"The Trail to the Leading Edge," Feature Article, March 2004

FEATURE FOCUS: Failure Analysis

the trail to the leading edge

Detective work in San Antonio led to understanding the loss of Columbia.

by James D.
Walker and
Donald J. Grosch

Here days after the world watched the orbiter Columbia disintegrate on reentry with the loss of all aboard, NASA enlisted the Southwest Research Institute to join the investigation into the causes of the accident.

The institute, in San Antonio, Texas, had a history with the Space Shuttle going back to the early 1980s. In that time, SwRI had conducted impact experiments on the orbiter's thermal protection system, experiments of the sort that would be needed now. Studies had included the effects of foam insulation, ablator, and ice striking both carbon-carbon materials similar to those that line the shuttle's leading edges and the silica thermal protective tiles that cover most of the rest of the craft.

The earliest available evidence led researchers in a direction that proved to be the wrong one. But physical experiments, computer simulation, and careful sifting through mounting evidence inevitably led investigators to the most probable cause of the Space Shuttle's failure.

Early in the investigation, discussion focused on the size, shape, mass, and velocity of a piece of foam that played lead in the launch film and video. Preliminary estimates of its size would prove very accurate. The volume of the foam that struck the Columbia would turn out to be at least 400 times greater than any of the foam projectiles SwRI had previously launched in earlier shuttle-related tests.

SwRI researchers decided to use the lab's large compressed-gas gun in the investigation. They constructed a barrel that exactly matched the dimensions of the suspect foam using off-the-shelf structural steel tubing, 5 1/2 by 11 1/2 inches in section. The barrel would guide the foam to the target.

Five days after the accident, SwRI launched down the barrel a block of foam equal in size to that of the piece in question, at various velocities.

Meanwhile, NASA Accident Investigation Team members continued refining their estimates of the size of the suspect foam and its impact velocity. They studied the films and began large-scale computational fluid dynamics simulations to examine the flight of the foam from the bipod ramp area, where the front of the orbiter attaches to the external tank, to the left wing.

These computations, along with the visual evidence, were used in determining the impact conditions for the foam. The Columbia Accident Investigation Board contracted SwRI to support the investigation with impact modeling to complement the extensive modeling work by the NASA Accident Investigation Team.

A First Place to Look

Descent telemetry from Columbia showed rising temperatures in the left wheel well. Because of this information—and because no images displayed the location of the foam impact directly—initial studies looked at the thermal tiles on and around the door of the left main landing gear. The team removed a corresponding door from the Space Shuttle Enterprise, which had been used only for atmospheric flight tests, and so never needed thermal protection. Shuttle technicians attached the light (0.15 grams per cubic centimeter) thermal tiles to the door as they would have for an operating orbiter.

The researchers set up a great many more data recorders and high-speed cameras than they had in prior studies to gather images of the impact from many angles. They then launched several impacts against aluminum plates to test the strain gauges, load transducers, and cameras.

Foam near the Columbia's external tank bipod ramp—which held the orbiter's nose to the large, orange tank—quickly emerged as a primary suspect in the disaster. Ballistics testing at SwRI would blow a hole 16 inches square in a leading-edge panel on the test stand.

Next, the SwRI team fired a 1.67-pound foam block, measuring 5 1/2 by 1 11/2 by 19 inches, at the Enterprise's landing gear door—five shots at around 530 mph. The impact tests struck the door at angles between 5 and 13 degrees.

Concurrently, SwRI developed detailed analytic and numerical models of foam insulation's impact on thermal tiles. The models provided a curve for distinguishing between damage or no damage, based on speed and impact angles.

Given an impact speed and angle for an incoming piece of foam insulation, the model determined whether or not tile material would sustain damage. The SwRI model agreed extremely well with previous tile impact tests and would agree with the five tests performed during the Columbia accident investigation of foam insulation impacting tiles.

Following the tests, SwRI evaluated the tiles nondestructively, and would later examine the reinforced carbon-carbon, or RCC, panels in the same manner to determine the extent and type of damage caused by the impact. In addition, three-dimensional optical scans provided detailed post-test geometries of tile gouging.

Experiments and modeling led to the conclusion that a foam impact on the underside of the wing could not have been the cause of the accident.

A Record of Disaster

Two developments led interest away from the underside of the wing and toward the left leading edge. Ground recovery teams found Columbia's modular auxiliary data system recorder. It contained data recorded from hundreds of sensors but not transmitted to the ground. The data made clear that the earliest thermal problems started not at the landing gear wheel well, but at the leading edge of the wing itself. Also, the groups that were analyzing the foam trajectory concluded that the foam had to have hit very near the leading edge.

Once focus shifted to the leading edge, a new test apparatus was built to duplicate the structure of the wing and to hold the leading edge panels as the orbiter had held them. Reinforced carbon-carbon, a very different material from the light silica foam tiles that lined most of the orbiter's body, made up the wing's leading edge. RCC is denser (1.6 grams per cubic centimeter) than the tiles and brittle. A series of 22 RCC panels lined the leading edge of the left wing. The simulated wing structure held panels 5 through 10.

High-speed images record the impact of the foam block against the leading edge.

The SwRI team performed two tests using Enterprise's non-RCC, fiberglass panels that had been instrumented. Diagnostics were worked out and comparisons were made with analysis. As many as 15 high-speed cameras that could film at 7,000 frames per second were placed inside the hollow leading edge to measure panel deflection during impact. Up to 240 channels of strain gauge, accelerometer, and load cell data were recorded during each test.

Next, the team shot a block of foam insulation at a speed of 524 mph at RCC panel 6, which had flown 30 missions on the Space Shuttle Discovery. The foam, striking the panel at about 20 degrees, broke the interior rib. The crack ran barely into the exterior leading edge. The team thought this amount of damage, unless a similar crack had grown during reentry, was insufficient to have led to the loss of the shuttle.

The reconstruction of the final minutes of Columbia, using the data recorders and recovered pieces, now focused on RCC panel 8 as the most likely failure site. The next tests were performed against panel 8—first using one of Enterprise's fiberglass panels, then a reinforced carbon-carbon panel 8 that had flown 26 missions on the shuttle Atlantis.

To reproduce the impact of the foam block against the leading edge, SwRI researchers fired the foam pneumatically through a steel pipe at the wing target.

Investigators launched a piece of foam at the RCC
panel 8 at 530 mph at an angle of about 20 degrees. The impact blew a large hole in the panel, some 16 inches square. Thermal analysis performed by the NASA investigation team indicated that a hole 10 inches across would have been enough to bring the orbiter down and would have corresponded consistently with the sensor data of Columbia's last flight.

As it did with the tiles, SwRI modeled the impact of foam insulation on RCC panels. A numerical model simulated the panel, and an analytic boundary condition simulated the pressure load generated by the colliding foam.

Comparison with the two tests performed against reinforced carbon-carbon panels led to estimates of failure stresses within the panel material. Parametric studies on the model determined the impact location that led to the most extreme stresses in the rib and the panel. Other computations investigated the effect of a rotating foam block striking the shuttle. Rotation of the foam nearly always increased the stresses on the panel's face and rib.

Volume 1 of the Columbia Accident Investigation Board's report announced the conclusion of the investigation: "The physical cause of the loss of Columbia and its crew was a breach in the Thermal Protection System on the leading edge of the left wing. The breach was initiated by a piece of insulating foam that separated from the left bipod ramp of the External Tank and struck the wing in the vicinity of the lower half of Reinforced Carbon-Carbon panel 8 at 81.9 seconds after launch. During reentry, this breach in the Thermal Protection System allowed superheated air to penetrate the leading-edge insulation and progressively melt the aluminum structure of the left wing, resulting in a weakening of the structure until increasing aerodynamic forces caused loss of control, failure of the wing, and breakup of the Orbiter."

Technologies developed at SwRI over many years were instrumental in reaching this conclusion.

A second conclusion in the report addressed policies and procedures within NASA.

The Columbia Accident Investigation Board report can be downloaded at www.nasa.gov. Chapter 3 of Volume 1 contains the accident analysis, and Appendix D.12, or Part 12 of Volume 2, discusses impact modeling.

James D. Walker is a staff scientist specializing in computational mechanics at Southwest Research Institute's Engineering Dynamics Department in San Antonio, Texas. Donald J. Grosch manages the Ballistics and Explosives Range for the department.