This document summarizes Matt Melis' presentation on lessons learned from the Columbia accident investigation and NASA's return to flight efforts. It discusses how ballistic impact testing and finite element analysis with LS-DYNA supported understanding the likely cause of damage to Columbia's wing from foam shedding at launch. Testing of materials like external tank foam, RCC panels, ice, and windows helped validate impact models. Full-scale testing of wing leading edges provided further evidence. The research informed modifications like redesigning the external tank bipod ramp and supported NASA's return to flight.

2. Lessons from Columbia
Ballistic Impact Dynamics
Project Management Challenge 2009
Daytona Beach, Florida
Matt Melis
NASA Glenn Research Center
Cleveland Ohio

3. Contributions of Ballistic Impact Research in
The Columbia Accident Investigation
&
NASA’s Return to Flight
Project Management Challenge 2009
Daytona Beach, Florida
Matt Melis
NASA Glenn Research Center
Cleveland Ohio

5. A Brief Overview of the Shuttle Launch System

6. A Brief Overview of the Shuttle Launch System

7. A Brief Overview of the Shuttle Launch System

8. A Brief Overview of the Shuttle Launch System

9. A Brief Overview of the Shuttle Launch System

10. The Columbia Accident

11. On January 16 2003, Columbia’s leading edge
was impacted by a piece of foam suspected to
have separated from the external tank bipod
ramp at 81 seconds into its launch.
Columbia was traveling at Mach 2.46, at an
altitude of 65,860 feet. The foam was calculated
to have hit the Orbiter at 700 – 800 feet per
second

12. Insulating Foam Separates from Bipod Ramp and
Impacts Left Wing of Columbia

13. Insulating Foam Separates from Bipod Ramp and
Impacts Left Wing of Columbia

14. Bipod
Image
Ramp from CAIB Report
Showing shuttle ramp and wing
Impact
Point

15. The Bipod Ramp

16. The Bipod Ramp

17. Redesign of the External Tank Bipod Ramp
Old Design New Design

18. The Orbiter Leading Edges

19. Reinforced Carbon-Carbon (RCC) Panels Protect
the Leading Edges of the Orbiter
22 panels per wing

20. RCC Panels 6, 8 & 9 of Specific Interest

21. RCC T-Seals Seal the Gap Between Panels

22. Leading Edge Panel Used for Full Scale Tests

23. Efforts of the
Columbia Accident Investigation Board
Forensic Reconstruction
Establish Probable Cause
Support Findings with Full Scale Testing

24. The Reconstruction Effort

25. The Debris Field

26. The Debris Hanger

27. The Debris Hanger

28. The Debris Hanger

29. Reconstructing the Left Wing Leading Edges

30. Reconstructing the Left Wing Leading Edges

31. Panel 8 Found with Significant Spatter Deposition and Erosion

32. Right Wing Leading Edges Exhibited No Significant Spatter

34. The NASA Impact Analysis Development Effort
With the Explicit Finite Element Code LS DYNA
Boeing, NASA GRC, JSC, LaRC
Extensive impact dynamics experience at Glenn & Langley

35. Impact Analysis Development Effort
For the Accident Investigation
• Demonstrate valid impact analysis capability with LS DYNA
• BX-250 External Tank foam characterization
• Reinforced Carbon-Carbon characterization
• Develop finite element mesh of panels and T-seals

36. LS DYNA Predicts Car Collisions
Courtesy of LSTC

37. LS DYNA Predicts Fan Blade Containment

38. LS DYNA Predicts Water Impacts
SRB Aft Skirt Splashdown with TVC Upgrade Tank

39. The NASA Glenn Ballistic Impact Lab

40. The NASA Glenn Ballistic Impact Lab
Large Vacuum Gun

41. The NASA Glenn Ballistic Impact Lab
2 Inch Vacuum Gun

42. BX-250 External Tank Foam Characterization

43. Ballistic Research Supporting the Accident Investigation
BX-250 External Tank Foam Characterization
High Speed Video of 90
Degree Impacts
No Vacuum
708 ft/sec
Vacuum
693 ft/sec

44. Ballistic Research Supporting the Accident Investigation
LS DYNA - explicit finite element impact analysis
LS DYNA Predicts 90 Degree
Foam Impact on Load Cell
LS Dyna is an industry
standard commercial finite
element analysis code
typically used to model
impact events

45. Ballistic Research Supporting the Accident Investigation
LS DYNA - explicit finite element impact analysis
LS DYNA Predicts 23 Degree
Foam Impact on Load Cell

46. Reinforced Carbon-Carbon Characterization

47. Ballistic Research Supporting the Accident Investigation
Ballistic Impact Tests on RCC Coupons

48. Ballistic Research Supporting the Accident Investigation
Ballistic Impact Tests on RCC Coupons

49. Ballistic Research Supporting the Accident Investigation
Ballistic Impact Tests on RCC Coupons
RCC Coupon Shows No Damage After 397 ft/sec Foam Impact

50. Ballistic Research Supporting the Accident Investigation
Ballistic Impact Tests on RCC Coupons
Foam Fractures RCC coupon in half at 695 ft/sec

51. Ballistic Research Supporting the Accident Investigation
Ballistic Impact Tests on RCC Coupons
700 ft/second Impact
400 ft/second Impact

52. Full Scale Impact Analysis with LS Dyna

53. Ballistic Research Supporting the Accident Investigation
Dyna - explicit finite element impact analysis
Full Scale Panel Analysis

54. Ballistic Research Supporting the Accident Investigation
Dyna - explicit finite element impact analysis
43,000 Panel Shell Elements
147,000 Foam Brick Elements

55. Ballistic Research Supporting the Accident Investigation
Dyna - explicit finite element impact analysis

56. Ballistic Research Supporting the Accident Investigation
Dyna - explicit finite element impact analysis
Panel 6 Edge Impact Case

57. Ballistic Research Supporting the Accident Investigation
Dyna - explicit finite element impact analysis
Panel 6 Edge Impact Case RCC Damage

58. The Full-Scale Testing Effort

59. Orbiter Leading Edge Full-Scale Tests
Tests conducted at Southwest Research Institute

60. Orbiter Leading Edge Full-Scale Tests

61. Orbiter Leading Edge Full-Scale Tests
Installation of internal high speed cameras

62. Orbiter Leading Edge Full-Scale Tests
Leading edge panels mounted after camera installation

63. Orbiter Leading Edge Full-Scale Tests
Phantom digital cameras
set up inside of full scale
test article

64. Orbiter Leading Edge Full-Scale Tests
High intensity lights required
both in and outside of test
article

65. Orbiter Leading Edge Full-Scale Tests

66. Orbiter Leading Edge Full-Scale Tests
External View of RCC Panel 8 Test

67. Orbiter Leading Edge Full-Scale Tests
Barrel View of RCC Panel 8 Test

68. Orbiter Leading Edge Full-Scale Tests
External View of RCC Panel 8 Test

69. Orbiter Leading Edge Full-Scale Tests
Internal View of
RCC Panel 8 Test

70. Orbiter Leading Edge Full-Scale Tests
Post Impact of Panel 8

71. Analysis Supporting Full-Scale Tests
Dyna – explicit finite element impact analysis
Latest Dyna Predictions
Correlate with Panel 9
Test

72. LS DYNA Analysis of Panel 8 Full-Scale Test

73. Return to Flight

74. Impact Analysis Development Effort
For Return to Flight
• Full development of analysis capability with LS DYNA
• Additional foams, ice, ablator material characterization
• Begin advance RCC model development (coating)

75. Ballistic Impact Research Supporting Return to Flight
Impact Studies on RCC for Model Validation
2 grams foam 2 grams foam
2054 ft/sec 2054 ft/sec
8 grams ice 8 grams ice
650 ft/sec 650 ft/sec

76. Ballistic Impact Research Supporting Return to Flight
Impact Studies on RCC for Model Validation
2 grams foam 2 grams foam
2371 ft/sec 2371 ft/sec
8 grams ice 8 grams ice
858 ft/sec 858 ft/sec

77. Ballistic Impact Research Supporting Return to Flight
Impact Studies on RCC for Model Validation

78. Aramis Displacement Measurement System
Photogrametric Technique Determines Full 3-D displacements

79. Aramis Displacement Measurement System
Photogrametric Technique Determines Full 3-D displacements
Point Displacement vs Time Displacement Contour Plot

80. Aramis Adapted to Full-Scale Wing Leading Edge Tests

81. Aramis Adapted to Full-Scale Wing Leading Edge Tests
ARAMIS Validates LS-DYNA Analysis Models

82. Full-Scale Leading Edge Test Setup with Aramis at SwRI

83. Aramis Data Validates LS DYNA Analysis Predictions
Full Field Displacements of Wing Leading Edge Impact Test

84. Aramis Data Validates LS DYNA Analysis Predictions
Principle Strain Comparison to Bonded Gauges
Aramis Indicated
2100-2700 Microstrain
Gauge Indicated 2100
Microstrain
Note Much Higher
Amplitude 2” From Gauge

85. Ballistic Impact Research Supporting Return to Flight
RT 455 ablator impact at approximately 300 ft/sec

86. Ballistic Impact Research Supporting Return to Flight
NCFI foam impact at approximately 800 ft/sec

87. Ballistic Impact Research Supporting Return to Flight
Tile Gap Filler Material Impact Testing

88. Ballistic Impact Research Supporting Return to Flight
Tile Repair Putty Material Impact Testing

89. Ice Formations on External Tank

90. Ice Research Supporting the Return to Flight
High Density Ice (no air bubbles entrained)

91. Ice Research Supporting the Return to Flight
Identification of Ice Microstructure

92. Impact Testing of Ice
Hard ice impact at approximately 800 ft/sec

93. Hadland Camera Captures Fracture Wave Propagation
700 ft per second ice impact 280,000 frames per second

94. Cordin Camera Captures Fracture Wave Propagation
600 ft per second ice impact at 480,000 frames per second

95. Ice Impact Testing on Full Scale Leading Edge

96. External Tank Impact Testing

97. Ballistic Impact Research Supporting Return to Flight
External Tank Impact Test Article with Acreage Foam

99. Orbiter Windows Impact Testing

100. Orbiter Windows Testing at NASA GRC

101. Ballistic Impact Research Supporting Return to Flight
NCFI Foam Impact Test on Orbiter Window
Rear View
Side View

102. Al3O2 Particles Measured and Weighed
Typical Aluminum Oxide Projectiles

103. Stereo Microscope Used to Load Gun

104. Tweezers Required to Place Particles in Gun Barrel

105. Ballistic Impact Research Supporting Return to Flight
Aluminum Oxide particles impact orbiter windows
70 degree, 127 ft/sec
90 degree 359 ft/sec
50 degree 118 ft/sec

106. July 26, 2005
Return to Flight
The most photographed mission

113. Forward and Aft SRB Cameras on STS-121

114. Chase Plane Video of STS-114 Launch

123. Rendezvous Pitch Maneuver

133. Orbiter Boom Sensor System

134. Orbiter Boom Sensor System

149. Additional Reading
- The Columbia Accident - The Challenger Launch
Investigation Board Decision
Final Report by Diane Vaughan
- To Engineer is Human
by Henry Petroski - Organization at the Limit
by William Starbuck and
Moshe Farjoun
- Lessons from Everest: The Interaction of Cognitive Bias,
Psychological Safety, and System Complexity by
Michael Roberto. Obtain from Harvard Bussiness Online