"Their hearts were full of enthusiasm, pride in country, faith in their God, and a willingness to accept risk in the pursuit of knowledge --- knowledge that might improve the quality of life for all mankind. Although we grieve deeply, as do the families of Apollo 1 and Challenger before us, the bold exploration of space must go on."
From a statement from the families of Columbia, Feb. 3, 2003
Chilled to minus 423 F, liquid hydrogen thunders through the Space Shuttle Atlantis' 12-in. diameter fuel lines at a rate of 300 gallons per second. Upon ignition, it produces a cumulative thrust of 37 million horsepower. And a spacecraft with a lift-off weight of 4.5 million pounds has accelerated from zero to 17,400 milers per hour in 8.5 minutes.
The 26th successful flight of Atlantis last October would not have been possible without Gas Tungsten Arc Welding (GTAW) to restore integrity to a hydrogen fuel line component. This may be the most important and famous act in the history of welding maintenance repair and operation. Florida Senator Bill Nelson certainly thought so. He entered this welding achievement in the Congressional Record.
United Space Alliance (USA, Houston) is NASA's prime contractor for day-to-day operations of the space shuttle system. This includes mission design and planning, flight operations, astronaut and flight controller training, and vehicle processing, launch and recovery. After every mission, USA performs extensive checks and verifies equipment performance.
In the course of normal operations, USA quality inspector David Strait noticed a small crack in Atlantis' fuel line bellows shield. Subsequent investigation uncovered a total of 11 cracks in all four shuttles. No crack was longer than 0.6 in. Engineers ultimately determined that they had probably existed for several missions, but following USA's "safety first" motto, every employee has the right to call a "time-out" that halts the normal flow of operations. This was one of those cases, and NASA and USA backed Strait.
The next day, a team of USA and NASA engineers, technicians and welding operators from at Kennedy Space Flight Center (Cape Canaveral, Fla.), Johnson Space Center (Houston) and Marshall Space Flight Center (Huntsville, Ala.) came together to put safety first and restore integrity to the fuel liners.
The Crack, or Finding a Needle in a Haystack
To allow for articulation and thermal expansion and contraction in the fuel lines, each line incorporates three metal bellows. From the side, these bellows resemble and function like the bendable portion of a drinking straw. To protect them from direct impingement by the fuel, each bellows has an upstream and downstream flow liner made from 0.05 in. thick 718 Inconel, an alloy with high strength and resistance to extreme temperature ranges.
The flow liners have a series of 1-in.-long, oval-shaped slots that equalize the pressure on either side of the flow liner. The slots also facilitate cleaning, acting as drain holes for any cleaning solution and drainage of fuel prior to main engine shutdown during flight. It is at the slots closest to the Space Shuttle Main Engines that quality inspector Strait found the cracking.
GTAW Is Best, Fastest Solution
USA and NASA examined a number of different repair techniques. This included GTAW repair, stop drilling, mandrel expansion and or micro plasma transfer arc (MPTA). MPTA, which is akin to a tiny oxy-fuel torch, was discarded because it could not ensure penetration on the 0.050-in. liner material. Stop drilling and mandrel expansion were discarded because they didn't return the flow liner to its original configuration and presented other operational uncertainties.
Repair techniques that did not restore the liner's original integrity may have required reconfiguring the engine test stand and "hot firing" the engines to ensure proper function. This would have taken up to eight months, crippling the schedule for the International Space Station (a critical delivery of water was scheduled, along with delivery of a 45-ft. long, 15-ton part for expanding the station's backbone).
Of all the techniques examined, GTAW repair offered the most benefits and the lowest risk. With GTAW, an arc is struck between the work piece and a non-consumable diameter tungsten electrode. When the arc is established and a molten weld puddle forms, filler wire is added manually; it melts when it touches the weld puddle.
By controlling heat (amperage) with a foot control, which functions like a gas pedal, welders can precisely control the heat input. They can add more heat to ensure 100 percent penetration to the flow liner a repair requirement yet control the heat to prevent burning through the liner or leaving debris (slag) on the backside where access is extremely limited. Cleanliness is important because the liquid hydrogen lines require a 400 micron cleanliness level: no particles bigger than 400 microns (0.016 in.) can be observed in the line after the repair process.
USA selected an AC/DC squarewave GTAW power source that allowed their welders to strike and maintain a steady arc through its entire amperage range (1 to 400 amps). In this instance, the welding operator needed a power source that would let him quickly establish the weld puddle and quickly react to changing weld puddle conditions during the repair process. Representatives from the GTAW power source manufacturer worked with USA to ensure flawless performance in this critical application.
USA used 0.035-in. filler wire and a 0.040-in. diameter, 2 percent thoriated tungsten sharpened to a fine point and polished smooth. Pure argon was used as a shielding gas and to purge the weld area. USA takes argon to maximum purity. By industry standards, "extremely dry" argon has no more than 2 parts per million (ppm) of water. USA reduces that to approximately .5 ppm with its system of its own design. Before the argon reaches the torch and purge line, it goes through two half-micron filers, a tube of desiccant, a heater and two more half-micro filters. This is necessary because oxygen, hydrogen, other atmospheric gases or water can contaminate the molten weld pool and lead to cracking. When welding 718 Inconel, the naked eye cannot see the effects of contamination. Under 500X or greater magnification, tiny cracks are apparent.
|Jerry Goudy of United Space Alliance performs a GTAW repair on a crack in Atlantis' fuel line.
To minimize gas turbulence (turbulence can pull in atmospheric contaminants), the GTAW torch included a gas diffuser. To ensure good shield gas coverage, the power source was set for 20 seconds of gas "pre-flow" and 45 seconds of "post-flow." Good post-flow also helped cool the weld puddle, which was critical.
With the purge, all the flow liner slots, except the one to be welded, were covered with welding purge tape. The argon tube was inserted into the bottom of the flow liner and an oxygen sensor installed at the top. Because argon is heavier than air, it sinks. When the sensor does not pick up any oxygen, the argon has forced all the air out of the flow liner. A second oxygen sensor was placed next to the weld to verify that no oxygen was in the immediate area during welding.
Within 24 to 30 hours of discovering the cracks, the USA and NASA team at Marshall Space Flight Center used in-house design and manufacturing capabilities to produce a very high fidelity simulator. This would enable the repair team to practice on the simulator, increasing the likelihood of success and minimizing the risk of damage to the actual flow liner.
Practice would be critical because the 12-in. diameter fuel line limited arm access and view. To see the weld, the only option was for the operator to look under his hand and behind the torch. Mirrors and cameras were tried, but that delayed operator reaction time to the point where it might result in a weld flaw. And here's the hardest part: weld travel direction was away from the operator and often out-of-position.
Better than New
Due to the limited technical data available, an empirical test program was developed and performed at Marshall to both certify the welding operator and the repair process.
More than 100 sample coupons were created with slots stamped in them, just as in the actual part. Baseline tensile (pull testing) and fatigue (cycle testing) values for these "original" parts were established. Then, coupons were run on a cyclic load-testing device until they developed cracks similar in size and location to those on the actual flow liners. The cracked coupons were then subjected to six different GTAW repair approaches and placed back into the testing machine to recreate the cracks.
In the chosen weld repair technique, the characteristics of life for the repaired part were equal to or better than the "as stamped" original configuration. Come launch time, the crew of the Atlantis would be assured of a sound craft.
The chosen repair technique for Atlantis involved seven steps:
1. Apply light edge preparation (approx.
0.002 grind with a fluted carbide cutter)
2. Apply weld repair (details below)
3. Autogenous (no filler metal added
"feather pass" with GTAW torch to
minimize stress concentration at toes of
the weld. This pass penetrates about
0.015 in. deep; the tungsten is held
approx. 0.002 to 0.003 in. away from the
4. Grind weld crown flush with carbide
5. Radius slot end, if required, to restore
initial radius (approx. 0.125 in.)
6. Apply autogenous heat pass to slot
7. Polish slot radius with hard rubber
sander to remove oxide and any sharp
Nothing About this Job Was Easy
Because the cracks could not been seen with the naked eye (most were 0.001 in. wide; a human hair is 0.003 in.), eddy current inspection was used to plot the path of the cracks. With this technique, an electrical current is sent through material and read on a screen. Cracks cause electrical discontinuities, which appear on the screen.
To remove surface oxides before welding (oxides would contaminate the weld pool), the operator set the machine to DC "reverse polarity" (electrode positive). The welding current, flowing from the part to the tungsten, would "lift off" any surface oxides. The operator then switched to "straight polarity" to make the weld.
Normally, an operator would start the repair weld at an edge and weld into part. This uses the mass of the part to dissipate the heat. On 718 Inconel, unfortunately, the contraction of the weld puddle as it cools pulls the metal. This causes the weld to crack at the edge of the slot.
The challenge for USA welding operators was to maintain penetration throughout the length of the weld without burning through as they reach the edge. Their technique involved striking the arc just past the start of the crack, then ramping up to about 15 amps of heat to establish the weld puddle. At this point, filler metal is added intermittently. As the operator approaches the edge of the slot, he reduces heat input to close to 1 amp but starts adding filler wire constantly. Capillary action pulls the filler wire into the crack for 100 percent penetration, and the filler wire absorbs heat. Maintaining the filler wire in the weld puddle at the end of the weld also absorbs more heat, and it produces an even contour without undercuts at the end of the slot.
Remember: All of this hand-eye coordination and superb timing takes place over an average weld length just three-tenths of an inch long.
After welding repair, the part goes through a five-point non-destructive test inspection:
1. 4mm boroscope (a visual inspection with up to 50X magnification)
2. Ultrasonic testing
3. Eddy current testing
4. X-ray testing
5. Cleaning and visual inspection
Through careful planning, training and just plain operator skill, these tests proved that the weld cracks were successfully repaired. USA hopes the weld repair techniques developed for the orbiter can migrate into other aspects of military and commercial aviation, as well as other machines that use similar metals.