNCSTAR
1-2B,
the
FLEX
WTC
fall
failure.
The
first
floors.
WAI
damaged
impact analysis. The second estimate of 20 failed columns was the number
Investigation
xci
Summary
Executive
used
and
The NIST base case impact damage of 3 severed and 4 heavily damaged
of
MIT
falls
damaged
3 heavily
falls in the
and
WAI
middle of the
MIT
WTC Impact
WTC
Investigation
MIT
7-20 Failed
4-12 Failed
Impact Analysis
WAI
&
23 failed
Impact Analysis
WAI
damaged
significantly
Plus 5
Damage
Damaged
Plus 10
20 Failed
damaged
damaged
5 Failed
Collapse Analysis
Damaged
Plus 3 Heavily
Less Severe
Damaged
Plus
Damaged
0 Severed
Heavily
Damaged
3 Severed
Severed
Plus 2 Heavily
Impact Analysis
Severed
Plus 4 Heavily
6 Severed
Impact Analysis
NIST
Severed
Plus 4 Heavily
Impact Analysis
Damaged
Plus 2 Heavily
found
in the
Damaged
estimated 7-20 columns failed (from 7 columns failed over a three-story length to 20 columns failed over
a single floor length).
WAI
columns
damaged
in their
FLEX
impact dainage of 5 severed and 4 heavily dainaged, as well as the more severe estimate of 10 severed
and
heavily
damaged
fell in
for the
damaged than
by MIT. The
the
MIT
and
less severe
WAI
impact scenario
studies.
number of failed core columns agrees remarkably well with the NIST
The WAI impact analysis predicted higher core column
Another reason
damage
columns from
WTC 2
that predicted
by
WAI
damage
prediction
may be
failure
attributed to
prediction in the
WAI
their
For
studies.
E.10
The
to
finite
element impact
The
analysis of the overall aircraft impact forces: This analysis indicated that the
momentum
transfer was more important than the strength of the impacting aircraft in determining the
xcii
NISTNCSTAR
1-2B,
WTC Investigation
Executive
The
effects
of the energy in the rotating engine components: This analysis indicated that the
potential for the rotational kinetic energy of the engine to significantly change the
static
is
significant effect
The
magnitude
small.
Summary
showed
(i.e.,
analysis
fail
core column. Application of this analysis indicated that the dispersed fuel cloud alone
not be sufficient to
NISTNCSTAR
1-2B.
WTC
fail
would
core colurnns.
Investigation
XClll
Executive
Summary
xciv
left
blank.
NISTNCSTAR
1-2B,
WTC
Investigation
Chapter
Introduction
(WTC)
effort
was
and
interior core
during the impact; and (3) estimates of debris damage to the interior tower contents,
The
towers to provide the following: (1) estimates of probable damage to structural systems due to
initial
initiation analysis.
WTC aircraft impact analysis is very complex, with large scale fracture and fragmentation of both
tower and
aircraft
believed to be the highest-fidelity simulations ever performed for this type of impact behavior using stateof-the art analysis methodologies.
subassembly
aircraft impact.
level,
and
WTC Investigation.
at
damage
to the towers
due
to
analyses were conducted to assess the effects of variabihty associated with various input parameters and
identify the
most
Based on the
results
damage
As
part of the
tower and
aircraft
and
were
identified
were
WTC 2.
WTC
aircraft
test results
of the
in the exterior
walls and core columns of the towers, weldment metal, bolts, reinforced concrete, and aircraft materials
strain-rate effects
Chapter
and
in
2.
aircraft global
the original
in
structural performance.
included the primary structural components of the towers in the impact zone, including exterior walls,
floor systems, core columns,
and connections.
aircraft.
NISTNCSTAR
1-2B,
WTC
in the
Investigation
aircraft structural
fuselage,
Chapter
The
WTC towers and Boeing 767 aircraft are complex structural systems.
all
In the global
model
both the aircraft and towers. This approach, however, results in very large models. The component and
to
while maintaining fidelity in the analysis. Therefore, a series of component impact and subassembly
analyses were performed.
(1)
and
(2)
develop the simulation techniques required for the global analysis of the
to
aircraft
aircraft
WTC towers, including variations in mesh density and numerical tools for modeling fluid-structure
interaction for fuel impact
finely
coarsely
impacting a
strip
impacted
and
WTC
An example
initial
1,
to the core
WTC towers.
and
trajectory,
aircraft nose.
In addition, a large
an aircraft engine
is
6, respectively.
aircraft orientation
two videos
utilized the
fuel.
An
by a segment of an
is
with
to relatively
shell
and progress
aircraft structures
start
to
These include
This task
flight
1,
impacted
which
WTC 2.
accuracy possible, the speed, horizontal and vertical angles of incidence, and
to estimate,
roll
The
initial
performed
the
at
etc.
levels, as reported in
Chapter
8.
analyses were to assess the effect of uncertainties on the level of damage to the towers after impact and to
determine the most influential parameters that affect the damage estimates.
Chapter 9 presents the results of the global analyses of aircraft impact into
global tower and aircraft models.
that
The analysis
WTC
and
damage
degraded their strength and the condition and position of nonstructural contents such as partitions,
workstations, aircraft fuel, and other debris that influenced the behavior of the subsequent fires in the
towers.
parameters.
They
also provided
influential parameters.
more and
less severe
less severe
variations of the
a range of
most
damage
estimates of the towers due to aircraft impact. Comparisons between the simulation results and
observables obtained from video and photographic evidence, as well as eyewitness interviews, were also
conducted. The chapter also provides a comparison with prior estimates of the damage obtained from
other studies.
NISTNCSTAR
1-2B,
WTC
Investigation
Introduction
As
reported in Chapter 10, these analyses included: (1) the analysis of the overall aircraft impact forces
relative
evaluation of the potential effects of the energy in the rotating engine components on the calculated
damage and
NISTNCSTAR
1-2B,
WTC
and
static
Investigation
Chapter
left
blank.
NIST NCSTAR
1-2B,
WTC
Investigation
Chapter 2
INTRODUCTION
2.1
An
important requirement for high fideUty simulation of the aircraft impact damage
constitutive
models
aircraft structure
aircraft.
is
the development of
(WTC)
towers and
for the impact analyses included the several grades of steel used in the columns, spandrels,
trusses
the
Boeing 767
and floor
WTC towers, the concrete floor slabs, and the aluminum airframe structure of
All of these materials played a significant role in the aircraft impact
aircraft.
damage
analyses. These materials exhibit significant nonlinear rate-dependent deformation and failure behavior
o\ er the range of strain rates expected in the impact scenario.
aircraft
in the
ftimiture
steel frame,
these materials
is
relatively small
constitutive behavior
The
primary structural
is
compared
insulation.
The
strength of
adequate.
(LS-DYNA
Version 971).
LS-DYNA
the
LS-DYNA
finite
element code
models and was capable of accurately reproducing the important material behaviors required
analysis. Material
fuel,
WTC towers
in
LS-DYNA
in this
material data from tests conducted by the National Institute of Standards and Technology (NIST) or
available in the public domain.
2.2
in
LS-DYNA
that
and
failure.
The primary
constitutive
model
model was
The
constitutive
in this
failure
of the
steel structures.
sufficient to
of steel used
in the construction
of the
WTC towers based on engineering stress-strain data provided by the NIST mechanical and metallurgical
analysis of structural steel (see
NIST
NCSTAR
1-3).'
The approach
to
This reference
is
to
this Investigation.
in the
Preface
to this report.
NIST NCSTAR
1-2B,
WTC
Investigation
Chapter 2
is
valid
is
up
to the onset
Extrapolate the true stress-strain curve beyond the point of necking onset.
Perform
iterative finite
The primary
until the
criterion
stress-strain
stress.
The
fine
test
true stress-strain
was
mesh
measured
maximum
model
Perfonn
a final finite
mesh
to the
using a coarser
test
level.
resolution
analyses).
same engineering
strain
that the
mesh
component
to failure
was
accurately captured in the constitutive model. In addition, the simulation of the material testing provided
a validation that the constitutive model parameters were defined accurately and that the model could
The
tensile tests
test conditions.
finite
column
materials.
ASTM
test
standard
test
(ASTM Designation A
370
in.
The
all
on
WTC
for testing
of the
of the
models used
and medium
fine
tests
is
described
in the
in.
The
following
sections.
Test Sample
Grip
Figure 2-1
Example
finite
ASTM
NISTNCSTAR
1-2B,
WTC
Investigation
most
from
370
2.2.1
In
ASTM
this plot
at the
strain
can be generated
displacement by the original length of the specimen. If the changes in area and length are small during
the
test,
especially
test,
is
stress
stress
when a general three-dimensional state of strain exists. As a result, the engineering stress
some materials approach failure, implying a weakening of the material. In reality, the stress
decreases as
to
are based
rise to
is
stress
and
strain
area
(i.e.,
necking).
These give
terms such as "Green"' and "Almansi" strain tensors. These are important in writing a computer
measure
The
strain
true stress
is
problems.
An
alternate approach
is
and
based on the load divided by the actual cross sectional area of the specimen and
equal to the engineering stress multiplied by a term to correct for the change in cross section.
f^r=^..,a + ^)
where
cXj-
and
NISTNCSTAR
cr
1-2B.
WTC
and engineering
Investigation
stresses, respectively,
(1)
and e
is
the engineering
strain.
Chapter 2
= ln( ) =
ln(l
j-
is
defined as
+ e)
(2)
This definition comes about from defining the incremental true or "naturaf strain as the cun"ent "change
in length" divided
This
is
with the definition of engineering strain that references the change in length, A/
in contrast
or
A/
(4)
After the onset of localization, the determination of the true strain in the necked region becomes
Development
2.2.2
the
to illustrate the
methodology
NCSTAR
NIST
Investigation
in the
tests
Tl and T2
in the figure).
roll
1-3)
Four
steel.
model development of
tests
tests
This particular grade of steel had a larger anisotropy than seen in most of
for constitutive
(NIST
more
neck geometry.
In this section,
local
specimens aligned with the rolling direction. The largest effects of the anisotropy were observed
in the
The
first
The
typical
approach was
in Section 2.2.1.
shown
in
from
test
LI was used
curve
Figure 2-3. This true stress-strain curve was then approximated by a piecewise linear curve in
tabular form as
shown
in
Figure 2-4. This tabular curve was the input used to specify the mechanical
The
final step
tensile test as
was
shown
in Figure 2-5.
model
to simulate the
adjusted until the simulation matched the measured engineering stress-strain response, including necking
and
of the stress-strain
cuwe beyond
the
maximum
engineering stress).
of the calculated and measured tensile behavior for the 75 ksi perimeter
this
shown
in Figure 2-6. In
example, the constitutive model was developed as an average between the longitudinal and transverse
properties. Results
stress-strain
steel is
A comparison
NIST
in the longitudinal
NIST NCSTAR
to the onset
1-2B,
of
WTC Investigation
necking. The difference between the longitudinal and transverse properties was in the ductility, where the
in the constitutive
modeling was
to use
in the exterior
column
steels is
about 1.22:1.
properties and ignore the orthotropic characteristics of the material in the impact analyses.
The
effects
produced by the variation of ductility between the longitudinal and transverse directions were assessed as
part of the uncertainty analyses.
120
"T
100
80
u)
60
I/)
Tes: Daia.
75
75
ksi
Tes
LI
ksi Tes! L2
75 ksiTes) Tl
76 ks) TeEi T2
40 r
True Slreas:
75kSiTes!Ll
20 -
0.00
0.05
L.
0.15
0.10
0.25
0.20
Strain
Figure 2-3. Test data and true stress-strain conversion for the 75 ksi steel.
120
100
(A
I/I
V)
90
75
ksi
Fit;
Model
]!
80
1
I
]j
,
70
"
1
0.00
0.05
0.10
0.15
0.20
rr
0.25
0.30
Plastic Strain
Figure 2-4. Tabular true stress-strain constitutive model curve for the 75 ksi steel.
NISTNCSTAR
1-2B,
WTC
Investigation
Chapter 2
Effective
Plastic Strain
Fringe Levels
A Ann _ n4
o.OuOe-ul
7.2uOe-ui _
6.400e-01
5.600e-01 _
4.800e-01 _
4.000e-01 _
3.200e-01 _
2.400e-01 _
1
.GOOe-01
8.0006-02
O.OOOe+00
Figure 2-5. Calculated tensile test response with necking for the 75 ksi steel.
120
-T
100 -
80
60
V)
0)
Test Data:
w
40
FEA
75
75
75
75
ksi Test LI
ksi
Test L2
ksi
TestTI
T2
ksi Test
Stresses:
75 ksi Fine
75
kSi
Mesh
Coarse
Wles^:
20
0.00
0.05
0.10
0.15
0.20
1_
0.25
Strain
The
of the
WTC towers.
10
NCSTAR
As
(suppliers)
of steel used
in
NIST NCSTAR
1-2B,
WTC
Investigation
test data
were used as
to failure.
is
shown
modeling
up
A comparison of this synthetic curve with the constitutive model behavior for the 75 ksi steel
in
Figure 2-7. There were observable differences in the curves, such as the representation of a
However,
produce a large
110
'
'[
MT
sufficiently large to
't
were not
the differences
Ml
100
80
yQ
Ft
0.00
0.02
0.06
0.04
0.08
0.10
0.12
True Strain
Figure 2-7. Comparison of the constitutive model and synthetic steel behaviors
for the 75 ksi steel.
2.2.3
which
test data
Summary
summary of the
steels are
to
of the
WTC tower
summarized
all
in
WTC tower
with the synthetic stress-strain curves, developed during the mechanical and metallurgical analysis of
structural steel, for the various exterior
column
steels is
shown
in
core column steels in Figure 2-11. In general, the true stress-strain curves developed from the
in
described in Chapter
NISTNCSTAR
to within
1-2B,
test data
approximately 10 percent.
in the uncertainty
analyses
8.
WTC
Investigation
11
Chapter 2
t
20
t_i
0.00
~
I
0.10
0.05
0.15
'
0.20
0.25
0.30
Plastic Strain
120
'35
100
O
55
80
60
0.00
0.05
0.10
0.15
0.20
Plastic Strain
Figure 2-9. Comparison of the constitutive model and synthetic steel behaviors
for the higher strength perimeter steels.
12
NISTNCSTAR
1-2B,
WTC
Investigation
0.00
0.05
0.15
0.10
0.20
Plastic Strain
Figure 2-10. Comparison of the constitutive model and synthetic steel behaviors
for the lower strength perimeter steels.
100 r
0.00
0.05
0.10
0.15
0.20
Plastic Strain
Figure 2-11. Comparison of the constitutive model and synthetic steel behaviors
for the core steels.
NISTNCSTAR
1-2B,
WTC
Investigation
13
Chapter 2
2.2.4
Elevated strain rates can influence the strength and ductility of structural materials. For the materials and
strain rates
compared
WTC tower impact analyses, these strain rate effects are expected to be somewhat small
of the
to the effects
of the baseline
(static) strength
Symonds
rate effect
and
model
failure
for the
is
governed by the
equation:
^.v='^,vti+(^)"n
where
cr,,
and
cr,.
(5)
and
zero, respectively.
parameters.
A series of high-rate characterization tests was perfonned on tower steels as part of the NIST mechanical
and metallurgical analysis of structural
0.001
s"'),
steel.
The
p = 6.7824
s"'
p,
at strain rates
were then
steels
exterior columns.
mesh
in the core
high rate
lines with
50
ksi,
75
fits
data.
in
in ksi:
summarized
in
Table 2-1
The 36
ksi,
s"
Gy,,)^
specific values used for each of the different tower steels are
and 42 ksi
fit
below
steels,
based on Equation
(5), are
ksi
in the
medium mesh
compared
to the
measured
and 100 ksi tower steels in Figure 2-12. In the figure, the dashed
using the
solid
capable of reproducing the rate effects for the range of data available.
Test results indicated that the influence of strain rate on the ductility of the tower steels did not follow a
consistent trend. Several grades of steel had an increased ductility at high rates (more
common
for
low
some had a reduced ductility at high rates (more common for high strength steels), and
some showed no significant effect of strain rate on ductility. The approach used in the constitutive
modeling is to ignore the changes in ductility produced by elevated strain rates. The effects of the
strength steels),
variation of ductility over the expected range of strain rates were assessed as part of the uncertainty
analyses.
14
NISTNCSTAR
1-2B,
WTC Investigation
11
140
'
1
120
100
^
80
<>.-0
.t*
^
o-
i*,..--*
4i
60
40
50 ksi Material Tests
100 ksi Material Tests
75 ksi Material Tests
A.
20
-3-2-10123
Log
Strain Rate
(s"^)
Table 2-1
Material \ idd
Young's
Specification
Modulus
Poisson's Ratio
Coefficient (C)
2.3
Strain-Rate
Exponent
(p)
Brick Element
Failure Strain
36
ksi
29.7 Msi
0.288
7.900e-H01
6.782e+00
0.32
42
ksi
29.7 Msi
0.288
1.360e-H05
6.782e+00
0.32
50 ksi
29.7 Msi
0.288
4.220e+06
6.782e+00
0.50
55 ksi
29.7 Msi
0.288
6.700e+06
6.782e-H00
0.64
60
ksi
29.7 Msi
0.288
3.950e+07
6.782e+00
0.56
65 ksi
29.7 Msi
0.288
1.270e-H08
6.782e+00
0.51
70 ksi
29.7 Msi
0.288
.270e-^08
6.782e-^00
0.62
75 ksi
29.7 Msi
0.288
6.710e+08
6.782e+00
0.56
80 ksi
29.7 Msi
0.288
2.440e+09
6.782e+00
0.49
100 ksi
29.7 Msi
0.288
3.430e+09
6.782e+00
0.53
FAILURE MODELS
A challenge for calculating the aircraft impact response and damage to the WTC towers was the wide
range of failure mechanisms that occurred in both the aircraft and tower structures. These failures
resuhed from severe loads and large scale deformations of the materials, as well as from exceeding the
strength of bolted, welded, and riveted connections. At the connection locations,
influenced by variations in material cross sectional geometry and material properties that produced stress
and
strain concentrations.
NISTNCSTAR
1-2B,
WTC
Investigation
15
Chapter 2
to
model
the impact
damage and
failure
LS-DYNA
finite
that occurred
at the
and
stress state.
When the
an element was exceeded, the element was eroded (deleted) from the calculation. The erosion of elements
allowed for the propagation of failure through a structure.
Calculation of the failure of structural components
impact analyses.
initiation.
Damage modeling
is
on a fine
further complicated
was
scale.
mesh can be
applied, and
at
a local
WTC towers,
approximately 2 million
elements. At this resolution, the gradients around a fracture cannot be accurately resolved, and the
damage
criteria required
to the aircraft
described.
Mesh Refinement
2.3.1
Effects
One of the significant modeling issues for the analysis of aircraft impact damage and failure is the effect
of mesh refinement. The global impact analyses of the aircraft impacts into the WTC towers were very
large analyses of
complex
structures
from the detailed component analyses. As the mesh refinement was reduced,
that the
extent of impact
to
it
be reduced significantly
was important
to ensure
captured.
A preliminary example of the effects of mesh refinement on the response was introduced earlier in the
WTC tower steels. Figure 2-1 showed both fine mesh and a coarser
mesh (referred to as medium mesh) versions of the ASTM 370 rectangular tensile specimen for plate
analysis of the material tests
material.
on the
A comparison of the
shown
in Figure
meshes
same
level of
specimen displacement.
16
NISTNCSTAR
1-2B,
WTC Investigation
(b)
The
effects
of further reductions
ASTM
in
370
in
example used
stress
and a single
in
in this section.
entire
To most
sufficient independent
behavior of the fme-mesh model, a failure strain of 0.18 was needed to model failure for
element. Note that this closely matches the engineering failure strain of 0.20 for this material (see
Figure 2-6).
is
stress-strain
shown
in Figure
mesh
was
maximum
strain.
The corresponding
medium mesh
resolufion,
(fine,
matching the
(shell
element model).
NISTNCSTAR
1-2B,
WTC
Investigation
17
Chapter 2
mesh
140
120
100
80
in
V)
60
40
20
0
0.00
0.05
0.10
0.15
0.20
0.25
0.30
Strain
Figure 2-15.
Mesh refinement
The medium mesh resolution used in the analysis of the tensile test described above corresponded to the
mesh resolution applied in the exterior column component impact analyses described in Section 5.2.1.
Therefore, the critical failure strain of 0.56 was carried forward to the detailed component analyses. The
coarser shell element subassembly and global impact models were typically developed with shell
in
is
described below in
Section 2.3.2.
18
NISTNCSTAR
1-2B,
WTC
Investigation
2.3.2
WTC towers immediately after impact, and inspection of the recovered exterior wall
Photographs of the
shown
weld zone
that failure
damage.
weld zone
was
very'
An example
As
were
exterior
web
is
shown
exterior
had
fractures
a
that
a result, the
(HAZ) was
The
failure
HAZ
little
effect
columns and
to develop
was
models and
required.
Source: NIST.
weld and
The
performed
at
The
specific
microhardness characterization locations are shown in Figure 2-17. The corresponding hardness
measurements across the base, HAZ, and weld materials are shown
NISTNCSTAR
1-2B,
WTC
Investigation
in Figure
2-18.
19
Chapter 2
Source: NIST.
300
<
250
m
500 g
200
Base
Weld
HAZ
150
0.6
0.4
0.2
Distance
(in)
200 g_
in Figure
HAZ
materials.
The flow
stresses
of the
increased from that of the base material by 12 percent and 18 percent, respectively. These shifts
con'espond to the relative magnitude of the average measured hardness in each material region.
20
NISTNCSTAR
1-2B,
WTC Investigation
The remaining
weldment
model was
failure
that could
test configuration
shown
in Figure 2-19.
aircraft
to
perfonn three-dimensional
The drop
test
in.
is
12
in.
wide and 2
in.
wide and
5 in. across.
exterior
column between
spandrels.
Figure 2-19. Drop test model for column weld fracture analysis.
The two
different
mesh
for the
similar to the
weld zone
medium
shown
The
in Figure 2-20.
model shown
in
Figure 2-1 3fb). This quarter-symmetry brick element column model contained 63,680 linear solid
elements. The coarse shell element model had approximately 4
1
in.
wide elements
in the
in.
shell
contained 144 linear shell elements. Obviously, the coarse shell-element model was not able to capture
weld zone
The drop
was simulated
test
first
to the
same extent
using both the brick and shell element models. The results of the
web
bound
in the
of the
in
Figure 2-21 (each black dot in the figure corresponds to a specific element plastic strain magnitude). The
figure
shows
model calculated
corresponding shell element model could calculate only a single strain value for the
as
shown by
NISTNCSTAR
1-2B,
WTC
peak
strains near
The
in.
Figure 2-21
Investigation
21
Chapter 2
An
average
critical strain
1
in.
wide single
example (55
element models
is
shell
ksi steel).
shown
in the
element weld zone. This weld zone failure strain was 8 percent in
in Figure
length of the weld failures were in good agreement. These analyses provided an upper
the engineering fracture criteria in the
effects
of a more
failure
for
and the
brittle fracture
that the
Medium mesh
little
Web
Flange
(a)
bound value
and
resolution
(b)
0.15
-]
0.10
e
a>
E.
113
0.05
O.OW
0.0
0.2
Ll
0.4
0.6
0.8
Distance From
1.0
1.2
L-
1.4
1.6
Web (in)
Figure 2-21. Calculated energy balance for the 2-D weldment models.
22
NISTNCSTAR
1-2B,
WTC Investigation
Medium
(a)
(b)
Coarse resolution
shell
element model
2.3.3
The priman'
exterior
0.875
columns of the
in.
WTC towers.
diameter grade
use
it
to
A325
at the
Within the impact zone, the connections were typically made using
steel bolts.
Initially, there
to
were no
test
develop a corresponding beam element bolt model for the majority of the impact analyses.
description of these bolt analyses are given in Section 5.2.3, along with the
of approximately 68
NIST NCSTAR
1-2B,
kip.
any significant
at a
metallurgical analysis
is
given in
strain rate
WTC Investigation
component analyses.
at failure
was approximately
0.1 8 in.
The
dependency.
23
Chapter 2
80
1
1
^^^^^
60
.s^
40
[
!
1
1
1
1
-T-T-T-rT
20
Bolt
Bo
Load
Load
Load
Soft 3
Bolt 4 Lload
Average Load
t
"
'
1
0.00
i.
0.15
0.10
0,05
0.20
0.25
Displacement (inches)
the
in the strength
may be
of the
bolt,
bolt model,
but
at a
boh
test data.
it
test data
were used
Figure 2-24. The bolt model shows a bilinear elastic-plastic behavior that
and yields
to the
The
beam model
beam element
was compared
test
data
is stiffer in
is
shown
of
in
higher stress level than the data. The inability of the simplified model to capture stress
model and
may
80
60
40
ns
20
Measured Bolt Behavior
Preliminary Boll Model
Q I
0,00
0 05
0 10
0.15
0.20
1_
0.25
Displacement (inches)
24
NISTNCSTAR
1-2B,
WTC
Investigation
2.4
in
LS-DYNA. Each
models
in
LS-DYNA
(soil
An
is its lateral
in the
WTC towers were not highly confined, therefore, a material model suitable for this loading condition was
needed.
is
Damage caused by
important.
low confinement regime. Inclusion of this damage growth provides a more accurate
16
(pseudo-tensor concrete model) was selected for modeling the concrete floor slabs. This model also
As implemented
in
LS-DYNA,
the pseudo-tensor
rates.
in
(1) a
simple tabular pressure-dependent yield surface, and (2) a complex model with two pressure-dependent
yield functions and a damage-dependent function to migrate between curves.
The
first
option
is
well
modeling standard geologic material behaviors such as a Mohr-Coulomb yield surface with a
suited for
Tresca limit and has been used successfully for the analysis of ground shock and soil-structure
interactions under high confinement.
of tensile
failure
The second
at
low confinement.
form:
in the
a^(p) = a,+
a^
maximum
(6)
+ a^p
in
is
By
undamaged and fully damaged concrete and an appropriate tabular interpolation between
damage behavior under low confinement can be captured. For a detailed description of the
NISTNCSTAR
is
1-2B,
referred to the
WTC
LS-DYNA
Investigation
user's
manual (2003).
25
Chapter 2
Pressure
damaged concrete
Material constitutive parameters for the pseudo-tensor model, Equation (6) and Figure 2-25, were
developed for both 3 ksi and 4 ksi compressive strength lightweight concrete, as shown
tensile cutoff with a
on the concrete,
the constitutive
500 psi
tensile strength
a simulation
was used
Table 2-2.
is
shown
Figure 2-26. The calculated compressive stress-strain response for the 3 ksi concrete
in
in
is
check
test to
in
compared
to
and
(psi)
a:
32
(psi)"'
a,,f
(psi)
aif
26
'or
ksi
J\ =4
ksi
750
1000
0.333
0.333
l.llE-04
8.33 E-05
0.0
0.0
0.385
0.385
NISTNCSTAR
1-2B,
WTC Investigation
Initial
2 percent compression
configuration
fc
= 3.8 ksi
= 2.3 ksi
fc
fc
3.0 ksi
Strain (%)
An
additional comparison of the calculated compressive behavior to the stress-strain curve for the
NISTNCSTAR
1-2B,
WTC
Investigation
is
shown
in Figure 2-28.
The agreement
is
good up
to the
27
Chapter 2
peak load
point.
The
specified
beyond
(NIST
NCSTAR
1-6), to
is
typical of concrete
3 ksi concrete
ksi, as
-1
1.0
1.5
2.0
strain (%)
28
NIST NCSTAR
1-2B,
WTC Investigation
4 -
0*^
0.5
fc
= 3.0
= 4.0
ksi
ksi
L_
0.0
fc
1.0
2.0
1.5
Strain (%)
wide
is
is difficult,
However,
is
general, elevated strain rate loading has a greater influence on the tensile strength than
strength.
and there
in the aircraft
on the compressive
floor slab
A collection of compressive strength data for elevated rates is shown in Figure 2-30 (Bischoff and Perry,
1991
There
et al., 1992).
approximately 30
s"'.
of 30
s'',
impact damage
to the concrete is
is
more
expected to occur
up
rate effects
to a rate
of
s''.
The
is
Beyond
aircraft
a rate
measured
is still
shown
in
were added
to the
model
in tabular form.
The
rate effects
to the available
NISTNCSTAR
1-2B,
WTC
Investigation
29
Chapter 2
range
static loading
4.0
<n
A AbramB (1917)
to
<l
<D
&
Jones
G>
3.0
>
a>
D.
E
o
4-
Millstein
&
Bertero (1972)
S Thoulow (1953)
Wotstein (1953.1955)
o Hotono A Tsutsuml (196D)
Ban & Muguruma (I960)
Tokodo (1959)
Tokedo a Tochikawo (1962o.b)
Lundeen (1963)
X Horibe fc Koboyoshi (1965)
+ Cowell (1966)
Hchity & Furr (1967)
O Hughes k Gregory (1972)
'55
Mahin
Richort (1936)
Evons (1942)
W Kolsulo n943.1944)
CEB RGCommcndalions-
Sobnis (I9B2}
tn
u
a*
>
1.50
1.25
a>
-7
10
10
10
10
10
10
10
10
10
in
Mater. Struct.
24 (1991)
.2
1-31*
3
o>
c
o
to
o
<S
w
o
1
-7
-6
-5-4-3-2-101234
I
(s^)
30
NISTNCSTAR
1-2B,
WTC Investigation
nH
1-6^
3
s
JCZ
c
o
JO
>-*
to
E
c
..0
'
>
Cr
-&
-e-"-
e--
-4-3-2-101234
11
,1,1
,1
,1, 1
Log
u M
I
Strain Rate (s
Mn
i i
,1,1
^)
2.5
No
material testing
constitutive
was performed
model parameters
Handbook (Brown
et al.
in the
open
and
literature.
Handbook (MIL-HDBK-5F,
develop the
The
principal sources of
Complete engineering
stress-strain curves
that are
were provided
commonly used
in the
MIL-HDBK-5F
in the construction
These curves were digitized for the various 2024 and 7075 alloys
for various
2024 and
as
shown
in
Figure 2-34, respectively. Representative stress-strain curves were then converted into true stress and
true strain, as described in Section 2.2.1,
and used
to
model
fits
are
shown
in Figure
2-35 and
Figure 2-36, respectively. Appropriate failure criteria for the aircraft materials were developed using the
fine
Insufficient data
effects
in Section 5.5.1.
and
ductility
aircraft constitutive
of the
aircraft materials
a result,
no
rate
models.
was included
in the
uncertainty analyses in Chapter 8 to account for the potential effects of elevated strain rates.
NISTNCSTAR
1-2B.
WTC
Investigation
31
Chapter 2
80
-I
60
^&P-'
...
...^
--^-^
*--*-*-*^(^
in
V)
40
"H 2024-T3 Clad Sheet
Longitudinal
20 r
\
L.
0.10
0.05
0.00
0.15
l_
0.20
Strain
60 i
-
(0
-i
-3
"T
-3
-i
-1
40
<E>
O20 I
0
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
Strain
32
NISTNCSTAR
1-2B,
WTC Investigation
120
100
SO
'(35
cn
60
40
20
0*
0.00
0.05
_i
0.10
0.15
0,20
L.
0.25
True Strain
Plastic Strain
NISTNCSTAR
1-2B,
WTC
Investigation
33
Chapter 2
2.6
No
material testing
was performed
aircraft
tires
and
interior furnishings
and the tower contents and nonstructural walls. In general, the primary influence of the nonstructural
As
is
a simple elastic-plastic
The
modeling of
deformation and subsequent erosion from the calculations as their distortions become large. The ability to
include material failure and erosion of these soft materials
is
analyses.
in the aircraft
explicitly but
were
included by distributing their mass into the airframe structures. These aircraft components included the
interior wall paneling
mass was
and
and a
to
models
failure strain
be maintained
elastic-plastic constitutive
containers).
in a simplified seat
model.
The mass of
A bilinear
behavior was assumed for the effective seats with a yield strength of 1 .45 ksi
of 40 percent. The relatively large ductility was used to allow for the seats and cargo
long to transfer their inertial loading to the impacted tower
aircraft
the tires
landing gear. Again, a bilinear elastic-plastic constitutive behavior was assumed for the
tire
on the
material with
contents.
made
primarily of gypsum.
nonstructural building materials (gypsum, wood, engineered materials, etc.) resulted in a wide range of
specified properties.
Wood
wood
maximum
to the
wood
and
(www.woodbin.com). The
grain,
on reported values. Gypsum board properties also vary with thickness and direction with
variation
reported
400
psi.
psi to
Modulus of Rupture on
970
psi
would be expected
to
have a significantly
lower strength.
The
constitutive
model used
for the building contents attempted to obtain an approximate strength for all
assembly of workstations,
34
from the
was used
aircraft debris
is
a yield strength of
transfer
etc. that
500
psi
and a
in the
model.
failure strain
failure
momentum
had occun"ed.
NISTNCSTAR
1-2B.
WTC Investigation
REFERENCES
2.7
for
High
and
Brown, W.
F.
Jr.,
Mindlin, H., and Ho, C.Y., (Eds.). 1991. Aerospace Structural Metals Handbook,
CINDAS/Purdue University
Publishers,
Volumes
& 4.
"DYNA3D
"LS-DYNA Keyword
April.
Militarv'
Handbook, 1987. Metallic Materials and Elements for Aerospace Vehicle Structures, U.S. Dept.
of Defense.
M1L-HDBK-5F.
NISTIR 5934,
Gaithersburg,
MD,
A Report of the
State-of-the Art,"
December.
Ross, C.A.. Kuennen, S.T., Tedesco, J.W. 1992. "Effects of Strain Rate on Concrete Strength," Session
on Concrete Research
in the Federal
March.
"TrueGrid Manual. 2001. Version 2.1.0,"
XYZ
results)."
Nos. 2 and
NISTNCSTAR
3.
1-2B.
WTC
Investigation
35
Chapter 2
36
left
blank.
NISTNCSTAR
1-2B,
WTC Investigation
Chapter 3
INTRODUCTION
3.1
This chapter describes the methodology used in the development of the structural models of the two
The
structural
components used
initial efforts
6, respectively).
focused on the
in the
in a
they could be used to generate the range of structural component models found throughout the towers.
Additional information on the models used for the component impact analyses are provided in Chapter
The subsequent
efforts
were
to use the
complete sections of the towers for the global impact analyses (Chapter
for the
LS-DYNA finite
element code
(LS-DYNA
LS-DYNA input
One of the
file for
for
TrueGrid included a
mesh
properties.
set
is
9).
complete
5.
damage
distributions.
The
limitation
this limitation,
it
was
was found
model
that the
to
impact zone to
combined
be
aircraft
and
tower models should not exceed approximately 2.3 million nodes. Decomposition of a model with a
larger
32-bit system.
require additional
memory beyond
at different levels
meshed with
that
can be addressed by a
was
to
miUion
in the
damage and
failure,
while
and
their
inertial properties.
for the
components developed
WTC
in the
and
WTC
towers' structure, a key aspect of developing the global models was automating the
process.
1.5
and components outside the impact zone were meshed more coarsely to primarily capture
structural stiffness
3.2
memory
components
the
files
were
all
developed
in a
(LERA), under a
mesh
generation
and Technology (NIST) within the framework of Project 2 of the NIST Investigation; and reviewed and
approved by NIST, were used (see NIST
NCSTAR
1-2
and NIST
NCSTAR
NIST NCSTAR
files for
1-2B,
mesh
WTC
to automatically write
(VB)
master level
generation.
Investigation
37
Chapter 3
A total of three VB programs were used for the tower mesh generation.
The
generation, including
column
splice
column
details such as
model
different
shown
in
(WTC
numbers
first
or
height.
to create a
generated the desired core column models. The master TrueGrid input
statements for pre-existing parameterized generic column generation
and material
file
files.
file that
developed for the subassembly and global impact analyses consisted entirely of shell elements. Different
mesh
Edit
File
:;
View
Insert
%3
#a
Format
5?'
in the global
files
Tools
Data
^^m- <y
Window
Help
^
CollFvi
Inteiioi
Columns
55
55
Twer
Llpper Floor
101
Lower Floor
weld2
14
Cancel
[Brade BoltjS
'Splic'e"'u6iT1
IK
0 312500
Generate TG
625
3/8
A325
7/8
^
'
^
rel
'i
17 Column
Lower Splice
|Fillet
W11
No
bolts
Figure 3-1. Model user interface for the core column generator.
the
same
The
framing
basis.
details.
The user
It
also controlled
interfaces for
mesh
programs
first to
generate a TrueGrid
beam
similar.
Figure 3-2 shows the user interface for the third program that generated a series of exterior wall panels.
This program was the most complex of the three and required the user to identify the tower, upper and
left
and
right (as
mesh
Similar to the other programs, this program extracted infonnation from the database and generated a
master TrueGrid
38
file.
Information not available in the database but included in the drawing books, such
NISTNCSTAR
1-2B,
WTC Investigation
as the
w eld
w ere included
specifications,
in the
VB program.
files
included the geometrv' and material specification for the columns, butt plates, spandrels, welds, bolts, and
spandrel splice plates.
Node
commands) were
also automatically
generated to define the connectivity of adjacent parts in the model. For example, the boundary nodes in
the e.\terior spandrels
to create a perfect
in the
make
connection
in the
were
not merged.
m
Toci;
at
Data
Wtoow
Help
100%
1?)
s
S
8
9
LeftPand
Sh^Parel
MJteiePands
PaneJNunte
.\
Top
Hoy
Bottom Hoof
BcttociHow
'\i
15
IS
"l7
Rnd Dfnensions
Left Panel
Te
19
3J8
R?iCPanei
Tct!
Root
22
BottanFtoor
23
24
25
26
27
3lel
=5
Wl
No boKs
4
7/8
BoHCaqe
A325
Bolls|o12 Field
'jp
13 Field
W11/
3/B
fillel
3.5
WU/
3;B_
3.2.1
conditions
group
at the
An example model
of the
WTC
as specified in the
model and
the
column
splices
is
shown
column
in
element type
2).
These
shell
shell
LS-DYNA
NISTNCSTAR
(LS-DYNA
Hughes-Liu element
in
LS-DYNA. The
Belytschko-Tsay
widely used for crash, impact, and metal forming applications. Results obtained with
these elements typically agree with those obtained with the Hughes-Liu element.
the
elements
elements are four noded elements with single point integration, and are
sections.
Both wide flange and box core columns were modeled with Belytschko-Tsay
shell
The boundary
Theoretical
1-2B,
WTC
Manual
Investigation
on
this
The reader
element type.
is
referred to
39
Chapter 3
floor.
far field
and 8
and a coarse
in. in
single
in the
in.
identical to
elements and
floor.
WTC
shown
flange core column connections were modeled by splice plates placed on the outside of each flange, as
shown
in Figure
bond between
(LS-DYNA
this
contact type
2).
If the
columns
the splice plate spanning the gap between column ends would be
40
model
resulted
from ductile
failure
in the
shear failure of the bolts or bolt bearing failure of the splice plates
model.
NISTNCSTAR
1-2B,
WTC
Investigation
Erect/on C/earonct
is
a.
of
The
Port
Splice Plates
The
thick
box column
cap was modeled with shell elements and was perfectly merged into the lower box column. The
connection between the wide flange column and the box column cap was an edge-to-surface tied interface
without failure
column being
(LS-DYNA
perfectly
contact type
bonded
7).
to the element
plate.
Failure of this
connection would occur only when deformations and strains of this connection are sufficiently high to
the elements in the
NISTNCSTAR
1-2B,
columns adjacent
WTC
Investigation
fail
to the joint.
41
Chapter 3
^\xfs^->/ of cd^e
"y
M-^__J.y^S preporc/tJon
!
J"\_
Column
cap/splice plate
The approach
in the
in the
for
in
WTC 2.
WTC
to generate
WTC
far
away from
concrete floor slab. The entire model was developed with Belytschko-Tsay shell elements.
can be
set
Mesh
density
independently from floor to floor to obtain higher accuracy in the impact zone and
computational economy
42
floor
simplification. Figure
model was
in the
suirounding floors.
A typical
NISTNCSTAR
1-2B,
mesh
WTC Investigation
density (floor 96 of
WTC
in this case)
shell
included core floor slab, floor beams, connections, and core columns over a height of one floor.
details
WTC
and without
same manner
as
floor slab).
illustrated in
above. Interior beams were connected with node-to-surface tied connections (contact type
to-surface contact algorithm constrains the nodes to
is
NISTNCSTAR
1-2B,
WTC
model were
Investigation
The node-
beam and
move
7).
is
An
shown
automatically
43
Chapter 3
3.2.2
The
exterior walls
WTC
beam connections.
94th-98th core.
Development
plates, typically at
each floor
common
floors.
level.
The columns
44
in the
WTC
each panel
wall model required the generation of a parameterized model for each panel type that
in
exterior
was located
in the
NISTNCSTAR
1-2B,
WTC Investigation
type 300 on the sides and panel type 210 in the comers. Examples of the models for these panel types are
shown
in Figure 3-9.
and 400
series panels
The
due
to the
(a)
(b)
NISTNCSTAR
1-2B.
WTC
Investigation
45
Chapter 3
The complete
exterior wall
model
in the
impact zone for each tower was generated by placing the various
panels in the actual locations with their dimensions and material specifications. These material and
geometric parameters were extracted automatically from the tower database using the third
to automatically generate the exterior wall
WTC
respectively.
The
The impact
3.2).
VB program
WTC 2 tower (south face) models are shown in Figure 3-10 and Figure 3-11,
figures
show
a mixture
in
improved accuracy of the impact response, and a coarse mesh was used
outside the impact zone for improved computational efficiency. All panels were primarily constructed
from Belytschko-Tsay
were used
to
element type
model
1 )
shell elements.
were used
to
model
As an example, each
zone were a
in.
151
(LS-DYNA beam
beam
142
133
46
(LS-DYN A
was 14
127
WTC
in.
in.
118
model
109
floors 91-101.
NISTNCSTAR
1-2B,
WTC
Investigation
free
column ends
Figure 3-1 1
WTC
model
VB
floors 75-86.
specify the boundary conditions and various connection details between the panels.
conditions
at the
The boundary
top and bottom of the exterior wall were constrained vertical displacements.
degrees of freedom and rotation about the vertical axis were not constrained. The free
displacements
at the
mesh
5.
model
to
have a
was approximately 10
rigid
to 11
s,
body
lateral
velocity following
would provide
the tower
less than
The
lateral
column end
impact event.
butt plates
and
exterior
to automatically
bolts
was compared
to the
to represent the
measured
This representation of the bolted column butt plate connections was only provided
panels were merged together to create a perfect bond between column ends.
The model
plate
were attached
interface).
The
is
shown
weld
tied
in
node algorithm
(LS-DYNA Type
7 tied
spot welds approximate the connection of the individual bolts connecting the spandrel
splice plates. Failure of these connections occurred through deformation of the splice plates and/or
The
NISTNCSTAR
1-2B.
splice plates
WTC
Investigation
was again
in Figure
3-13. The
47
Chapter 3
influence of the spandrel splice connection on the impact response and exterior wall
investigated using engine
in
Chapter
damage was
5.
med-to-med
Medium
connection
coarse-to-med
connection
(spliced]
(spliced)
coarse-to-
coarse conn,
(merged nodes)
Coarse
^""In
panels
Spliced
^^-r^
connection
ifli|
48
panels
in
NISTNCSTAR
1-2B,
WTC
Investigation
3.2.3
The approach
to the
development of the
truss floor
model w as very
comer two-way
trusses.
truss
for
segments of long-span
individual truss floor segments spanned the distance from the exterior wall to the core.
truss floor
two
segment used
trusses
model
in the global
were modeled
is
trusses,
shown
in Figure
An example
of a
floor trusses
accounted for the potential for buckling of the bottom chord under impact loading.
combined
(secondary) trusses, using Belytschko-Tsay shell elements for the truss upper and lower chord
components. Belytschko-Tsay
shell
and bridging
was adapted
much
larger
in this
in the slab
model
model
trusses.
for
an offset tied
and
model
size requirements.
Development
merged
size.
The model
beam elements
beam
would
tower model
shown
in
impact tower model with these detailed floor segments was not practical due to model size limitations.
result in
in the
is
Figure 3-15 and Figure 3-1 6 show truss floor connection details
The models
a resuh,
respectively.
As
at the exterior
shell
using the tied interface algorithm. The failure of the seats only occurs as a result of exceeding the
ductility
NISTNCSTAR
1-2B.
WTC Investigation
49
Chapter 3
Truss
Beam
Core Perimeter
Beam
The primary
was
and the
of the
building under wind loading. These dampers, however, were of low mass and the arrangement of the
the
damper
to the
truss),
connections, had virtually no strength in the transverse direction. Under impact conditions, the aircraft
applied transverse forces to the damper assembly due to the
Chapter
50
7).
Also, due to the short duration of the impact event (less than
s),
aircraft (see
NISTNCSTAR
1-2B,
WTC
Investigation
in the analyses.
As
a result, the
in the
to
impact analyses.
A model assembled for an entire floor (floor 96 of WTC 1), including the core and exterior wall
structures, is shown in Figure 3-17. A comparison of the plan view and the associated framing plan
drawing for
this floor is
shown
in
Figure 3-18.
WTC
Framing Plan
(Ref
DWG
No. SA-104)
FEA Model
<6~5_
jT'i:-
of
The
iT
ji-
i-"'
Port Authority of
-f-j-t..^
^- .^
New
..
v.,
HJft.
York and
New
Jersey.
Figure 3-18. Model plan view and framing plan drawing of the 96th floor of
NISTNCSTAR
1-2B,
WTC Investigation
WTC
1.
51
Chapter 3
3.2.4
The
Interior
The
live load
mass was
The
distributed
in the
distribution of the
gypsum
walls
drawings and other infonnation gathered as part of the NIST reconstruction of the thermal
NIST
develop the approximate placement of workstations over the truss floor area. The resulting model of a
floor with interior contents is
shown
in Figure 3-19.
Heavy
and counterweights inside the core's elevator shafts were not included
way of determining
in the
Workstations Modeled
over Truss Floor Area
The
were scaled
WTC
1,
and
live loads
superimposed dead loads. The densities of the tower contents (workstations and gypsum walls) were
scaled by the appropriate ratios to get the desired distribution of live loads in the core and truss floor
areas.
The
to
obtain the desired superimposed dead loads. These additional loads were important for obtaining an
accurate mass distribution in the towers and inertial effects in the impact response.
was based on
on the
floors inside
The
used
live load
25 percent of the design hve loads resulted in 19.7 psf uniformly distributed over the core and 16.2 psf
uniformly distributed over the outer truss floor area. Using the approximate floor areas of 8,700
core and 31,250
ft"
ft"
in the
outside the core produced 170,000 lb and 500,000 lb of live load for the core and
to the
and
1 1
.5
core producing 3 15,000 lb and 359,000 lb of superimposed dead load for the core and outer floor areas,
respectively.
structural
52
components
(slab,
columns, beams,
etc.)
in the
NISTNCSTAR
1-2B,
WTC Investigation
components such
as insulation, cladding,
to the
columns
and windows.
A summary of the live loads and superimposed dead load magnitudes and density scale factors are
provided
in
Table 3-1
was
The
total
1.35 million lb resulting in a total weight for the floor of 4.41 million lb (after the
Summary
Table 3-1.
of
live
loads (per
floor).
Area
Weighting
Added Weight
Density
(ft^)
(psf)
(lb)
Multiplier
8.694
36.2
3.15E5
1.628
31,257
11.5
3.59E5
1.218
8.694
19.7
1.71E5
0.852
31.257
16.2
5.06E5
0.447/0.757'
39.951
N/A
1.35E6
N/A
Load
Total
a.
The
partitions
(partitions
floor, as
shell elements.
shown
in
97,284 shell elements. To include the complete distribution of the building contents over five floors in
the global impact
models
included the partitions and workstations only in the region of each floor directly in the path of the aircraft
this
WTC
global impact
to
model
included only 160,410 nodes and 148,858 shell elements for the partitions and workstations in the impact
path over five floors (savings of about 300,000 elements).
All of the contents in the truss floor regions rested on the floor with a contact algorithm used to prevent
penetration. Gra\ itational acceleration and friction included in the contact analyses resisted sliding
motions of these contents. The bottoms of the gypsum walls in the core were rigidly connected to the
floor slab using
in the
As
a result
core walls were not constrained. This approximation did not have a significant impact on the response in
the core.
An
momentum
The
majority of the core wall damage in the simulations was failure and erosion rather than displacement or
3.2.5
The
global
in the
The
floor
96 of the
NISTNCSTAR
WTC
1-2B.
was
shifted to be centered
on the floor
system
is
shown
in
An
Figure 3-20.
Z=0
at
WTC
Investigation
53
Chapter 3
associated with contact or small interference dimensions of components relative to large vertical
The multiple
floor global
in
WTC
is
shown
WTC
model.
in Figure
3-2 1
The model
included the complete floor inside and outside the core, the exterior walls, and core structures for
floors 92 through 100.
in the longitudinal
and
lateral directions
and found
to
have
little
was
in Figure
the top
vertical axis.
The
shown previously
at
The higher
resolution exterior
The
WTC
shows how
global impact
the
exterior wall
removed
is
shown
in
workstations) were not modeled over the entire floors in the impact zone, but rather only in the path of the
aircraft
distribution
debris,
to
in the
(partitions
Figure 3-23.
in
As
and workstations)
in the path
this
a resuh, the
less.
The
of the aircraft
have a significant effect on the impact behavior since these components do not contribute significantly
to the surrounding
contents outside the impact zone were not sufficiently stiff or rigidly attached to the tower to have a
54
NISTNCSTAR
1-2B,
WTC
Investigation
impact.
^
Figure 3-21. Multi-floor global impact model of the
The
Side 100
WTC
tower.
separate view of the truss floor structures adjacent to Side 100 for floors 94 through 97
is
shown
in
Figure 3-24. The sections of the truss floor in the path of the aircraft impact were modeled with the
detailed truss floor
modeled using
model described
a far-field truss
in
model with a
significantly reduced
mesh
resolution as
shown
in
Figure 3-25. This far-field truss floor model was designed to have the appropriate inertial properties and
structural stiffness.
structure
The
modeled with an
dimension of 30
to accurately
identical
in.
model
floors
local
columns
compared
to an
NISTNCSTAR
that occurred
1-2B,
WTC
of the
truss.
shown
damage
The
floor slab
in.
was
approach as the floor slab in the impact zone, but with a typical element
the core
and a beam element along the lower cord. These element dimensions were on the order of 30
The
truss
direct
Investigation
WTC
from
in
in. in
that in the
55
Chapter 3
Floor:
(a)
Oblique view
II
11
tg
i.i
lli
i.iJ.i.Li J
li
,i,,,U
Floor:
.Qli
.i.l.
l.i
i,l
1.1
100
LJ--
Higher mesh
density in core,
floors 93-98
(b)
56
WTC
NISTNCSTAR
1-2B,
WTC
Investigation
in
the
NISTNCSTAR
1-2B,
WTC Investigation
in
WTC
the
WTC
57
Chapter 3
model
floor global
WTC 2
is
in
the
shown
WTC towers.
in
included the complete floor inside and outside the core for floors 77 through 85. The exterior wall panels
at the
at the
75.
same
WTC
model. The higher resolution exterior wall panels in the impact zone, shown previously in Figure 3-11,
can be seen on the impact face of the
The
WTC 2 global impact model with the exterior wall removed is shown in Figure 3-27. This figure
the WTC 2 model was optimized to reduce the mesh size and eliminate computational
shows how
shown
in the
damage
path of the aircraft impact and debris cloud. These components are
separately in Figure 3-28. Similarly, the truss floor structures near the impact zone were
in greater detail as
adjacent to Side
The surrounding
modeled
seen in Figure 3-27. These detailed sections of the truss floor were positioned
400 (south
far-field truss
and 82.
model shown
To summarize, the approach adapted for the development of the global impact models for towers WTC
and WTC 2 was similar. The structures were modeled primarily with shell elements with some beam
elements in the detailed truss floor sections and bolted connections between the column ends in the
Some brick elements were used at the exterior column butt plates
summai7 of the model size and element types for the global tower models
in the
is
58
NISTNCSTAR
1-2B,
WTC Investigation
NISTNCSTAR
1-2B,
WTC
Investigation
59
Chapter 3
direction
Table 3-2.
Summary
in
of Nodes
3.3
WTC
WTC
Number
the
models or the
1
WTC towers.
Tower Model
1.300,537
1.312,092
47,952
53,488
1,156,947
1,155,815
2,805
2,498
REFERENCES
"LS-DYNA Keyword
April 2003.
XYZ
60
September 2001.
NISTNCSTAR
1-2B.
WTC
Investigation
Chapter 4
INTRODUCTION
4.1
The
finite
aircraft
was constructed
(3)
meshing of the
structure.
The
focus of this effort was on gathering sufficient structural data and including adequate detail in the aircraft
model so
that the
distribution
of the
aircraft
A fundamental
that could
memor\'
for
domain decomposition,
the
combined
this
aircraft
model
clusters.
finite
LS-DYNA
uses addressable
aircraft
mass
distribution,
and debris
into
(WTC)
towers.
to
model
beam elements
aircraft
components
elements in the airframe provided higher fidelity simulation of the impact response and fragmentation
beha\
ior.
As
a result of the
model
explicitly.
size constraints,
Where modeling
some of the
details
and smaller
structural elements
applied to ensure that the impact strength and break up behavior were maintained.
An example
structure
u as
first
fine
mesh of the
detailed
wing
A section of the
aircraft
behavior for aircraft structural failure and fragmentation upon impacting the exterior wall of the
towers.
failure criteria
were modified
how
this
The following
WTC
to obtain similar
wing
similar
mesh
the
fine, detailed
methodology used
for
aircraft.
model developed
modeling each major component including the wings, engines, fuselage, empennage, and landing gear are
provided. The final section provides an analysis of the fuel distribution in the aircraft
at the
time of
impact.
NISTNCSTAR
1-2B,
WTC Investigation
61
Chapter 4
4.2
Structural data
were collected
PW4000
and Whitney
Company
website.
for the
CAD drawings,
Boeing 767-200ER
aircraft
literature, electronic
Engine Reference Manuals, American and United Airlines, and the Boeing
The following
is
a description of
how
model
development.
4.2.1
Two
electronic surface
was not
To
able to provide reference information about the data used in the surface
was compared
result,
Digimation
model development.
to electronic
two-dimensional
CAD descriptions obtained from the Boeing Company web site, as shown in Figure 4-1. The first model,
VP331 shown in Figure 4-1 (a), did not match the CAD descriptions well. However, the second model,
VP 157 10 shown in Figure 4-1 (b), matched the CAD descriptions, showing only minor discrepancies in
1
The
VP 157 10
used
was subsequently
in the aircraft
model generation
to
(a)
Model VP3311
(b)
drawings of a 767-200ER.
Aircraft Inspection
4.2.2
No
CAD
Model VP15710
Boeing 767-200
aircraft
was available
767-300ER.
Figure 4-2.
On
angle,
the
is
in
and parts of the structure are reinforced. Aside from these differences, the two
identical.
62
NISTNCSTAR
1-2B,
WTC Investigation
Aircraft
r-'-f
Used
with permission.
ARA personnel
767-300ER
to take
CAD
shown
-4
:~
Enhanced by NIST.
fi innt
Model Development
in
aircraft.
to the
Photographic documentation of
aircraft.
For the landing gear, global dimensions were taken along with ultrasonic thickness measurements of the
major
structural
components, as shown
in
the photograph by the blue tape positions in Figure 4-4. Locations, dimensions, and approximate masses
of various non-structural components such as the overhead luggage containers, overhead electronics, and
under floor electronics were also documented, as shown
NISTNCSTAR
1-2B,
WTC
Investigation
in
Figure 4-5.
63
Chapter 4
(b)
(c)
Non-structural contents
Source: NIST.
64
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
(a)
(b)
Ultrasonic
Model Development
Nose gear
measurement
locations
Source: NIST.
NISTNCSTAR
1-2B,
WTC
Investigation
65
Chapter 4
Fore-Afl Length
Shell
3-5
Is
Vi" fiberglass
honeycomb
Luggage {gre^r:) and
W,
jO-OSM)
electronics
{1}
2c)
fill
remainder of space
Listed load limit
is
871b per
40 FT 5
IN.
'
"
(12.32M)
-21 FT 11
(from GrabU's
(b)
Used
with permission.
Under
Pratt
Ul9FT
10IN.
floor electronics
Enhanced by NIST.
4.2.3
'
mass
locations
in
the fuselage.
& Whitney PW4000 engine and the General Electric CF6-80 engine
were on the aircraft that impacted the WTC towers (FEMA 2002) For this reason, the Engine Reference
Manuals were obtained from Pratt & Whitney for the PW4000 turbofan engine. These manuals contained
Initial
detailed information
66
PW4000
in Figure 4-6.
A detailed finite
NISTNCSTAR
1-2B,
WTC Investigation
Aircraft
Used
Model Development
with permission.
PW4000
American Airlines
flight
that
1 1
impacted
WTC
was
Electric
PW4000
all
in the
& W hitney
Pratt
PW4000-94
Engine
Fan Blade Diameter
94
Length
153
Dr>'
Weight
(in.)
(in.)
9,400
(lb)
Pratt
& W hitney
JT9D-7R4'"
94
153
General Electric
CF6-80C2"'
93 (m.)'
(in.)
(in.)
8,885 (lb)
161-168
(in.)'
9,135-9,860 (lb)
d.
e.
a.
b.
c.
counter weights.
in.
also found
NISTNCSTAR
1-2B.
WTC
is
much
Investigation
in. in
length.
67
Chapter 4
Airline Data
4.2.4
The Boeing Company provides some general infomiation on the layout of their aircraft at their website
(www.Boeing.com). Useful data on the seating layout and dimensions and the Unit Load Devices (ULDs
or cargo containers) were available, as shown in Figure 4-7 and Figure 4-8. The cargo, fuel, and
passenger and crew masses were also provided by the airlines for American Airlines
(Condon
&
of 6,625
lb
was assumed
It
(UAL175)
lb.
Economy
class seating
flight
1 1
(AAl
1)
22
ULDs
was estimated
when empty.
EXTEBIOH
198 IN (5.03Mj
96 IN
(S
44M)
LOOKING AFT
15 IN.
{0,48M)
2 m.
62
IN.
10M)
(1
i
1
1,
..
/^TVP
DOOR
<t.07BYtS3M)
ac
8V
IB
EMEBQE^Y
38
EXIT
B*. (0.61
DOOR
(BH
t,
LH)
BV 0.9m)
.-jfe^i3ii9iiiiiiiiia iisaiaiaiaaiiaaaiaai
[u
W
^^aaaaaaaaaaaaaaaaaareBaaaaaaaaaaaaa^-^
<I
07
8Y
it?
22
Used
PASSENQEBS
AT 32-\H
(O fltM)
PITCH
with permission.
Figure 4-7.
68
98M|
Economy
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
10-2 DATA:
CONTAJNER WEIGHT
LOADED WEIGHT
VOLUME
Used
Model Development
CROSS-SECnON
M)
with permission.
Figure 4-8. Location and characteristics of the unit load devices (www.boelng.com).
\A
175
14,720 lb
9.4101b
Freight
7.972
16,970 lb
1.1501b
Luggage: Carry on
1,620 lb
1,010 lb
Catering
5,234
30,696
Total
4.3
UAL
11
lb
,390 lb
lb
28,780
lb
lb
LS-DYNA
model of the
aircraft
meshed using the TrueGrid software (TrueGrid Manual 2001). The complete model for the Boeing 767200ER is shown in Figure 4-9. A summary of the model parameters is presented in Table 4-3. The total
empty mass represents an average
cited
lb.
seats,
Company
to the
Fuel was distributed in the wings as shown in Figure 4-10. Analysis of the fuel distribution
with a
tip deflection
aircraft
shown
of approximately 52
photographs of the
NISTNCSTAR
WTC
1-2B,
Investigation
is
ULDs.
discussed
in
in.
1989) for an
includes the
Figure 4-1
A cubic
to
was used
damage documented on
by NIST.
69
Chapter 4
70
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
Model Development
AA
L'AL 175
11
70.000
70.000
562.000
562.000
60.672
60.672
740.000
740,000
No.
SPH
Fuel Particles
Total
Nodes
Total
\\
eight (Emptv')
183.500
lb
183.500
lb
ULD'Cargo Weight
12.420
21.580 lb
10.420 lb
Fuel Weight
66.100
62.000
283.600
lb
lb
lb
21.660
277.580
lb
lb
lb
NISTNCSTAR
1-2B,
WTC
Investigation
71
Chapter 4
4.3.1
in the aircraft.
for the
fuselage,
structural
completely using Belytschko-Tsay shell elements. Models for the landing gear and engines were
primarily developed using shell elements, but contained
developed
in a
model
was
to
size
parameterized form where the mesh resolution was detennined by a single element
the
some
was selected
and resolution as the model development and impact analyses progressed. The objective
in. to
in.
in.
and 2
in.
for small
components,
wing or
fuselage skin.
hubs, the
main
strut
and
truck,
in red
72
in the
is
shown
in
shown
in
modeled
are
all
all
in detail in the
The
with blue edges). The density of these containers was scaled to include the weight of the cargo.
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
(b)
Side view
Figure 4-12.
Empennage model
(c)
for the
Model Development
Oblique view
767-200ER
aircraft
model.
Figure 4-13. Retracted landing gear components for the 767-200ER aircraft model.
NISTNCSTAR
1-2B,
WTC
Investigation
73
Chapter 4
Figure 4-14. Underside of the 767 airframe model (skin removed) showing retracted
landing gear.
Figure 4-15 shows the model of the wing structure, including the center wing, which attaches the port and
The wing
stringers
were not
explicitly
modeled
model. The stringers have a z-section geometry with typical dimensions of approximately
and
two
in.
web with
a thickness
wing
of approximately
ribs as
shown by
part
1/8 in.
in.
of the
flanges
number 178
in
Figure 4-16.
To account
for the
weight and strength of the riveted skin/stringer construction, an 'effective' wing skin was used, as
discussed in Section 4.3.2.
Rib 35
(a)
(b)
Figure 4-15. Complete wing structures for the Boeing 767 aircraft model.
74
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
Model Development
Figure 4-16.
Wing
structural diagram of a
Boeing 767-200.
A model of the fuselage was assembled with a stringer and frame constmction supporting the external
skin, as shown in Figure 4-17. A tied interface was used to connect the stringers to the frames and skin
using the tied surface to surface contact algorithm in
LS-DYNA.
surface are constrained to nodes on a master surface provided they are within a certain distance of the
was
is
Due
to
shown
model
in
wing
to the keel
The wing
forward and
aft
modeled
without the detailed stringer/frame construction. Instead, the weight of these components was smeared
into the skin
by increasing the skin thickness and scaling down the strength by a factor of 40 percent,
as
NISTNCSTAR
1-2B,
WTC
Investigation
75
Chapter 4
Wheel Well
UAL
tlie
was increased
to include the
structure not
in
total
weights
AA
1 1
and
75 were due to differences in passenger and cargo weight. In both cases the weight of the cargo,
to
account for
the weight of small structural details, such as stiffeners, not specifically modeled, as well as hydraulic
lines
and fluid pumps, actuators, inboard flaps and outboard ailerons, flap and rudder connections, and
a smaller
amount
( 1
.5)
account for hydraulic fluid and smaller structural components not included in the model. The weight of
the fuselage
was adjusted
the fuselage
was
to
components
is
many heavy
reasonable as
conditioning,
aircraft.
power
units,
cargo floor, actuator motors, insulation, hydraulics, galley, and lavatories). These structural and nonstructural
4.3.2
in detail
due
to the constraints
size.
6, respectively).
tip,
The
full
wing contains 35
The wing
76
on model
ribs,
section
shown
and subassembly
with rib
in
level analyses
model developed
and
rib
in this section
Figure 4-19.
NISTNCSTAR
1-2B,
WTC Investigation
Aircraft
Model Development
^p
p n c i It.'
Component
Factor (AAll)
Wings
2.0
Empennage
2.0
Fuselage
(AAll)
37.000
lb
8,350 lb
103.050
6.68
Landing Gear
1.5
8.400
1.2
20.100
(UAL175)
Factor (UAL175)
2.0
37,000
2.0
8,350
lb
lb
lb
103.0501b
6.68
lb
lb
1.5
8,400
1.2
20,100
lb
lb
ULD
1.43
12.400 lb
2.50
21,650
lb
Seats
1.29
28,200
lb
0.78
17,050
lb
Fuel
1.0
66.100
lb
1.0
62,000
lb
Total
W eight
283,600
277,600
lb
4^
(a)
lb
-r
(b) Internal
Figure 4-19. Wing section model for component level and subassembly analyses.
The main
spars,
wing
ribs,
included in the wing section model. The internal structure of the wing
as
shown
in
slats,
(i.e., ribs,
To
spars, nose
beams,
all
etc.),
connection between the external skin and the internal structure, the skin was attached to the leading edge
ribs, the spars,
and the
ribs
is
trailing
two
to
account for
The wing
stringers
shown by
in the
model
part
would require a
slat actuators,
mass.
this nonstructural
structure of the
rear spars as
ribs,
large
in
front
and
if
modeled
explicitly.
The individual
stringers
were not expected to have a large effect on the impact response due to their relatively small cross section
compared
to
tower structures. As a
were not
wing model
as
they add significant complexity and size to the model. Instead, the stringers were included in an
'effective"
The
w ing skin
effective
in the
model used
for the
to
construction with a reduced overall model size. For this purpose, a simplified wing section model,
NISTNCSTAR
1-2B,
WTC
Investigation
77
Chapter 4
containing a unifoiTn stringer-skin construction and a simple rectangular cross-section, was also
to evaluate different
corresponding skin strength and weight. The wing section component models utilized Belytschko-Tsay
shell elements.
and stringers
(tiebreak)
Rigid
constraint
at floor
Stringers attached to
locations
ribs (tiebreak)
(a)
(b)
Impact configuration
Figure 4-20. Simplified wing section model and impact analysis used for the effective
skin model development.
A suitable effective skin was developed by first modeling a wing segment containing the stringer and skin
shown
construction, as
in Figure 4-20(a).
and bottom
skin.
In this
ribs
in the
section model.
was impacted
spars,
The
was assumed
all
A single
in the simplified
wing
LS-DYNA.
meshed with
model
Ribs, stringers,
surface-to-surface tiebreak contact algorithm uses a biaxial failure criterion governed by the
equation:
\2
-I-
where an and
are the
>1
(7)
are the
When this
failure criterion is
met, nodal displacements between tied nodes are no longer constrained, except to restrict penetrations of
78
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
were used
0.063
for
14
3/16
where
in. plates,
Assuming
in.
in.,
mode remained
and
1/4
rivets.
in.).
The
Various
to
same
rivet diameters
was used
largest
the
be
0.
The
User's Manual
shown
is
for 5/32
in.
rivets in
in the aircraft
in.
LS-DYNA Keyword
in
diameter
shown
refen-ed to the
this algorithm.
is
Model Development
rivet
The
in.,
a^.
C"
55
L - 8nun
Source: Reprinted from Finite Elements in Analysis and Design, Vol. 30,
Langrand, B., E. Deletombe, E. Markiewicz, and P. Drazetic, "Riveted join
modeling for numerical analysis of airframe crasfiworthiness," Figure 1 on page
23 and Figure 12 on page 36, Copyright (2001), with permission from Elsevier.
Figure 4-21. Riveted joint configuration and failure data under mixed-mode loading.
Two
alternatives
alternative
was
were considered for an effective skin of the riveted stringer/skin construction. The
of the skin to include the mass of the stringers. The second
first
option was to increase the density of the skin to include the mass of the stringers. In both cases the skin
yield strength
to
effective skin
wing segment
with that of the riveted stringer/skin construction. In both options, the effective skin was only riveted to
the front and rear spars and not the ribs. This behavior most closely resembled the skin behavior seen for
the
model with
column
is
stringers
shown
NISTNCSTAR
in
1-2B,
and
skin.
Damage
to
Figure 4-22.
WTC
Investigation
79
Chapter 4
Effective Skin
Effective Skin 2
damage
The
to
evaluate the effective skin performance. Yield strengths for each of the effective skins were modified
until the residual kinetic
energy of the wing segment debris most closely matched that of the model with
closely
shown
in
match the kinetic energy of the riveted stringer/skin construction. Also shown
2,6x10-
I
1
'
'
'
'
'
'
'
'
'
'
'
'
'
'
9B~
'
'
'
'
'
'
'
'
'
'
'
Tiebreak (Rivets)
Tied
Stinger/Skin
Not Tied
Effective Skin
'
'4.
li"
Stinger/Skin
-i Stinger/Skin
2.4
to
constmction.
tied) stringer/skin
'
made
Effective Skin 2
2.2
LU
2.0
0)
k
1.8
.6 H"i
,1
,1
,i,ji
il,,i,,i
Time
10x10
(s)
80
NISTNCSTAR
1-2B.
WTC Investigation
Aircraft
Model Development
is
shown
in
Figure 4-24. Both effective skin options simulate reasonably well the damage to the exterior columns.
also showTi physically in Figure 4-25.
This
is
(i.e.,
construction or the
first
damage
However, the
to the exterior
aircraft
first
the stringer/skin
was
models.
Time
(s)
Stringers
Figure 4-25.
NISTNCSTAR
1-2B.
Damage
WTC
&
Effective
to an exterior
Investigation
column from
Effective
81
Chapter 4
4.3.3
The
Pratt
&
The engine
damage
PW4000
Whitney
is
to the
Used
to capture the
&
Whitney
identified
As
was given
to the
turbofan engine.
details.
drawing
In addition, the
structural
engine dimensions were used to detennine the scale factor for the drawing. The simplified
common
Used
with permission.
Figure 4-27.
82
Figure 4-26.
components were
Known
in
with permission.
drawing had
shown
an important component of the aircraft with the potential to produce significant impact
is
digitization procedure.
An
Enhanced by NIST.
PW4000 engine
NISTNCSTAR
1-2B.
WTC Investigation
Aircraft
rr
Used
with permission.
Figure 4-28.
Once
Model Development
Enhanced by NIST.
PW4000 engine
was captured,
was imported
into
TrueGrid and
used to generate surface definitions and part geometries for the engine model. The engine model was
shell elements.
dimensions between
and 2
in.
in.
to
locations to capture details of the engine geometry. Brick elements were used for
at
many
some of the
thicker
hubs and the roots of the compressor blades. The various components of the resulting engine model are
shown
in
Figure 4-29.
A summary
NISTNCSTAR
1-2B.
WTC
Investigation
in the
engine model
is
given
in
Table 4-5.
83
Chapter 4
9,560
54,788
Total
After the
Nodes
101,822
7,873 lb
9.447
known primai^
structural
lb,
lb
at
7,873
lb.
lb
in the
and 9,860
listed at
lb.
These
initial
in the
in
model. To account for the difference, the density of all of the material models
used for engine components was increased by 20 percent. This effectively smeared the missing mass
in
proportion to the original mass distribution in the model. The resulting adjusted engine model mass was
9,447
lb.
4.4
An
fire initiation
was
the distribution of
UAL
flight
75 and
AA flight
1 1
at the
estimated that flight 175 contained approximately 62,000 lb or 9,1 18 gal of fuel
at
evenly distributed between both main tanks."' American Airlines estimated that flight
66,081 lb or 9,717 gal of fuel
at
1 1
left
contained
shown
Fuel tank locations and capacities for the Boeing 767 are
integral fuel tank
fuel tanks
skin, ribs,
and center
is
rib 3
9.
5.
12,
Ref.;
84
from
to rib 34.
shown
is
The center
rib to rib 3.
from
onboard the 767-200ER, a main tank, a surge tank, and auxiliary tanks. The auxiliary tanks
at the
in Figure 4-30.
and spars serve as the fuel tank. There are three classes of
rib 3 to rib 3
wing
1
auxiliary tank
A dry bay
is
in Figure
4-3
is
2003,
"NIST
1 1
Litigation
C&F
DTB/MH28079."
NISTNCSTAR
1-2B.
WTC Investigation
Aircraft
Si>
A E
WHEEL
MB:Ei.S
GAU.ONS
12.000
-tOOER
!2,000
1T.2$1
8.310
31457
45.i25'
f-JSMli.t P1.!JQ5
OtJSWIEe TO Mtit
UTERS
4.5S0
FtRE
STANOARO
-200 EH
-390 ER
Model Development
i^Mt>ACH
AFT
i-A*iC?3K*^
t^iKE Alt
rOISWA^
WritiLi A!>
1
1
Used
45.425
with permission.
in a
NISTNCSTAR
1-2B,
WTC
Investigation
in
85
Chapter 4
wing but
acts as a
two-way
is
fuel baffle to
minimize
fuel slosh.
rib
allows
some
ribs 6
and
9,
bamer running parallel to the rear wing spar. In addition, baffle ribs (ribs 5 and 18) include
of fuel dams that act as a one-way valve, allowing fuel flow inboard toward the sump areas (low
include a fuel
a series
point of tank). According to the statements from both airlines regarding fuel distribution,
that the surge tanks
and
all
were diy
auxiliary tanks
at the
it
is
most
likely
time of impact.
Overall tank dimensions and geometry were estimated from aircraft stmctural references, the surface
CAD
is
good approximation
is
shown
in
Figure 4-32.
wing
As shown
section.
in
dimensions, the fuel tank capacity as a function of the distance along the wing buttock line (Figure 4-34)
was
calculated.
The
actual
modified to match
this
is
maximum
is
to
be approximately 6,500
gal.
6,070 gal (Figure 4-34), so the calculated fuel capacity distribution was
value, as
shown
in Figure
at the
time of
impact.
935.12
Dimensions
in
inches
Dimensions
in
inches
86
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
Model Development
Cross-section
Plane
Wing
skin
Approximate Fuel
^^^^
Rib 18
rib locations.
NISTNCSTAR
1-2B. WTCInvestigation
87
Chapter 4
The exact
The
was complicated by
perfonned extreme
flight
maneuvers prior
to
could cause the fuel to flow inboard quickly. High loads on the wings due to the extreme flight regime
the time of impact
For simplicity,
it
at the
was assumed
it
is
reasonable that
could flow
time of impact.
at the
capacity at the outboard baffle rib and the fuel onboard are approximately the same, a good
was
that the
rib for
full
AA
at
inboard of baffle
flight
1 1
first
estimate
This
A full
rib
is
Fuel
flight.
REFERENCES
4.5
Badrocke, M.
& Gunston, B.
Aircraft
2,
www.boeing.com.
Condon
1'^
NY,
2001.
Langrand, B., E. Deletombe, E. Markiewicz, and P. Drazetic. 2001. "Riveted joint modehng for
numerical analysis of airframe crashworthiness," Finite Elements in Analysis and Design, 38.
88
NISTNCSTAR
1-2B,
WTC
Investigation
Aircraft
"LS-DYNA Keyword
Model Development
April 2003.
Midgett.
J.
T. 2003.
Airlines and
"NIST
XYZ
WTCl- 156-1.
NISTNCSTAR
1-2B.
WTC
Investigation
September 2001.
FEMA 403,
May, 2002.
89
Chapter 4
90
left
blank.
NIST NCSTAR
1-2B,
WTC Investigation
Chapter 5
introduction
5.1
The
failure
phenomenon of the
aircraft
to (1)
start
(2)
required for the global analysis of the aircraft impacts into the
towers.
The
of key components of the tower structure and progress to relatively coarsely meshed beam and shell
element representations to be used for the subassembly and global models. This was done to develop
reduced
finite
component with
fine details
fidelity global
at the start
fuel.
of the project:
An
exterior
column impacted by an
aircraft
engine
An
interior
column impacted by an
aircraft
engine
An
exterior
column impacted by an
aircraft
In addition to the
were identified
fuel
that
in
constitutive
the tasks
above
to
Once preliminary
to
produce the
it
was
found, for example, that the load generated by an impacting engine would totally overwhelm an interior
or exterior column. Reduced shell element models based on this severe loading would match the detailed
brick model, but the subtle response from lesser loading might not be as accurate.
To
capture the
more
column components, wing section impactors were used as an alternate to the engine
for the majority of component analyses. It was found that impact of an empty wing section produced
damage to the exterior columns that was near the failure threshold. Similarly, a fuel-filled wing section
subtle response in the
NISTNCSTAR
1-2B.
WTC
Investigation
As
a result, the
91
Chapter 5
impacting an exterior column and a fuel-filled wing section impacting a core column were the primary
same
time.
As
columns (Sections
a result, the
wing
5.2.1
the final
for the
tower and
aircraft structures
at the
a prelimJnaiy
was
to
that
is
interior core
of a lesser
fidelity than
methodologies for the tower components. The preliminary wing model was sufficient for obtaining a
representative impact load.
Analysis Methodology
5.1.1
LS-DYNA
structures
is
the
LS-DYNA
(LSTC
2003).
The
at
finite
explicit finite
(LS-DYNA
element code
the
).
DYNA3D finite
initial
Version 971
the code
has been extensively developed and supported by the Livermore Software Technology Corporation
(LSTC) and
LS-DYNA
Particle
is
blast,
that
dispersion
damage
is
crucial for assessing the impact loads inside the tower structures
to the
in this report
fine
mesh
fuel
analyses of
and debris
mechanical systems.
The
provided in the
is
detailed
is
LS-DYNA
component analyses
is
the
Theoretical
A significantly
Manual
solid
detailed description of
(1998).
LS-DYNA.
in
element type
in
LS-DYNA
due
hourglassing or zero energy modes. There are several methodologies for controlling hourglass
LS-DYNA. The
were included
plastic strain)
strictly
were tracked
had a
criteria
to
modes
in
fine
mesh. As a
Damage
criteria
result,
(such as
damage and
maximum
of damage and
impact simulations.
92
was
its
modes.
to
is
stress,
and the
strains in the
in the
NISTNCSTAR
1-2B,
WTC
Investigation
associated mass of the elements remains with the nodes in the calculation. If adjacent elements have not
reached the failure surface, the nodes remain attached to the structure. If all of the elements connected to
a specific node have failed, the node
becomes a
the calculation or remain in the calculation with associated inertial properties and potential for impacts
against other structural components (free nodes remain in contact algorithms).
As
the
size
Belytschko-Lin-Tsay shell elements. These are four node shell elements with single point integration.
LS-DYNA and is
is
Hughes-Liu element
in
forming applications. Results generated with the Belytschko-Lin-Tsay element typically agree with those
generated using the Hughes-Liu element.
introducing
damage and
As used
was through
the constitutive
to
erosion.
example
is
the interface
this
model
internal
4.3.2).
modes
in the analysis.
connection and failure during the impact event was the tied interface with
failure criterion
was
move
failure.
to
In this
An
aircraft.
and shear
move independently
failure stresses
Overall contact in the impact analyses was modeled using the automatic single surface contact algorithms
in
LS-DYNA.
Interacting
automatically generated by
LS-DYNA.
a material
list,
was used
in the
contact algorithm that determined the contact stiffness based on stability considerations, time step size,
was found
to
formulation for modeling the complex contact behaviors in large impact and crash simulations.
5.2
As described above,
the
proceed
Detailed models of the aircraft components, and in particular the wing components,
were not
in parallel.
initially available
structures.
Therefore, the
many of these
due
to the
initial
aircraft
all
following sections describe the development and analysis of the tower components.
5.2.1
Exterior
Various exterior column component impact analyses were performed with different objectives. The
preliminary impact analyses were performed on a single column using a highly refined
elements. These analyses were used to investigate details
NISTNCSTAR
1-2B,
WTC
Investigation
mesh of brick
to
develop analysis
93
Chapter 5
techniques that could be applied to the subassembly and global impact analyses. For these analyses, a
simplified wing-section impactor
was used
design was not yet available. Subsequent exterior column impact analyses were perfoiined using less
refined models of the exterior
columns
panel configurations.
in full
More
detailed
were developed and used in these analyses with the primary objective of studying the impact response of
the aircraft
analyses.
wing
An example
merged together
42
specific
ksi.
(i.e.,
failure criterion.
The
and rear
had
in the
the forward
fluid
at
as
The
in Section 5.5.
shown in Figure 5-1. The model of the exterior column consisted of three floors of a single
column with a relatively fine mesh of brick elements. The plates that make up the exterior
470 mph,
exterior
structures
fuel
spar.
was added
The
fuel
constitutive
a negligible strength
and
plastic strain
a simple constitutive
model
elements was allowed (250 percent strain) prior to removal from the calculation. This allowed the
fuel to transfer
its
momentum
to the
Figure 5-1. Preliminary model of a wing section with fuel impacting an exterior column.
94
NIST NCSTAR
1-2B,
WTC Investigation
at 5
ms
is
shown
in
corresponding column damage shown in Figure 5-2(b). The subsequent impact loading of the wing
The
shown in
at the front and back web plates
side plates had initiated. The
plates,
and the
of the front
failure
web were
of the
position of the internal gusset plate at the location of the lower edge of the spandrel above the impact
point.
Although
this analysis
many
of these damage mechanisms were observed in the recovered exterior panels from the impact zone as
shown
front
in Figure 5-3.
web
free
flanges.
The
exterior
weld
until
web with
Figure 5-2. Although the exact loading history that created the observed damage
qualitative
it
is
the internal
shown
in
agreement of the calculated and observed damage modes provided a level of confidence
in the
Effective
Plastic Strain
Fringe Levels
2.500e-01
2.2508-01
2.000e-01
1.750e-01
1.500e-01
1.250e-01
I.OOOe-01
7.500e-02
5.000e-02
2.5008-02
O.OOOe+00
(a)
(b)
Column damage
Figure 5-2. Calculated wing with fuel impact response on the exterior column.
NISTNCSTAR
1-2B,
WTC
Investigation
95
Chapter 5
Source: NIST.
be applied
exterior
to the global
damage
in
the recovered
still
to
WTC
shown
exterior columns.
damage
This empty wing section model was selected to produce significant column damage
failure.
properly.
to
at
The
removed.
an impact speed of
fuel.
series of these
to further refine
modeling methods. These calculations used the updated constitutive models, described
with preliminary failure
criteria.
The
exterior
ksi steel.
fine
in
Chapter
2,
but
model
with a coarser mesh of shell elements were developed. These models included a specific description of
the
different properties. In the fine brick element model, the failure strain for the
base metal, weld metal, and Heat Affected Zone (HAZ) were
until they
showed a
ductility.
all set at
A comparison of parameters
for the
uniform plastic
strain
magnitude
two models
is
damage and
similar failure
Table 5-1.
96
mesh of brick
modes
is
shown
The
in Figure 5-A.
column model on
to obtain a
NISTNCSTAR
The
1-2B,
match of the
matched the
WTC
Investigation
response, were failure strains of 12 percent and 2 percent for the base metal plates and weld zone,
respectively. Contours of resultant displacements are
shown on
the
response was similar in both magnitude and damage mode. The reduction
min
on the
and
in
to 9 min. This
local stress
in
overall
strain concentrations
failure behavior.
Coarse Shell
Model
473,208
352
352
9.192
9,192
1.0 in.
1.0 in.
12%
34%
Number
of Beam Elements
0.0625
64%
Weld Zone
Failure Stram
64
Calculation
Time (CPU)^
in.
min
9 min
624 min
Simulation of 0.035
444 min
Elapsed Time
a.
Model
8
3
%
min
9 min
CPUs.
Fringe Levels
2.000e+001
1.800e+001
_m
1.600e+001_
1.400e+001 _
1.Z00e+001 _
l.OOOe+001
8.000ei000_
6.000e + 0[IO_
4.000e+000_
2.000e + 000_S
O.O0Oe+OOO_W
(a)
(b)
Coarse
shell
element column
An
additional exterior
analysis
was
identical
NISTNCSTAR
1-2B,
column impact analysis was performed using the coarse shell element model. This
to the coarse shell element model shown above in Figure 5-4, but the failure
WTC
Investigation
97
Chapter 5
criteria
were obtained from the mesh refinement study and weld zone
Section 2.3.1 and Section 2.3.3. These were the failure parameters used subsequently in the global impact
analyses described in Chapter
set at
The updated
9.
is
for the
two
different
failure
modes, many of the features, such as extent of weld failure and wing
some guidance
as to the
bounds on uncertainties
in these analyses.
Fringe Levels
2.000e+001 _
1.800e+001 _
1.600e+001 _
l-OOOe+OOl _
1.200e+001 _
1.000e+001 _
8.000c+000_
B.000e+000_
4.0DOe+000_
Z.000e+O00_
O.OOOe+000
(a)
Coarse
shell
element model
(b)
Coarse
shell
element model 2
5.2.2
fuel
in the core
described in this section. Subsequent analyses of engine impacts focused on exterior panels and the
columns
tower structures to
98
resist
and
5.3.
Most
sufficient
NISTNCSTAR
1-2B,
WTC Investigation
for core
columns impacted by
initial finely
fuel-filled
to the exterior
column analyses,
for both
to coarser shell
was
to
element models.
Figure 5-6 (wide flange core column) and Figure 5-8 (box section core column) compare the fine brick
shell
between the
fine brick
kinetic energy
comparisons
shell
factor for the impact behavior in this example, and the primary objective
shell
was
element model could accurately reproduce the impact defoiTnations. The figures indicate that the
response of the coarser shell models was very similar to that of the fine brick models. Therefore, the shell
element formulation and mesh refinement of the coarse model were sufficient to capture the impact
damage mechanisms
(a)
in this
component impact
scenario.
(b)
Coarse
shell
element column
Figure 5-6. Wide flange core column response comparison, showing contours of the
NISTNCSTAR
1-2B.
WTC
Investigation
Chapter 5
Fringe Levels
6.000e+D01
5.4D0et001
J
I
fl.BUOe+BOl _i
4.200e+001 _
3.600e+001 _
3.000et001 _
2.400e+U01 _
I.BOOe+nOI _
1.2n0e+tl01
(a)
Figure 5-8.
(b)
Coarse
shell
6.0D0e000
O.OOOetOOO
element column
of the
100
NISTNCSTAR
1-2B,
WTC Investigation
0.000
0.005
0.010
0.015
0.020
Time
0.025
0.030
0.035
0.000
0 005
0.010
0.015
Time
(s)
0020
0.025
0,030
0.035
(si
Figure 5-9. Displacement and kinetic energy comparison for box core column wing
impact analysis.
Bolted Connection Modeling
5.2.3
There was a wide variety of different connections required for the assembly of the tower models.
Where\ er
from the
in the
model and
failure
Photographic and structural debris evidence clearly demonstrated that the external column connections
played a significant role
in the
mode of column
at exterior
failure
column
failed in several
Source:
1-2B.
WTC
Investigation
in Figure
2002).
FEMA 2002.
NISTNCSTAR
in
shown
(FEMA
in
101
Chapter 5
Source: NIST.
to
impact analyses that accurately captured the capacity and failure modes of the connection. Various
connection component models developed included both fine models of these complex connection
components
bolt
(e.g.,
models and
node and
a surface
failure criterion is
segment
beam element
two nodes or
(shell
exceeded.
LS-DYNA to
A tied
as simple
utilized
many of the
in the fine
beam element
bolt
beam elements
calibrated to
in
match
the fine
Section 2.3.4.
Bolted joints, such as in the spandrel connections, were approximated with a tied node algorithm that
constrained degrees of freedom of adjacent nodes and element faces. Various options were investigated
in these
fidelity
final selections
exterior
column
The
beam elements
beam element
bolt model.
in
102
shown
butt
a preliminary engine
in Section 5.2.4.
The
NISTNCSTAR
1-2B.
profile
was a
WTC Investigation
between the
and a
(a)
(b)
(butt plates
beam
same time
as the
brick element bolts. Failure of the bolts occurred at a time of approximately 3.0 ms. These connection
in the
corresponding brick and shell models of the exterior column component impact
analyses
shown previously
cases, as
shown
same
was
in
in
(a)
the
failure
mode
plates.
is
at
(b)
Beam element
models
for both
at the
models
bolts
joint.
common
rows of
Typical failure of these spandrel joints in the impact zone resulted from
would be
six bolts
shown
on
in the
photographs
in Figure
5-11.
Modeling the
response mechanism was a bolt bearing shear failure. In the exterior columns, the bolts failed primarily
NISTNCSTAR
1-2B.
WTC
Investigation
103
Chapter 5
in tension,
and
of the contact between bolts and butt plates were not as important for capturing the
details
failure behavior.
As
a result, an alternate
made up of shell
shown
elements, as
in
Figure 5-14. Connections corresponding to individual bolts were treated by tying single nodes on the
used
to
make up
Use of a
Two
sometimes
(a) Detail of
connection
(b) Multiple
panel connections
The
first
node
(shown
it
in green in Figure
to
in the
in red in Figure
plate.
equal to one half of the combined thickness of the spandrel and splice plate and was included in the
automatic contact definition. Both material definitions had the same constitutive properties. The
in the
following engine
5.2.4
An example
WTC
on the order of 2
in.).
medium mesh
(i.e.,
in
24
at floor
density
was used
shown
initial
is
96
(the
this
bolted at the free column end, and no boundary conditions were applied to represent the floors. Velocity
104
NISTNCSTAR
1-2B,
WTC
Investigation
mph
(b)
after
shown
in
Figure 5-1
6.
The
plot
shows an
Impact response at 80
ms
NISTNCSTAR
1-2B.
WTC
Investigation
Chapter 5
600
Initial
I
j
mph
Velocity 500
500
a
E
400
mph
3
C
300
re
o
o
200
>
100
0
0.00
0.02
0.01
,1,1
,1
0.04
0.03
,1
0.06
0.05
Time
Figure 5-16. Engine velocity history for the exterior wall impact.
A second engine
impact analysis was performed with similar conditions except that the impact location
in a reduction
is
little
shown
in
exterior wall
shown
this case).
17.
the
damage
in the plot
as seen
The analyses
5-18 shows
differed
merged spandrel
most
between spandrels.
maximum
failure strain
in spandrel failure
strain are
at the
in Figure
column connection
5in
Engine core velocities for the three engine impact analyses are compared
representative engine component.
The
splice plate
model resulted
in a
74
in Figure
mph
5-19
for a single
engine core. The splice model resulted in the largest reduction in the speed of the engine. In this case,
the entire spandrel
was
carried
away by
change
in
splice plate
connection model did not introduce a large computational cost and resulted
106
effect
on the
engine velocity during impact and penetration of the exterior wall. However, the spandrel
in a
more appropriate
NISTNCSTAR
1-2B,
WTC
failure
Investigation
mode
for the spandrels in the impact zone. Therefore, the spandrel splice connections
were maintained
in
(a)
(b)
Figure 5-17. Example engine impact analysis with different impact locations.
(a)
(b)
NISTNCSTAR
1-2B,
WTC
Investigation
107
Chapter 5
520
"T
pi
r T
r~T
f~i
Merged Spandrels:
Spandrel Centered Impact
480
Spliced Spandrels:
x:
Q.
E
460
2^
o
o
<>.
0)
>
440
0)
-&
lU
420
Typic
400
for
33Q
L_i
L_J
000
L_J
0.02
al
_1
L_J
Time
1,1
I,
0,06
0.04
(s)
Figure 5-19. Engine velocity history for the exterior wall impact.
5.3
interior
exterior
columns
to shorten the
in
is
mph
against a set of
shell elements.
exterior
run time necessary for the complete impact scenario. The simulation included
90 ms
560
both above and below. The core column models were several floors
at
tall to
to additional panels
at the ends.
The
calculated impact
damage
is
all
shown
at a
time of
of the columns
with a residual velocity of the engine of approximately 224 mph. The deformations of the column
included large lateral displacements, which would be significantly reduced
of the
concrete floor slab were included in the model. The deceleration profile of the major engine debris
fragment
is
shown
in Figure
5-21. The majority of the engine structure had been broken into fragments
were not
likely to
size, strength,
failure to a
This example illustrates the large level of damage produced by a massive aircraft component, such as an
engine,
at
and, therefore,
made
it
difficult to
to
overwhelm a
single core
column
108
NISTNCSTAR
1-2B,
WTC Investigation
Chapter 5
700
1
j
J
1
Mn
Exterior Panel
Impact
Box Column
Interior
Impact
c
UJ
Interior
200
Wide Flange
Column Impact
100
0
0.00
I
'
0.01
0,02
0.03
Time
0.05
0.04
0.06
0.07
0.08
(s)
Figure 5-21. Example engine impact analysis with exterior and interior columns.
5.4
The
The
failure
in the aircraft
the extent of damage and the spread of fuel and debris through the structure.
The
structures could also be significant in the analysis of the subsequent fire behavior in the towers.
result, floor truss
models used
in the global
As
Component models of a
section of the composite floor assembly outside the core were generated and used
in the
was achieved by
direct impact
by
a simplified impactor.
impact analyses. The models included the concrete slab, the metal decking, and the primary and
bridging trusses. The dampers were not included in the models for the reasons explained in Secfion 3.2.3.
Detailed floor component models used a combination of brick elements for the concrete slab,
beam
elements for the truss round bar diagonals, and shell elements for the remainder of the structures,
including the truss upper and lower chords and metal decking. This model
simplified
shell
is
shown
in
size of the
Figure 5-22.
in
110
NISTNCSTAR
1-2B,
WTC
Investigation
Beam Elements
6,928
3.440
230,778
148,256
39,000
Nodes
372,084
48,971
Total
CPU Time
16.796
(4.7 h)
2,482
(0.7 h)
Elapsed Time
26.553
(7.3 h)
4.454
(1.2 h)
NISTNCSTAR
1-2B,
WTC
Investigation
111
Chapter 5
Combined Concrete
and Metal Decking
Beam &
for
Shell Elements
Truss Structures
The concrete
constitutive
model used
in the brick
previously in Section 2.4. The design for the floor truss structures specified that material yield strength
was
to
be taken equal
constitutive
to
50
truss structure
members
model developed
for the
50
ksi
tower
steel,
all
of the
components and metal decking. The exact strength of the metal decking could not be
confinned; however, the relative contribution of the metal deck was small compared to the
much
thicker
concrete slab (see Figure 5-24). Therefore, the uncertainty on metal deck strength was not expected to
The coarse
floor
model used an
effective material
model
for the
combined concrete
slab
and metal
decking so that these parts would not need to be meshed separately. As the concrete constitutive model
described in Section 2.4 (pseudo-tensor model) was developed for brick elements, and does not work for
shell elements, a
yield strength
piecewise plasticity model was used for the effective slab-decking behavior.
A tabular
unconfined compressive behavior for the concrete. The combined slab and decking stress-strain curve
was compared
combined
to the concrete
floor slab
was dominated by
of the 50
ksi material.
in
low
strain levels
(below
until a strain
percent strain).
was equivalent
to
This combined slab material model had the same rate effect
in
shown
in
Figure 2-32.
112
NISTNCSTAR
1-2B,
WTC Investigation
'
'
'
0<!^
0.000
0.005
0.010
0 015
0.020
I_J
l_l
0 030
0.025
Strain
Figure 5-24. Constitutive behavior for the combined concrete and metal decking.
in the
impact.
to the
similar to the
response to an aircraft engine, but without the complications of the detailed engine impact behavior,
mesh
floor
model
is
shown
is
in Figure 5-25.
shown
in
in
An example
analysis with a
the
calculated plow impactor deceleration profile for the brick and shell element models are
compared
in
Figure 5-27. The comparison shows that the simplifications did not significantly change the impact loads
in the floor assembly.
NIST NCSTAR
1-2B.
WTC
Investigation
13
Chapter 5
NISTNCSTAR
114
1-2B.
WTC
Investigation
(a) Initial
(b)
configuration
Figure 5-26. Floor assembly impact analysis with shell element concrete slab.
NISTNCSTAR
1-2B.
WTC
Investigation
115
Chapter 5
12x10
10
-I
a
S
5 !i
I
O
O
<
'ft
0.00
*'|-J
1_
0.05
-1
0.10
0.15
0.20
Time
Figure 5-27. Comparison of brick and shell element floor assembly impact analyses.
5.5
The
aircraft
component models
to
be developed.
Having established acceptable tower component models, the exterior panel model type 300 was then used
to evaluate
an aircraft wing section component model. Construction of the wing section model was
failure
and
failure criteria
modeling
criteria are
modeling methodology
developed
in this section
is
model of
were implemented
in the
Element
model
size, tied
airframe.
5.5.1
A significant challenge in developing an aircraft model suitable for perfoiining the global analyses
reduce the number of nodes and elements in the model to a manageable
size.
was
to
because the detailed geometry of the fracture and tearing of aluminum that occuiTcd during impact could
in the
coarse
116
WTC tower.
NISTNCSTAR
1-2B,
WTC Investigation
Development of the w ing section component model was previously described in Section 4.3.2. The
meshed wing component model had a mesh fine enough to capture the tearing of the wing. In this
analysis, the fine
finely
modeling of the panels and a reasonable runtime appropriate for the uncertainty analyses. The
resuhing model for the exterior panel had 10,400 shell elements (columns and spandrels), 264 solid brick
failure
beam elements
model
is
shown with
Figure 5-28.
The wing
impacted with a
way between
in
its
The
wing
exterior panels.
This was done to impart a less symmetric loading on the exterior columns, which was more
of the
the
fine
aircraft
in.
or
roll,
realistic for
for a total
model
size of approximately
245,000
shell elements.
Material properties and constitutive models discussed in Section 2.5 were used in the component model.
of the 0.5
NISTNCSTAR
in.
1-2B,
(MIL-HDBK-5F
WTC Investigation
at failure
failure
element met
117
Chapter 5
was approximately
7 hours
on
2 processors.
Fine
The
debris
118
was
mesh used
section impacting
spars, ribs,
significant material
Wing Mesh
Figure 5-3 1
initial
component model.
free to pass
two
in
initial kinetic
energy.
Damage
result,
is
shown
NISTNCSTAR
1-2B.
in Figure 5-32.
WTC
Investigation
Figure 5-31. Calculated impact of a fine-mesh empty wing section with two exterior
panels at 442 mph.
NISTNCSTAR
1-2B.
WTC
Investigation
119
Chapter 5
(a)
(b)
Back
4 percent of the
fracture of the
damage
initial kinetic
welded
wing
meet the
structure
energy.
and
be an order of magnitude larger than that for the fmely-meshed wing section.
still
to
size constraints
needed
damage
to the exterior
web
front
large-scale
(building interior)
in. to 5 in.
shown
is
in
The
total
model
wing section
When
the coarse
the
wing
The
structures
absorbed
initial kinetic
in the
energy.
Much
in the panels.
in
material
to
8 percent of the
of the additional kinetic energy loss for the coarse wing section model was
The damage
to the
initial kinetic
internal energy
energy.
Additional wing section-panel impact analyses were used to determine failure models for the coarse
model
in order to
to
analysis).
the
damage
mesh
in the calculation.
to the exterior
120
competing mechanisms
mesh
Reducing the
in the
wing section
fine
mesh
analysis).
NISTNCSTAR
1-2B,
The
final
WTC Investigation
debris.
was
to
found
shown
needed
The
to
initial
It
was
approach,
2.3.1).
rn
|
wing model
to the panels,
damage
80
Element Erosion
Element Erosion
Mesh
sheet
Fine
sheet
Coarse Mesh
20
L_i
0.00
0 10
0.05
020
0.15
L.
0.25
0.30
Plastic Strain
model.
"1
r-]
r r ri
I
I
I
80
rz
Element Erosion
...A--
i5
60
(A
(A
-*- 2024-T3
in
clad sheet
Mesh
Fine
Coarse Mesh
40
o
zs
20
Element Erosion
_^
r
0
0.00
0,10
0.05
L_J
0.15
0.20
0.25
0.30
Plastic Strain
(b)
Figure 5-33. Failure modeling approaches for the coarse wing section model.
NISTNCSTAR
1-2B.
WTC
Investigation
121
Chapter 5
was applied
that
columns prior
was
to
that elements
being eroded.
first
method
at
is
shown
in
eroded
An
for treating
to
strain,
degraded tme
transfen-ed to the
material failure.
value of plastic
given
at a
An example
fine
mesh
Beyond
failure strain.
this strain, a
of 7 ksi was used. This level was selected to balance the need for a reduced strength
to that
of the
initial failure
model,
Figure 5-33(a).
Wing
sections
momentum
of these elements
to the
columns prior
failure approach,
shown
to element erosion.
in Figure 5-34(b),
Element
were much
larger.
impact
this
(a)
(b)
for
two material
failure criteria.
A final parameter investigated in the coarse wing section component analyses was the modeling of wing
structure connectivity
(i.e.,
tied interface
without failure, between the internal structure and skin) and with a tiebreak interface. The surface-
to-surface tiebreak and surface-to-surface tied contact algorithms were discussed in Section 3.3.2.
found
elements.
122
When
between the
a tied contact
internal structure
was used,
the
It
was
NISTNCSTAR
1-2B,
WTC Investigation
fine
mesh most
(a)
Fine
tied
mesh
to that
wing..
tied
Coarse mesh-
(b)
(b)
in
meshed
is
wing
shown
mesh
section
fragmentation.
model
is
shown
in Figure 5-37.
The
in
Figure 5-32. The tabulated true stress-plastic strain data used in the final model for the various aircraft
aluminum
alloys are
shown
in
Figure 5-36.
model did
a reasonable job
challenge of simulating the tearing and fragmentation of aluminum with such large elements. In both
damage extended over approximately one floor, and the columns were still intact. The front
plate was pushed inward, but did not fail. The side flanges were also bent inward, but did not fail. The
back plate sustained minor damage. Most of the difference in damage appeared to be in the side flanges,
resulting in larger internal energy absorption in the panel. The larger elements of the wing section and
their limited ability to conform well with the column geometry resulted in more damage to these flanges.
cases the
Despite the coarse wing section having a lower residual kinetic energy, the wing broke up in a similar
NISTNCSTAR
1-2B,
WTC
Investigation
123
Chapter 5
meshed model,
as
shown by
therefore, less
mass and
had
100
Oi
(A
(A
O
o
Element Erosion
1
0.05
0.00
0.10
L.
0.20
0.15
Plastic Strain
Figure 5-36.
Aluminum model
damage
analysis.
coarse-mesh
failure
As
aircraft
components.
will
aircraft
components
9,
the
pattern predicted in the global analyses agreed reasonably well with that observed in
photographs. This
is
even true
5.5.2
in the
outboard parts of the wings where there was no fuel and the exterior
fail.
it
is
was responsible
significant effect
interaction
is
767 wing
is
from the
fuel in
its
estimated that each aircraft had approximately 10,000 gal of fuel onboard.
for large distributed loads
on the damage
inflicted
it
on the building
structure.
it
At the
Upon
WTC towers
could have a
Modeling of the
fluid-structure
necessary to predict the extent of this damage and the fuel dispersion within the building to
124
for
initial
fire
dynamics modeling.
NISTNCSTAR
1-2B,
WTC Investigation
= 0.0 s
= 0.02 s
= 0.03 s
0.01 s
Mill
= 0.04 s
Figure 5-37. Calculated Impact of a coarse-mesh empty wing section with two exterior
panels at 442 mph.
NISTNCSTAR
1-2B,
WTC Investigation
125
Chapter 5
(a)
(b)
Back
(building interior)
Figure 5-38. Impact damage to the exterior panel from a coarse-mesh empty wing
section at 442 mph contours of effective plastic strain are shown.
A number of approaches to solving Fluid-Structure Interaction (FSI) problems are available in LSDYNA. One approach is the standard Lagrangian finite element analysis with erosion, where the fuel
modeled using
is
a deformable mesh. This approach accounts for the inertial effects of the fuel, but does not
mesh
distortion.
The
Arbitrary-
Lagrangian-Eulerian (ALE) method has been developed as one good approach to solve fluid and solid
material interaction. With this methodology, fluids are
materials are
modeled with
for
interact.
An alternative approach is to use mesh-free methods such as Smoothed Particle Hydrodynamics (SPH).
SPH modeling for fuel effects has the advantage of a smaller mesh size and, potentially, much faster run
times than ALE analyses. Both ALE and SPH methods were applied to the analysis of fuel impact and
dispersion and are compared here.
5.5.3
for
small wing segment was from rib 14 to rib 18, the outboard baffle
considered to be completely
filled
SPH
Figure
shell
126
fuel particles
elements for
this
and
10,825
section
rib.
gal).
For
fuel.
The
segment was
ALE mesh for the fuel, shown in blue. The fuel was modeled with
ALE elements for the fuel and surrounding air region, shown in
model was
NISTNCSTAR
1-2B,
WTC
Investigation
(a)SPHmesh
(b)
Figure 5-39.
ALE mesh
wing segment.
For these component analyses, the impacted structures were the same two exterior panels used
empt\'
Section 5.5.1.
in
changed
was
Model
beam elements
of water
is
62.3 lb
ft
'
where
size
model the
to
fuel,
although
in the
in
is
was
it
later
approximately 48.3 lb
ft"\
Solid
ALE
Elements
Mass
Figure 540.
into.
In
initially
ALE
for the
An ALE mesh
"empty" regions.
2.1 12
110.825
is
w as
that of a
pitch,
mesh was
ALE
lb
yaw, or
roll.
shown
is
in
and
at
500
mph
at
mesh
needed for
is
air to interact
with stationary
filled
normal impact
shown
in Figure
fail
1.014
trajector}'
23.733
surrounding the wing segment and the panels was needed for the fuel to flow
analyses, material
54.096
16.854 lb
Component
^^ ing
at a
exterior panels
For comparison,
5-38, but
it
is
important
specifically discussed
still
the columns.
NISTNCSTAR
1-2B.
WTC
Investigation
127
Chapter 5
ALE
Exterior panel
of
air
region
WTC tower
Figure 5-40.
Wing segment,
fuel,
V=
500
mph
(223 mis)
128
NISTNCSTAR
1-2B,
WTC
Investigation
ALE
and
SPH models
using the
MAT_NULL material
Gruneisen equation of state model. The pressure in the Gruneisen equation of state
is
model and
estimated as
P=
for
(1)
(2)
complete description of these models. These material models were used in order
constituti\ e beha\ ior of the fuel,
momentum
is
However,
is
the
an exchange of
air
model parameters.
Fuel (Water)
Air
62.3
0.0737
1.82E-05
3.66E-07
MATNULL
Density
p^
(lb ft')
EOSGRUNEISEN
C
6.48E-04
(in s)
s,
1.921
S:
-0.096
S?
0.0
0.35
To
Eo
2.14e+05
(ft-lb)
0.172
C. (Btu Ib F)
0.240
To(R)
1.028
Cp (Bni
Vo
1.0
Fluid-structure interaction
SPH. The
algorithm.
first,
when
using an
The second,
NISTNCSTAR
1-2B.
WTC
for an
ALE mesh
SPH
fuel
Investigation
for fuel,
mesh,
is
the
The
129
Chapter 5
reader
SPH
is
referred to the
LS-DYNA user's
manual
for a
the
ALE
and
was not
ALE
and
SPH methods
5^8 show
shown
in Figure
are
5^2
initial
5-44. Figure
to impact.
impact speed.
5^5
through
close-up plan and side view comparisons of the fuel dispersion and wing break up
predicted by the two fuel modeling methods. While both modeling approaches gave comparable results
for the
damage
and wing break up than when using ALE, as can be shown clearly
data,
it
The
in the side
difficult to evaluate
is
cuiTcnt
SPH model
that the
solution.
SPH method
is
more
ALE
method
as
it
mesh and does not need to rezone after each time step, as is done in the ALE method.
ALE method requires a mesh for both the fuel region and the air zone into which the fuel
requires a smaller
In addition, the
ALE
analysis.
SPH
that the
ALE
much
as 10 times
SUMMARY
5.6
The
set
mph
in this chapter
demonstrates
how
an engine traveling
at
analyses are not as useful as desired in deteraiining the sensitivity of modeling parameters in the impact
response.
of twenty
interior core
The
8.
WTC towers and Boeing 767 aircraft are extremely complex structural systems, and including all of
the overall
130
is
and towers
model
to
would
result in very
damage.
NISTNCSTAR
1-2B.
WTC Investigation
NISTNCSTAR
1-2B,
WTC
Investigation
131
Chapter 5
= 0.04 s
Figure 5-43. Calculated Impact of a coarse mesh wing section laden with fuel modeled
using SPH particles.
132
NISTNCSTAR
1-2B,
WTC
Investigation
NISTNCSTAR
1-2B.
WTC
Investigation
133
134
NISTNCSTAR
1-2B,
WTC Investigation
Time =
0.039999
ALE
Figure 5-48.
A series
analysis of structural
damage and
of component impact analyses were performed using models of tower exterior columns, core
columns, and floor assemblies with wing section and engine models as impactors. These models were
used
to
develop the simulation techniques required for the global analysis of the aircraft impacts into the
WTC towers.
The following
500
mph
results
exterior wall
and
failure
in a penetration
of the
majority of the engine core remained intact through the exterior wall penetration
slab, the
with a reduction in velocity between 10 percent and 20 percent. The residual velocity and
after penetration
fail
a core
column
in
normal impact of the exterior wall by an empty wing segment outboard of rib 14 of the
failure.
This
is
significant
damage
to the exterior
showing the
exterior
damage
to the
A fiiel-filled wing section impact resulted in complete failure of the exterior columns.
This
initial
of the
momentum
is
prior to impact.
Three different numerical techniques were investigated for modeling impact effects and
dispersion of fuel: (1) standard Lagrangian finite element analysis with erosion, (2)
Particle
Hydrodynamics (SPH)
analysis.
Of these
approaches,
analysis,
SPH
and
(3)
Smoothed
Arbitrary-Lagrangian-Eulerian (ALE)
NISTNCSTAR
1-2B,
WTC
Investigation
135
Chapter 5
REFERENCES
5.7
Hallquist, J.O.,
2002, World Trade Center Building Performance Study: Data Collection, Preliminaiy
"DYNA3D
UCID- 19156,
Revision
"LS-DYNA Keyword
2,
986.
April 2003.
"LS-DYNA
136
XYZ
May
1998.
September 2001.
NISTNCSTAR
1-2B,
WTC
Investigation
Chapter 6
INTRODUCTION
6.1
are a transition
9.
in
Chapter 5 and
two primary
into
sets:
preliminary and final subassembly analyses. The preliminary subassembly analyses were used mainly to
investigate different
size,
stability,
and
impact response. These preliminary subassembly analyses were perfoiTned concurrently with the
in
Chapter
Therefore,
5.
many components
in the preliminary
subassembly model were not those ultimately used in the global impact models. As a
constitutix e properties, failure criteria,
The
final
is
in
Chapter
9.
The
constitutive
of these
to
The
final
to
subassembly model
8.
6.2
A preliminary
tower structures
in the
shown
in
Figure 6-1
v\
slab
detailed
model of the
(WTC)
1.
the
for this
in the
in the
model with
as
floor
model contained
many
An example
later.
subassembly analyses were performed using the same modeling methodologies as the global
slab
some of the
those used subsequently in the impact zone of the global impact analyses.
was used
result,
model used
floor
94
members
were
shell elements.
The
final
approximately 793,000 nodes, 404,000 brick elements, 385,000 shell elements, and
model contained
1,000
beam
elements.
NISTNCSTAR
1-2B,
WTC
Investigation
137
Chapter 6
138
NISTNCSTAR
1-2B,
WTC Investigation
The impact response of the tower subassembly when impacted by an engine at 500 mph is shown in
Figure 6-1 (b) and Figure 6-2. Engine impact was centered on the spandrel at floor 96. The trajectoi-y of
the engine was normal to the exterior panel. The engine significantly penetrated the floor structure and
caused significant damage to the core column (503) upon impact.
The impact response of this preliminary subassembly model was influenced by many of the preliminaiy
modeling assumptions applied. Unconstrained boundary conditions on the side of the floor slab
influenced the deformation of the truss floor structures. In addition, the concrete in this analysis
assumed
to ha\ e an
included
in the
The
result
rate
was
still
were useful
to
in
developing simplified models for the bolted column connections, as described in Section 5.2.3, and to
refine the
FINAL
6.3
The
final
final
SUBASSEMBLY ANALYSES
subassembly model was developed using structural components from the impact zone on the
north face of
WTC
(Side 100).
in
response of structural connections, material and failure models, and other issues affecting global impact
analyses.
floors
tall,
first
included in the subassembly model were column numbers 502-504 and 602-604. The exterior wall in the
NISTNCSTAR
1-2B,
WTC
Investigation
all
that
column numbers
in the
139
Chapter 6
column
17 through 125.
The
structural
components
in the final
panels, core framing, truss floor structures, and interior contents (workstations). These structural
components
are
shown
The concrete
in this
model.
Ext. Panel
Numbers
\
118
140
NISTNCSTAR
1-2B,
WTC
Investigation
(a)
(b)
(c)
Workstations
in
1-2B,
WTC
Investigation
NISTNCSTAR
in
141
Chapter 6
The
vertical displacements
were constrained
at
the top
lateral
to
at the free
realistic
at
in Section 6.2.
model reduced
greater width of the exterior wall and truss floor stnactures in this
In addition, the
6.3.1
&
the Pratt
Whitney
in the
PW4000
The
was
to investigate the
was described
impact response to a
A few
analyses are described here to illustrate the characteristic engine impact behavior and the effects of
specific
The
model parameters on
initial
initial
mph and
velocity of 413
shown
is
in
of 4 degrees from the exterior panel normal and a vertical approach angle of 7.6 degrees below the
horizontal.
ft
below
impact did not engage a significant portion of the truss floor structures. The calculated impact response
of the subassembly
is
shown
subassembly model,
it
plowed through
in Figure 6-6.
at
impact behavior
downward
trajectory.
and continued
into
and eventually
at different
downward
As
times
is
shown
trajectory.
its
in Figure 6-7.
the
downward it impacted
The engine motion was redirected by the
As
At a time of 0.25
is
shown
in
primarily from the penetration of the exterior wall and the floor slab and truss of floor 97. This
deceleration
was
in
Section 5.2.4 was that of only exterior panels without any floors attached or boundary conditions to
approximate their
structure
142
effect.
This difference
in
deceleration
was due
of the floor
structure.
NISTNCSTAR
1-2B.
WTC Investigation
Figure 6-6. Calculated response for the baseline engine impact analysis.
NISTNCSTAR
1-2B.
WTC Investigation
143
Chapter 6
r"'i
ir
1
(a)
Time = 0.00 s
(b)
Time = 0.05 s
(c)
Time = 0.25 s
Figure 6-7. Baseline engine impact and breal< up behavior (side view).
144
NISTNCSTAR
1-2B,
WTC Investigation
500
400
\
c
UJ
100
0 '
0.00
0.05
0.10
0.15
Time
0.25
0.20
0.30
(s)
Figure 6-8. Speed history for the baseline engine subassembly impact analysis.
Building Contents Strength Effects
Two
As
calculation used identical impact conditions, but the building contents had virtually
The deceleration
Figure 6-9.
The deceleration
ms was
exterior wall. In this phase of the impact response, the engine speed
of 413
was
mph
to
are
no strength
compared
(0.5 psi).
in
approximately 375 mph. The majority of the subsequent deceleration to a time of 0.25
mph
of the strength of
30
mph
1
500
total loss
the
was
low strength
NISTNCSTAR
1-2B.
WTC
Investigation
145
Chapter 6
500
400
Low
Strength
300
o
o
o
>
Core Column
Impact
200
UJ
100
0.00
0.05
0,10
1
i
0.15
Time
0 25
0.20
L.
0.30
(s)
Figure 6-9. Effect of the building content strength on engine impact behavior.
A pair of subassembly
engine impact analyses was performed to investigate the influence of the floor
shown
was
in Figure 6-10.
in
and
(c)
significant interaction of the impacting engine with the truss floor structures, as
on the
initial
at
a time of 0.25 s
core structures.
ksi.
However, due
to aging
was
was used
in the
at the
actual strength
was approximately 4
ksi
(NIST
To demonstrate
NCSTAR
1-6).
This
the specified 3 ksi strength to demonstrate the influence of the concrete strength on the impact behavior.
146
NIST NCSTAR
1-2B,
WTC
Investigation
(a)
Time = 0.00
-a
(b)
Time = 0.05 s
(c)
Time = 0.25 s
Figure 6-10. Engine impact for concrete strength evaluation (side view).
NISTNCSTAR
1-2B,
WTC Investigation
147
Chapter 6
1.
was introduced by
a 5
identical speeds at
mph
ms when
compared
in
The speed
trajectory, a larger
and 4
a slightly different
When the
50
the engine
rapidly.
A comparison of the engine impact damage to the truss floor structures with the
concrete
is
shown
impact damage.
in
and 4
ksi strength
hole was created in both truss floors that extended from the outer wall back to
truss.
148
3 ksi
Figure 6-12. The comparison shows that the two tmss floors had very similar
NISTNCSTAR
1-2B,
WTC
Investigation
Time=
Fringe Levels
0.2
2.000e-02 ___
1.800e-02
1.BOOe-02_a
1.400e-02_
1.200e-02_
1.000e-02_
B.OOOe-03 _
6.000e-03
4.000e-03
2.000e-03
O.OOOe + 00
(a)
damage
damage
(b) 3 ksi
damage from
is
this study.
NISTNCSTAR
1-2B.
WTC
Investigation
to
be 4 ksi for
149
Chapter 6
6.3.2
Fuel-filled
model
discussed in Section 6.3.1. The development of the wing section impactor was described previously in
Chapter
is
4.
shown
in
Figure 6-13. The fuel-structure interaction in this analysis was modeled using
SPH
particles,
Figure 6-13. Final subassembly model for wing section impact analysis.
first
is
shown
in
Figure 6-14.
50 ms, the wing section penetrated the exterior wall and was torn apart by the interaction with
20
The
ft
to
aircraft fuel
30
150
ft
a relatively
significant
to the core
formed
damage
At a time of 0.25
to the
s,
full
height of the
of the subassembly.
NISTNCSTAR
1-2B.
WTC
Investigation
r
1
(a)
Time = 0.00 s
(b)
Time = 0.05 s
(c)
Time = 0.25 s
Figure 6-14. Baseline wing section impact and break up behavior (side view).
the
shown
0"^
Ib
ft.
in
in
Figure 6-14,
is
different in
Figure 6-7. Both impactors started with similar kinetic energies; the
initial kinetic
energy of 5.0x10^
Ib
ft
the workstations,
a path
through the center of the subassembly. The wing caused extensive floor buckling despite having no
initial vertical
component
NISTNCSTAR
1-2B,
WTC
to
its
Investigation
section,
damage
151
Chapter 6
due
is
to the
more than
little
to
no damage
The
to the core.
break up and expansion of the wing components and fuel cloud. This
the
more
The difference
in the
impact behavior for the engine and wing section can be shown by the comparison of
momentum
momentum
momentum
resulted in greater
damage near
shown
histories,
to the
in Figure 6-15.
initially
tower
in a
much
However,
wing section
comparable impact
for
mass, the engine had a significantly larger energy entering the core and corresponding potential for
damaging
.2
[T
1
i
1
1
VZ
1-
Bc
1.0
o
E
o
Engine Impactor
ro
'
a.
0.6
3
o
V)
on
T3
0.4
i
o
0.2
z
QQ
C__l
0.00
0.05
Time
Figure 6-15. Impactor
Weld Zone
momentum
0.15
0.10
0.30
0.25
0.20
(s)
Ductility Effects
column weld
toughness on the impact response. This weld existed between the four plates comprising the column
cross sections.
As
described in Chapter 2, the ductility of the weld zones was developed based on
experimental characterization of the steel behavior, as well as simulation of the weld fractures in column
was
maximum
set at 8 percent.
elevated strain rates and under the very high stress concentrations
152
at the
crack
tip
fractures at
can result in a
NISTNCSTAR
1-2B,
WTC Investigation
from
Two wing
first
of a more
brittle
weld
ductility
Of the
for
due
to
column welds.
brittle
weld behavior
is
shown
in Figure 6-1 6.
similar with only subtle differences in the fragmentation of the failed columns.
calculated global energy balance and the absorption of energy by the exterior panels are
Figure 6-17.
was used
crack extension. This low ductility would more accurately simulate the
Damage
to
section impact analyses were performed to investigate the influence of the weld ductility on the
effects
was of interest
it
approximately 8.9x10^
Ib-ft
shown
The
in
was absorbed
in the
in the
two
analyses was nearly identical, indicating that weld ductility had a neghgible effect on the impact behavior
and damage
to the tower.
(a)
Baseline response
NISTNCSTAR
1-2B.
WTC
Investigation
(b) Brittle
damage
for the
weld response
153
Chapter 6
60x10
c
UJ
JO.
0.00
0 15
Q-10
0,05
Time
0.25
0.30
(a)
12x10
0.20
(s)
10
c
iU
m
c
c
Weld Strength
Brittle
Effects:
Welds {s,=1%)
Welds {c,=8%)
Ductile
3
iU
0.10
Time
(b)
!)
0.15
'
0.20
0.25
0.30
(s)
Figure 6-17. Effect of the weld strength on wing section impact response.
154
NISTNCSTAR
1-2B,
WTC
Investigation
SUMMARY
6.4
a transition
analyses,
more complex
level analyses
The subassembly analyses were primarily used to investigate different modeling techniques and
associated model size, run times, numerical stability, and impact response. The final subassembly model
analyses.
used structural components from the impact zone on the north face of WTC
in the final
The
components
structural
and
The subassembly model was impacted by an aircraft engine and by a segment of a fuel-filled wing. The
subassembly model was used to investigate the effect of a number of modeling parameters on the
response and damage estimates. For the engine impact simulations, these parameters included the
strength of both the building nonstnactural contents and the concrete slab. For the
simulations, the effect of the ductility of the exterior
investigated.
The deceleration
was
was dominated by
the
rather than
their strength.
Varv'ing the strength of the floor concrete slab from 4 ksi to 3 ksi did not result in significant
change
effect
wing impact
by
column weldment on
in the
impact response.
It
Varying the
ductility
in the exterior
weld
damage
ductility
columns from
8 percent to
percent did
had a negligible
effect
on the impact
response.
6.5
REFERENCES
"LS-DYNA
NISTNCSTAR
1-2B,
WTC
Investigation
1,
in
1997.
155
Chapter 6
156
left
blank.
NIST NCSTAR
1-2B.
WTC Investigation
Chapter 7
INTRODUCTION
7.1
Trade Center
initial
included impact speed, aircraft orientation and trajectory, and location of aircraft nose
\'ideos captured the
World
hit the
flight
1 1
(AA
at
Two
impact.
WTC tower, and several videos captured the United Airlines flight 175 (UAL 175) aircraft
that impacted the WTC 2 tower. In addition, there is a large body of photographic evidence that was used
impacted the
were analyzed
The
and orientation
and
roll
was performed
in
two
The
steps.
first
step
was
to
perform an analysis of the video footage of the two impact events. This analysis compared the various
videos and used visual references and
The second
step
was
impact damage to refine the details of the impact position, orientation, and trajectory (Section
impact orientation and trajectory parameters are defined
Two vectors
at the
Figure 7-1.
in
is
no reason
to
assume
that these
The
shown
enough resolution
orientation
is
measured clockwise
ft.
was not
In the
sufficient to
two videos
WTC
AA
north side of WTC
angle of
is
in the figure.
roll
1 1
Since
UAL
still
WTC
and
to the
of the damaged
and one
coincident,
angle around structure east as shown in the figure, and a lateral angle, which
angle, as
The
7.3).
still
frame photography
2 structure again was used as the primary source to determine the impact location.
impact conditions.
NISTNCSTAR
1-2B,
WTC
Investigation
157
Chapter 7
Tower
Axis
7.2
7.2.1
Videos Used
The
first
in
the Analysis
An
selection of appropriate
WTC tower
impacts has been collected by the National Institute of Standards and Technology (NIST). The available
videos
in the
NIST
WTC collection were reviewed to select the best video footage of the aircraft's
approach and impact with each tower. NIST had already digitized the footage, which was stored
with
AVI
analysis.
foiTTiats.
the towers
aircraft
were measured
in
in
Still
in files
in the
wing
tips, aileron,
each frame of the videos. Adobe Photoshop was used to determine the
conditions.
158
WTC
the analysis.
and
The
to analyze
impact
NISTNCSTAR
1-2B.
WTC Investigation
Format
\'ideo
WTC
NTSC
VI
Impact
Description
Footage taken
Lispenard
WTC
PAL
\'2
at
ground level
streets.
at the
NTSC
\'4
NTSC
WTC 2
WTC 2
at
NTSC
V6
NTSC
WTC 2
WTC 2
towers.
WTC 2
NTSC
\'7
Footage taken
Libert)'.
\'8
NTSC
V9
NTSC
WTC 2
WTC 2
FDR drive,
heading
west just before the Brooklyn Bridge. Footage taken north and
east of the towers.
in
Table 7-1
lists
is
Any image
digitized
is
common
NTSC
It
in
was
PAL
assumed
that the
images per second. The digitized images had sizes of 720x480 pixels (NTSC) and 720x576 pixels
(PAL). The original video footage was assumed to have an aspect
to
ratio
ratio.
to
7.2.2
discussed
in the
aircraft.
quality and limited video footage available produced greater uncertainty using the
following
is
as
following sections, provided more accurate estimates of speed and orientation. The
a discussion of the
complex motion
The
in this analysis to
determine the
aircraft
developed for other applications (Cilke 1995). Figure 7-2 depicts the analysis procedure. The image
coordinates of the moving object (the aircraft) and two stationary positions on the structures within the
field
NISTNCSTAR
to
known
i.e., all
1-2B,
WTC
Investigation
The
The
was
159
Chapter 7
extending from the camera to the measured object. Note that the position of the object along the vector
camera
to the
test conditions,
field
of view
is
by a
was
measured object
The global
in
result
set
the real
object.
Note
that in ideal
camera
percent to
1.5 percent.
Step
known.
Step
No.
2:
No. 2
second camera.
Figure 7-2. Procedures to measure object motions using multiple cameras.
160
NISTNCSTAR
1-2B,
WTC Investigation
For the
WTC
locations.
aircraft
motion analysis, various locations on the two towers were used as fixed reference
WTC
on the
Figure 7-3 shows the reference locations used on the towers. The tower reference positions were at the
center of each beveled comer. Similarly, Figure 7-4
is
shows
by using
to
is
fidelity, the
While the
NISTNCSTAR
cameras had
a plan
is
1-2B.
WTC
Investigation
WTC
161
Chapter 7
A2,A6,A10
A1,A5,A9
True
North
WTC
Structure
North
yA3,A7,A11
A4,A8,A12
B2,B 6,B10
B1,B5,BV
WTC
B3,B7,B11
B4,B8,B12
WTC
towers.
With the camera locations estimated, motion analyses were performed using the complex motion analysis
technique. For the
WTC
aircraft impact,
from the V2 video. The analysis produced a speed of 435 30 mph for the
WTC
the data
time of
aircraft at
impact.
For the
V5, and
V9
other two cameras to determine the motion of the aircraft prior to impact.
effective with this analysis technique.
Videos
V6 and V7 viewed
less
few frames,
which could not be intersected with many frames from the other cameras. Videos V3 and V8 were both
taken from the north. The
As
towers.
the
two
a result, the cameras' data did not intersect with the other cameras' fields of view. Data
helicopters,
were too
last
V3 and V8,
from
similar.
later in
to a tight set
aircraft
derived from three components. First, there was a significant amount of scatter in the measured image
coordinates.
to the
low resolution of the images and the motion of the camera fields of view. The scatter in the
image data contributed to approximately 2 to 4 of the image uncertainty. The tips of the aircraft
relatively
162
NISTNCSTAR
1-2B,
WTC
Investigation
to
an estimated
uncertainty of 4 in the roll angle. Second, the uncertainties in the camera locations contributed to the
uncertainties in measixred angles. Since there
in the
than the vertical positions, the measured horizontal azimuths had larger uncertainties (2). Third, the
potential distortion in the field of
The
to
initial
that the
aircraft
V4
footage.
UAL
traced to three possible causes. First, the range of the camera could only be estimated. If the camera
close to the object motion, the range of the
of the object
in the field
in
was
left to right,
in the
tracking
camera
UAL
fields
175 as
it
edge appeared
(Section 7.2.3).
of view. As a
result, the
to decrease
V5
initially
appears from
to the center.
result, a simplified
field
measured image distances increased. Third, and most important, there were
measurable distortions
east
a significant effect
motion. Second, the scale of the image was determined from the dimensions of the towers
uncertainties in the
from
to
in the field
of view. As a
WTC 2
aircraft
and
trajectory vectors.
7.2.3
of view with the aircraft's apparent fuselage length. Figure 7-5 depicts the
simplified procedure to determine the aircraft speed. For several videos of the
of the nose.
tail,
and wing
tips
the nose
and
were measured. The apparent length of the fuselage within each image
and
tail
to the
length of the aircraft determined the aircraft speed (there are constant time steps between frames).
Finally, a geometric correction
was made
if the
Image 3
Image 4
Speed
(U,*L,)/2
NISTNCSTAR
1-2B.
WTC
Investigation
163
Chapter 7
adjustment
As
in
aircraft is
was assumed
to
155
ft
as
shown
in
ft.
The
apparent fuselage length was a result of the relatively low resolution of the video footage.
by a single
pixel, the
value until the object has entered by a significant fraction. The low resolution could not accurately
capture the shape of the aircraft nose and
tail,
and the
aircraft
than the actual Boeing 767-200ER nose. The average length of the fuselage
the fuselage).
As
It
in the
was assumed
(1 ft at
each end of
was approximately
ft
Used
with permission.
164
NISTNCSTAR
1-2B.
WTC
Investigation
The
WTC
was used
to
this
The second video, V2, could not be used to obtain an accurate measure of speed as the aircraft
was tra\ eling away from the camera. The simplified analysis produced a speed of 451 mph 30 mph,
which is 16 mph higher than the value obtained from the complex motion analysis technique. Both of
technique.
WTC
WTC
impact speed
fall
As
aircraft
is
The
UAL
scatter in the
for
UAL
175
measured
displacements, the aircraft length within the image, and uncertainty in the actual aircraft length as seen in
unknown
orientation.
introduced due to the lateral fuselage orientation relative to trajectory. The uncertainty in this value was
due to the
aircraft
3 degrees
in speed,
an uncertainty of
was assumed.
in orientation
UAL
\'ideo Reference
V4
573
mph
V5
556
mph + 27 mph
\'6
535
mph
23
mph
V7
523
mph
31
mph
\'Q
557
mph
53
mph
542
mph 24 mph
55
mph
weighted average of the mean values. The measurement precision (the reciprocal of the variance) was
used as a weight factor on the mean values.
mean could be
for
AAl
actual
calculated by
mph =
and 542
bound on
The
for
AA
is
v\
for
1 1
and
shown
in
UAL
is
16
away from
mph
WTC
175, respectively.
was increased
to
Table 7-3.
was used
The
to obtain
to
this
aircraft
mph 30 mph,
higher than the value obtained from the complex motion analysis technique. Both of
the camera.
in the
175.
traveling
which
WTC
the individual
UAL
technique.
was
summing
mph
the uncertainty
14
If
WTC
aircraft
impact speed
fall
in
Table 7-3
is
As
NISTNCSTAR
1-2B.
WTC
Investigation
165
Chapter 7
Summary
able 7-3.
measured
of
aircraft
AA
Impact Speed (niph)
Vertical
(WTC
UAL
1)
Approach Angle
3 below
10.6
horizontal
Approach Angle
80.3
4 clockwise from
4 below horizontal
(heading downward)
Structure North
6 clockwise from
Structure North"
25 4
wing downward)
is
6 clockwise from
Structure North"
a.
horizontal
from horizontal
(left
4 below
19
Structure North"
Roll Angle
2)
(heading downward)
(heading downward)
(Velociry vector)
(WTC
175
542 24
443 30
(Velocity vector)
Lateral
11
38 4
UAL
UAL
impact speed. The results of the sirnplified motion analyses from each camera for
The
uncertainties in
175
displacements, the aircraft length within the image, and uncertainty in the actual aircraft length as seen in
the images due to
unknown
orientation.
introduced due to the lateral fuselage orientation relative to trajectory. The uncertainty in this value was
due
to the aircraft
3 degrees
in orientation
was assumed.
(the reciprocal
used as a weight factor on the mean values. If measurements were independent, the uncertainty
mean could be
for
AAl
actual
summing
calculated by
and 542
bound on
mph
mph
14
the uncertainty
for
AA
is
for
was
1 1
the individual
UAL
175.
UAL
and
shown
in
was increased
75, respectively.
mean speed
for
UAL
Therefore, this speed was used in rurming the global impact analysis, discussed in Chapter
refinement of the analysis and associated uncertainties produced the slightly lower
mph
as discussed above.
Because
to
Table 7-3.
542
in the
this difference in
speed
is
less than
mean
Subsequent
9.
value of
unceitainty range, the speed used for the baseline impact analysis was not modified.
7.3
Estimates for the aircraft impact locations, orientations, and trajectories were further refined based on the
damage
patterns
documented on the
exterior of the
WTC towers.
to visualize
project the impact points of the wings, fuselage, engines, and vertical stabilizer onto the exterior wall of
166
NISTNCSTAR
1-2B,
WTC
Investigation
A damage pattern was then estimated and compared to that obtained previously
each tower.
from
components (wind
profile.
For
AA
11
of WTC
components of AA
was found
1 1
is
.
in Figure
is
shown
needed
in the lateral
to the figure to
to
fit
the
damage
0 degrees.
An example
shown
from horizontal =
impact condition
(fuselage orientation
is
in
Figure 7-8 where the vertical approach angle was 10.6 degrees
8.6 degrees)
= 1 80
and the
degrees).
within approximately
ft.
NISTNCSTAR
1-2B.
WTC
lateral
The
critical in
Investigation
roll
tips
was
were known
to
angle of approximately
time of impact.
167
Chapter 7
/'\
/"\
^
'
~<>
^\
y''^
fSi /5\
(1A (;
iSl
(8) rs^ (St
f!f\
f
<7
>
fS?)
.'g\
VVV V
--7
VV
S-
VVV
S"
'v-
[CCfl
'iiiiijddolcco:^
i|||eddii[l[[*
3j3D0ttCr[[i
IjUDDIClLP
nnuDDttCLLiP
Tjiinoiirrff^
iniiniij
Figure 7-8.
An
[cra
S308
B]J]
[ClO
to the face
of
WTC 2
is
shown
components of UAL
[H
0]J]
IDO
in Figure 7-9,
175.
From
its
ODjIDDO
is
ODOCCCO
IIDKOB
orientation
(vertical
OjDICLII
in
along the trajectory axis so that the projection of each aircraft component onto the tower face represents
aircraft
component with
was observed
in parts
of the
aircraft that
little
structural
defonnation
to experience structural
in Figure
7-10, which were based on video analysis alone, would cause the
starboard wing tip to miss the building and are, therefore, not physically reasonable. Also
figure are the estimated impact locations for the
wing
shown
in the
not align well with the observed impact damage. Translation of the aircraft alone does not account for the
shown
in the figure.
ft
west was needed, along with a specific relationship between the trajectory and orientation,
impact pattern
initially
168
to
in
ft
further
match. The final impact points, defined as the location where the nose of each aircraft
NISTNCSTAR
1-2B.
WTC
Investigation
fi\ fSi
fis
\s) f^.
!?
Vs.=
1^3)
4^
i^)
f^- ('gi
f^i fp\
(f
1%
^
i-rv
vs,i
>
f'^
/i^
s^} s^l V5; !|) (1^
i
!
^9^,
k9)
4^
i^nia
Wing tip
n
%
1
Engine
1!F
sir
...
Damaqe
ir
-TTiiiirLTriiai ilrrfr'
P11
03]
00
Wing
tip
[ll]D1]ll]l31Il]:il|t
-CD
Engine
'mmTM
CD
NISTNCSTAR
1-2B.
WTC
Investigation
UAL
169
Chapter 7
It
was found
between the
aircraft trajectory
maintained in order to achieve an impact pattern consistent with the damage observed on the south wall of
WTC 2.
up).
The fuselage
The difference
orientation needed to be 1
An example
where the
vertical
above the
was
A second
angle of 17 degrees,
is
moved from
Summary
Approach Angle
10.6
shown
in
Figure 7-11,
and
time with a
this
(WTC
1 1
UAL
approach
2)
+ 2 below horizontal
15
clockwise from
2 clockwise from
Structure North"
1
approach angle
-3 clockwise
from lateral
approach angle
approach angle
25 + 2
wing downward)
(WTC
(heading downward)
approach angle
Relative to Trajectory
175
542 24
is
far horizontally
Structure North"
Relative to Trajectory
Structure North
wrong
3 below horizontal
180.3
(left
(heading downward)
Approach Angle
38 2
T able 7-5.
Aircraft
AA
11
(WTC
2.0
1)
ft
175
(WTC
23.1
2)
ft
aircraft trajectory
The uncertainty
1.6.
ft
east of
0.6.
ft
above
81st floor
in the vertical
above
96"' floor
centerline
between
tower s
Location
west of
centerline
UAL
WTC
Vertical
Horizontal
relationship
degree nose-
(Velocity vector)
these parameters.
is
443 30
(Velocity vector)
The
Figure 7-12.
in
a.
175
AA
Roll Angle
UAL
shown
Table 7-4.
Lateral
(i.e.,
in projected
Vertical
approach angle
approach angle was 6 degrees (fuselage orientation from horizontal = 5 degrees) and
was maintained.
vertical
approach angle and the fuselage orientation from structure north was
in the lateral
4 degrees to 12 degrees, as shown in Table 7-3, and the fuselage orientation from horizontal from1
degree to 7 degrees. As a
in the veitical
in
approach angle and the fuselage orientation from structure north was similarly reduced, as shown
Table 7-4. The impact points of the wing
corresponded to an uncertainty
170
in the roll
tips
were known
to within
angle of approximately
approximately 2
ft.
in
This
degrees.
NISTNCSTAR
1-2B.
WTC Investigation
^-^
nil
'
13]]
'
133]
vy
-V ^'
033
i
i
Bl
II3jJJJj
'
Umil
0331
OjITODJJ]
33113]!
033]
133113-1
i33r]:|t]l]l33111
Wing
tip
]333331i]3li33I31Il
]3]li
I3]n33ri:i3ji
Engine
-aDl33I1333333331I33II33333IlJ3B33333S
-COi:-3in333]333I33I!3ll3]I]33
31]
]3I133]133333I33IIj3
ni33D333]3333133I3]
mijjii]]]]]!]]!]]^'
Oni:3I13333]333B33
33133111
I]
mmi
|3333j333I33I
mrnmrnnimi
[ioi]]i]]]]]j]]i]:
]]ij]]]ji]]|]j]D]jij:i
DDJ]j|j]|J]|JiI]J|]jJJ]|
An
Figure 7-11.
orientation
(vertical
ijjj
on
nj
j]Lmi
Wing
CJ^
LJ-:
'
iimi
mi
mi
331]3II]3]
03I
13]I]3)
mi
13311331
UJ]
]n]I33E33l3ji:]
tip
^
Engine
.
Jll3333Ji:
"S>
aOi:-:jJ3j:3333]lD3133l3JI33
aDD3JlJ3]]3:33n3]I3J
OD33:i33I133333113JS334||l33
I13H33H
-- nni]]i]]]]3iijj|]i
QDJj:1IJ333]333D3U
<E'
Figure 7-12.
An
1-2B,
J333333]n3Ii
[]DI1]|]]]]]I]]|j]|]]]]i
ii]]]]]j]j]]|]]|]]]]]n]j!
--aDij]ij]i]]ij]i]]iJ]]]]
]]i]]]]ji]]i]]]]Ji]]|]]ii
orientation
(vertical
NISTNCSTAR
[If
WTC
Investigation
171
Chapter 7
Although the
lateral
approach angle of UAL 175 had a nominal value of 15 degrees, additional observable
infonnation was used to define a most probable flight condition. Figure 7-13 shows the top view of
WTC 2 with the engines and landing gear in their pre-impact location.
trajectory of the starboard engine of
15,
UAL
175 with an
initial lateral
it
Also shown
is
the projected
at the
this
possible that the tower structure and/or contents deflected the engine from
WTC
This
1.
in
Chapter 9 used
UAL
its initial
175
lateral trajectory.
This configuration did not cause substantial deviation in the trajectoiy of the starboard engine.
lateral trajectoiy
was, therefore, the most likely and was adopted for the global analyses.
7.4
by other
studies.
analysis
was
initial
to provide
may
between the
Many
not have been available in the previous analyses. In this section, a comparison
aircraft
in this
full
of these data
is
presented
provides an opportunity to review the methodologies applied, as well as assists in the determination of the
uncertainties in the impact conditions.
data
evaluation of those estimates of impact conditions and determination of their uncertainties could not be
made.
Table 7-6 compares the results of the motion analyses for the
and
172
MIT
AA
1 1
analyses utilized the Doppler shift of the engine noise to determine the aircraft speed.
NISTNCSTAR
1-2B,
WTC
The Hart-
Investigation
AA
1 1
at
than the analyses presented here (NIST analysis in Table 7-6). If the Hart-Weidlinger analysis had the
approaching
aircraft
One
analyses.
at a steeper angle,
it
V2
in detemiining the
AA
Table 7-6.
AA
(WTC
much
closer to the
1) aircraft
MIT
and NIST
1 1
11
a speed
AA
1 1
FEMA"
Weidlinger''
Mir
NIST
470
500
429
443
51
30
+ 30
Lateral
Vertical
50
a.
4"
4.3"
0.3
6.2"
10.6
25"+ 2
20.7"
b.
c.
/ -
UAL
Table 7-7 compares the results of the various motion analyses for the
Weidlinger and the analyses presented here were consistent with the exception of the
MIT
angle.
The
lateral
approach
lateral
estimates of impact speed are low compared to the other analyses. However, assuming a
MIT
estimate of the
UAL
524 mph.
UAL
Table 7-7.
175
(WTC
2) aircralt
FEMA'
Weidlinger''
Mir
NIST
590
550
503
542
38
24
Vertical
Approach
.A.ng]e
(downward)
wing down)
11.7"
15"
15
2.7"
0"
2
2
382
30.1"
b.
Building Performance Study, May 2002. Analysis methodology or data source not available.
Levy. M.. and Abboud, N., 2002, "World Trade Center - Structural Engineering Investigation," Hart-Weidlinger.
c.
a.
SUMMARY
7.5
Institute
WTC
towers.
The
initial
to
determine the impact conditions for the two aircraft that impacted
and
incidence, roll angle of each aircraft, and the location of nose impact with each tower.
to calculate the
vertical angles
The
first
NISTNCSTAR
1-2B,
was used
WTC
method
length of the aircraft in the video to calculate the impact speed. Finally, analysis of the impact
the face of each tower
of
Investigation
and
trajectory. This
damage on
was done by
173
Chapter 7
matching the projected hnpact points of the wings, fuselage, engines, and
exterior wall of each tower to the observed
damage
pattern.
REFERENCES
7.6
of Air-Delivered
Inc., Letter
Report to the
FEMA,
2002, "World Trade Center Building PerfoiTnance Study: Data Collection, Preliminary
Foecke,
T.,
and
Pitts,
W.,
NIST
Kausel, E., "The Towers Lost and Beyond," Massachusetts Institute of Technology,
http://web.mit.edu/civenv/wtc/index.html.
Abboud
N., 2002,
May
2002,
Weidlinger Technical Report, Prepared for Wachtel, Lipton, Rosen, and Katz, August
Lipton, E., and
J.
Tower
to Fall
Was
Hit
at
1.
New
23.
174
NISTNCSTAR
1-2B,
WTC Investigation
Chapter 8
Uncertainty Analyses
INTRODUCTION
8.1
The objectives of the uncertainty analyses were to assess the effect of uncertainties associated with the
aircraft and World Trade Center (WTC) towers on the level of impact-induced damage to the towers and
to determine the most influential modeling parameters that affect the damage estimates. Uncertainty
from the following key parameters:
Aircraft impact parameters: aircraft speed, horizontal and vertical angles of incidence,
orientation,
2.
Material properties: high strain rate material constitutive behavior and failure criteria for the
mass and
3.
Aircraft
4.
Tower parameters:
relative to
5.
An
aircraft.
stiffness properties
and the
mass
distribution, connection
is
may
random
known
is
modeling
as
errors, are
methodologies on the calculated response. All of these variables do not necessarily have a significant
effect
to the
model and
in the finite
list
applicable to a limited
to
to a
be justified. Engineering
experience and intuition alone are not sufficient. Therefore, parameter screening was conducted using
design of experiments methodology. Screening was
first
conducted
at the
influential parameters
and
reduce the number of parameters to a more manageable number for the global impact analyses. These
A Plackett-Burman
J. P.
screening of parameters in the analyses discussed in this chapter. Plackett-Burman (P-B) designs are twolevel designs, run at high
and low values for each parameter. For k parameters, Plackett-Burman designs
parameters held
at their
i.e.,
the least
is
1-2B.
WTC
Investigation
One
all
NISTNCSTAR
k.
First,
175
Chapter 8
at a
in the presence
of each other's
variation.
8.2
Two
was
sets
that of an aircraft
it
collection of
column damage
empty
determine differences
difficult to
parameters.
aircraft
to engine impact.
The second
discussed in Chapter
5,
a single
in
columns was,
As
of uncertainty analyses
made
first set
set
more
sensitive
measure of
in
Section 5.5.1. Finally, the subassembly uncertainty analysis was that of the configuration discussed in
Chapter 6 impacted by an
Engine
8.2.1
WTC
first
but were
stiTJCture,
moved
same
shown
shown
in
closer together.
for a
sample configuration
in
in
in the
is
The assembly of core columns was selected from the core column airangement on floor 96 of
The
1.
aircraft engine.
initial
each simulation as appropriate for the engine interacting with the tower core
structures.
rigidly constrained
around
to a steel plate
their edges.
in the
to initiate failures at
to
A more rigid
structure.
Table 8-1
The
flight
speed of flight
was used
UAL
in the horizontal
First, these
176
to half the
in
and
was more
accurately determined.
column
shown
two reasons.
vertical
were selected
to
be appropriate for
NISTNCSTAR
the first
at the ceiling
1-2B,
WTC
height
Investigation
Uncertainty Analyses
enically.
for horizontal
and
vertical
symmetry considerations.
(a) Initial
(b)
configuration
Impact response
at 0.20 s
tower
failure strain
was considered
in
A magnitude of
A large variation in
in the table
was
this
in the
study, final data were not yet available for strain rate effects in the tower materials, so this large variation
was
selected.
factor of 0.
NISTNCSTAR
1-2B.
WTC
2,
was used
Investigation
177
Chapter 8
two
Engine
Parameters
Tower
umn
Parameter
ID
Uncertainty Parameters
Parameters
Minimum
Baseline
Maximum
Value
Value
Value
mph
mph
mph
Speed
0.00
ft
0.00
ft
2.03
ft
0.00
ft
0.00
ft
3.00
ft
Material Strength
Failure Strain
392
485
579
100%
%
50 %
150%
10%
100%
1000%
Material Strength
85
100%
Failure Strain
50%
10
10
Erosion Parameter"
Contact Parameter^
12
Friction Coefficient
13
0.0
0.3
0.6
65
Parameters
Model
Flight
many of the
100
00
100
135
%
150 %
115
%
%
1000%
Parameters
a.
Discrete parameter.
is
was
the
that
However,
calculation.
first
still
in the
erosion parameter, the free nodes can be maintained or eliminated in the contact algorithms. Deleting
nodes
may
momentum from
subsequent
impacts. However, maintaining the deleted nodes in the contact algorithms can sometimes lead to
computational
momentum,
results.
instability.
lost the
majority of its
in the uncertainty
little
effect
initial
on
the
this effect.
Finally,
coefficient
model
between contacting
contact.
is
stiffness.
LS-DYNA
and the
and
is
is
most
internal contents).
A more complete
LS-DYNA
uncertainty stems from various sources. Reference values for the coefficient of friction for
178
to
the
is
used
nodal mass and global time step to calculate the local interface
steel
friction
parts.
et. al.
1990).
However,
in this
large
aluminum on
NISTNCSTAR
1-2B.
WTC Investigation
Uncertainty Analyses
possible complexities that served to modify the effective friction coefficient. These factors included,
among
others, painted
and unpainted surfaces, erosion of metal and painted surfaces, and changes in
properties of paint and metal from heating. In addition, the ability of the contact algorithms to smoothly
calculate the interface normal and frictional forces
was
coefficient
is
A fractional factorial 2' "^ experimental design was selected for this study, as shown in Table 8-2,
with
the intent of identifying the significant parameters affecting the impact response.
factor
IDs are
maximum
of the
table,
left
hand
The uncertainty
side.
Minimum and
shown by
0.
Table 8-2. Fractional factorial 2^^'^ experimental design (with centerpoint) for the enginecore column impact analyses.
Runs
10
11
12
13
ID->
1
.1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
.1
-1
-1
-1
-1
.1
-1
-1
-1
-1
-1
-1
1
1
10
11
-1
12
-1
-1
-1
-1
1
1
14
-1
15
-1
-1
-1
-1
16
17
-1
The reduction
-1
-1
it
-1
The baseline
The complex
variation in the
damage
-1
strength
area.
were used
to calculate the
was assumed
The main
to
be the
factors that
change
full plastic
in load
load level or
reduction in load carr>'ing capacity were the reduction in cross sectional area from failed material,
deflections of the
column
or distortions that
modify the
The
first
under the impact load. The load carrying capacity was reduced by the
cross-sectional areas.
The second
column produced by
the impact
ratio
that
of the remaining to
had
failed
initial
damage. The
effect
of the centerline
displacement was determined using classical solid mechanics analysis of an axial compressive load,
centered on the column ends and aligned with the original longitudinal axis of the column. At the point
of maximum centerline displacement, the axial compressive load resulted in both a uniform compressive
moment equal to the force multiplied by the centerline displacement. The column was
assumed to fail when the maximum compressive stress in the column reached the yield stress. For
simplicity, the bending was assumed to be aligned with the direction of the maximum principal moment
force and a
NISTNCSTAR
1-2B,
WTC
Investigation
179
Chapter 8
The
sectional shape.
reduced the load carrying capacity of a column was the local distortion
cross-section
To
damage
classified in integer
scale
to failure).
The cross-section
Example
cross-section
damage was
the
to the cross-section
damage
and
due
damage
to categorize the
in the cross-
difficult to quantify
to
distortion of each
was used
was
damage magnitude
in
to 5,
(large distortions
for
of the
each column.
Effective
Pliistic
Stiaiii
The
total
damage
to a core column.
reduction in load carrying capacity of a single column was calculated by multiplying the
reductions from each factor: the global centerline displacement, cross-section area reductions, and cross-
The
section distortion.
summing over
all
the
total
reduction
columns
in the
in
was calculated by
model.
A main effects plot for the engine-core column impact analyses is shown in Figure
effect (parameter 10)
on coluinn
8-3.
3),
The
strain rate
column
materials failure strain (parameter 9), and engine materials strength (parameter 5) had the largest impact
on the residual load carrying capacity of the assembly of core columns. The uncertainty
effects
had the
largest effect
of this
set
for the
maximum
in these analyses.
factor of 5,
it
is
variation. If the
magnitude of uncertainty
parameter
effect
to a factor
of 2, not
10,
less.
Horizontal impact location (parameter 3) was the second most significant parameter affecting the residual
load capacity of the columns. The importance of the horizontal impact position was probably magnified
by the problem geometry used. For the baseline and minimum value analyses, the impact was aligned
horizontally with a
180
the impact
trajectory.
For the
would
maximum
NISTNCSTAR
1-2B.
WTC Investigation
Uncertainty Analyses
location
the engine passed between columns. For this impact scenario, significantly fewer
columns were
0.35
Engine
Ver. Impact
Maf'l
Engine
Core Col
Failure Strain
Strength
Location
0.30
0-25
Set
10
-
12
11
Core Column
Strain Rate Effects
13
\*
Contact
Parameter
Interpretation of this strong significance of the horizontal impact location for the subassembly or global
and
initial
tower exterior were not easily related to the resulting trajectory of the engine through the core. This
to the vertical
momentum,
impact location.
in
in nearly a
dominant
50 percent increase
effect
in the
engine impact
columns. The relative importance of some parameters can be easily seen. The impact location was more
at
at
higher speed. The impact response mechanisms controlling the relative importance of other
contributing
damage than
a head-on impact
(impact trajectory and orientation both aligned with the engine axis of rotation). Similarly, the impact of
engine debris caused significantly less damage than a relatively intact engine.
As
a result, the
changes
in
material ductility and strength could introduce engine impact responses that resulted in significant
changes
in the
NISTNCSTAR
1-2B,
WTC
and
Investigation
181
Chapter 8
the friction and erosion parameters were found to have an insignificant effect in this
compared
to a
result of the
more observable
much
damage
it
was
to the core
8.2.2).
likely to
have sufficient
momentum
problem geometry
This
potentially a
is
in the relatively
Based on these
effect in the
low strength
Remaining portions of
columns.
results, the
8.2.2
Tower
Tower
Engine materials
set (possibly
secondaiy significance).
exterior panel
models discussed
in Section 5.5.1
overwhelm
was expected
level of
in
Figure
in the
to the exterior
fuel-laden
The wing
to the exterior
failure.
columns of the
This
WTC
fuel.
182
NISTNCSTAR
1-2B,
WTC
Investigation
Uncertainty Analyses
Thirteen uncertainty parameters were selected for the wing-panel impact analyses, as listed in Table 8-3.
The
flight or trajectory
parameter uncertainties were obtained from the video and photographic analysis of
and
failure strain
probably
known
with greater certainty than the aircraft materials due to the detailed testing perfonned by
National Institute of Standards and Technology (NIST). The tower material strengths were, therefore,
how
long the nodes and mass of an element remained in the calculation after material failure. Therefore,
was included
in this
this
example as
is
refened to
UncertaintA Parameters
Speed
Flight
Parameters
Maximum
Value
Value
mph
443 mph
521
mph
4.0
-4.0
Material Strength
65
100%
135
Wing
Failure Strain
150%
Rivet Strength
50%
50%
100%
Parameters
100%
Weight Factor
1.5
2.0
3.0
Material Strength
100%
115%
Failure Strain
100%
50%
10%
%
200 %
Erosion Parameter''
10
Erosion Strain
11
0.2
0.3
0.4
Contact Parameter"
12
Friction Coefficient
13
0.0
0.3
0.6
Lateral
85
Tower
Parameters
Model
Parameters
a.
Baseline
Value
413
Approach Angle
Minimum
0.0
100%
50
%
%
50
Di-screie parameter.
is
known from
significant uncertainty in the weight of the wing, since not all of the
wing was,
Chapter
The
is
exterior panels
4.
rivet
in the
minor
structural
estimates, as described in
Therefore, a significant variation in the wing section weight scale factor was considered here.
connection strength between the interior structure and skin was also varied. Rivet sizes used
in the
was, therefore,
considered.
NISTNCSTAR
1-2B,
WTC
Investigation
183
Chapter 8
in
Three deterministic parameters, or modeling parameters, were considered. These parameters were the
erosion parameter and two contact parameters, contact type and friction coefficient. These parameters
wing section
in
Section 8.2.1.
column and
The primary
uncertainty analyses were very similar, as listed in Table 8-1 and Table 8-3, respectively.
differences were the uncertainty parameters used for the impact conditions.
for the engine-core
the core.
The
bounds on the
factorial
hand
side.
The
to interior structures
and
tower materials
that
on the tower
similar, with
were reduced
steels.
to
It
move
number on
the
as
of the debris
at a sufficient
of the
residual linear
by placing
lateral
experimental design was used for the wing-panel impact analyses as shown in
damage
and
lateral
field
distance from the panel such that contact with the wall had no
184
NISTNCSTAR
1-2B,
WTC
Investigation
Uncertainty Analyses
-1
-1
2^^'^
-1
-1
.1
10
11
12
13
>
1
-1
'1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
-1
1
-1
-1
10
-1
11
-1
-1
-1
-1
-1
12
-1
-1
-1
-1
-1
13
-1
-1
-1
-1
14
-1
15
-1
16
-1
17
-1
-1
-1
-1
-1
-1
-1
A main effects plot for the wing-panel impact analyses is shown in Figure
effects plot, all responses at the
and
denoted as the
To
8-6.
generate a main
shown
The
in the figure.
uncertainty in the wing weight had the largest effect on the response of the wing-panel study. However,
the relative uncertainty in the total aircraft
aircraft
of the wing
that
itself.
The
4.
with the baseline scale factor of 2 to 3 (a 50 percent change in wing mass), for example, resulted in only
an
change
8 percent
in fuselage weight.
was
as
much
also dominated
fuel,
2
^
1.4x10
13
*
1
A
*
5
*
'
1
'
6
I"
7
-
10
"
Impact
Wmg
Wing
Panel
Erosion
Ductility
Weight
Ductility
Parameter
Trajectory
1-
12
11
-
Contaci
Algorithm
1.2 i-
1,1
1.0
0.9
0.8
0.7
06
0 5
I
0.4
Impact
Wing
Rivet
Sp(J
Strength
Strength
Friction
Coefficient
Erosion
Strain Rate
Effects
Panel
Strength
Strain
10
11
12
13
Factors
Figure 8-6. Main effects plot for the wing-panel impact analyses.
NISTNCSTAR
1-2B.
WTC
Investigation
185
Chapter 8
Secondary uncertainty factors included the erosion parameter, impact speed, wing material yield strength
and
ductility,
tower materials
ductility,
and the
friction coefficient.
in these analyses,
a larger
it
caused
computational stability problems in some analyses. The negative effect of maintaining the free nodes in
contact outweighed the benefit of keeping this
Based on these
results
momentum
be potentially significant
in the
in the global
impact calculation.
in the following
listed in order
of
importance):
Impact speed.
Tower
Friction coefficient.
Erosion parameter
8.2.3
materials ductility.
Figure 8-7 shows the configuration selected to perfonn the engine-subassembly uncertainty analysis. The
subassembly model
is
the
same
a strip of the tower with a width of three exterior wall panels and extending through the truss floor system
to the core for a height of three floors. Eleven parameters
in this analysis, as
shown
in
flight
were selected
horizontal approach angle, and lateral approach angle were varied in this analysis. Uncertainties from the
AA
impacting the
1 1
WTC
tower were
The
used. Analyses of these impact conditions and associated uncertainties were described in Chapter 7.
moved
by 3.28
ft
as
shown
in Figure 8-8.
The
approach angle was considered important because this parameter contributed significantly
magnitude on the
The
lateral
approach angle
in this analysis
determined the path of the engine through the building contents and influenced the potential for an impact
on a core column. Horizontal impact location was not varied because variations in the
lateral
approach
186
NISTNCSTAR
1-2B.
WTC Investigation
Uncertainty Analyses
Speed
Tower
Value
mph
4 4
1
443
mph
Maximum
Value
472
mph
X'ertical
Approach Angle
7.6
10.6
13.6
0.0
2.0
4.0
Material Strength
100%
135%
Material Strength
100%
115%
Failure Strain
100%
150%
100%
200
100%
160%
Approach Angle
-3.28
ft
%
85 %
65
10
50%
10%
60%
Erosion Parameter"
11
Parameters
Model
Baseline
Value
Impact Location
Lateral
Engine
Parameters
Minimum
\'ertical
Flight
Parameters
Parameter
ID
0.00
ft
3.28
ft
Parameters
a.
Discrete parameter.
NIST NCSTAR
1-2B.
WTC
Investigation
87
Chapter 8
'
(a)
Maximum
(b)
in vertical
Mininimum
Tower
materials failure strain and strain rate effects were considered in this analysis because of their
importance as indicated
in the
was reduced
to a
more appropriate
were obtained and evaluated. Engine material strength and the engine material
study for the
same
uncertainty analyses.
Tower
material strength
was
is
set
were included
in the
was
assumed weight
Chapter
3.
The
in this
as in the component-level
columns, floor tmsses, and concrete slab had not been investigated
An
same way
A description of
in
160 percent of the baseline weight. Finally, the only modeling uncertainty parameter included was the
erosion parameter. Neither variation in the contact parameter or the coefficient of friction parameter had
a significant effect in the component-level uncertainty analyses,
and
were not
188
columns
after penetrating
1 1
factors
was
through the tower exterior and truss floor region of the building.
NISTNCSTAR
1-2B,
WTC Investigation
Uncertainty Analyses
XI
X2
X3
X4
X5
X6
X7
X8
-1
X^t
X10
XII
ID->
1
2
:
-1
-1
-1
-1
-1
-1
1
...
-1
-1
-1
-1
-1
-1
-1
10
-1
-1
-1
-1
-1
12
-1
-1
-1
13
11
1
-1
impact analyses
shown
is
in
tower materials strength, and the engine vertical approach angle had the largest effect on the
residual kinetic energy of the engine materials. Parameters with the smallest effect
1-
>
j-^
10
TrafGCiory
Vi. tmpaci
Yaw
"
7x10'
and strength.
set
M^l'i Strength
Tower
Liv6 LOAtS
Ouctiltly
Wplght
TrajcTory
Engine
Tower
Siratn Rale
Pitch
Ual'l Set
Strength
EMccls
Sl3ee<l
u
9
Erosion
Parameter
>-
10
11
Factors
Figure 8-9. Main effects plot for the engine subassembly impact analyses.
The
vertical
impact location had the largest effect on the impact response. For the uppermost engine
shown
in
to
glance off the floor slab and continued into the core on the same
Figure 8-10. In four of the six runs where the impact point was in the lower position
was
significant
engine trajectory.
NISTNCSTAR
1-2B,
WTC
Investigation
189
Chapter 8
= 0.02 s
= 0.06 s
= 0.10 s
= 0.30 s
results, the
1.
uncertainty analyses:
Tower
Engine
Of secondary
Of the
materials strength.
vertical
approach angle.
importance, but
still
Impact speed.
Engine
Tower
lateral
approach angle.
materials ductility.
190
NIST NCSTAR
was obviously
1-2B,
WTC
Investigation
Uncertainty Analyses
important for the subsequent response of that engine. However, the global impact analyses included the
aircraft.
vertical
the aircraft, as well as the trajectory (roll, vertical approach, and lateral approach angles), and the aircraft
orientation influences
where various
aircraft
Therefore, modillcation of the vertical impact location for a specific engine in these subassembly analyses
has a
much
vertical
impact
location of the entire aircraft. In addition, changing impact parameters, such as the aircraft pitch, changes
the vertical impact point of the engines.
8.3
Based on
in
to
be significant
each analysis are summarized in Table 8-7. Secondary parameters are identified with a
Based on
these results and on engineering judgment as explained below, the following parameters were selected for
variation in the global impact analyses:
Impact velocity.
Tower
vertical
w hat
analyses.
The
A change
this
magnitude of the
this debris,
its
may
increase the
in the global
aircraft
in
vertical
secondary importance
was affected by
vertical
significant in the
in the
in the global
in the
component
and
simultaneously decrease the damage due to another component, resulting in less pronounced overall
effect.
NISTNCSTAR
1-2B.
WTC
Investigation
191
Chapter 8
Table 8-7.
Summary
Column
Component
Engine
Subassembly
Impact
Exterior Panel
Component
/.
Impact velocity
Flight
Parameters
Vertical
-/
Approach Angle
/.
Approach Angle
Lateral
-/
Aircraft
V.
Parameters
Engine materials
set
Tower
materials strength
Tower
Tower
V.
Parameters
Tower
^-
effects
V.
Friction coefficient
Model
Erosion Parameter
Variations in the strength and ductility (failure strain) of materials had a similar effect on the amount of
energy absorbed.
An
is
The
was
typically
known more
where the
value of the failure strain needed to be assigned based on the model resolution and failure criteria used.
As
and tower
in the
The uncertainty
in the
weights associated with building contents (corresponding to service live loads) was
in the
more
and
debris dispersion. In addition, the partition walls were significant for controlling the subsequent spread of
fire
was included
as an
The wing
section weight
was found
to
total aircraft
exterior panel
parameter. The uncertainty in the total aircraft weight was, however, significantly smaller than the
unceitainty in the weight of the empty wing section impactor.
An
uncertainty of 5 percent
was used
for
set
and the
friction coefficient
to
be sufficiently significant
in
any of
the uncertainty analyses and were, therefore, not included in the global analysis. Similarly, the erosion
192
to
be significant only for the wing-section impact analyses. The importance of this
NISTNCSTAR
1-2B,
WTC Investigation
Uncertainty Analyses
this
example as
wing-section impactor and the large amount of material failure and erosion as the wing was segmented by
the exterior columns.
The
analyses and potentially had negative consequences for the stability of the global impact analyses led to
the elimination of this parameter from the global impact analyses.
REFERENCES
8.4
"LS-DTOA Keyword
April.
Oberg.
E.. F.
Inc.,
NY,
Plackett, R. L.,
23''''
2202-2203.
and Burman,
J. P.,
Biometrika, 33.
NISTNCSTAR
1-2B,
WTC
Investigation
193
Chapter 8
194
left
blank.
NIST NCSTAR
1-2B,
WTC
Investigation