Launch failures: Titan Groundhog Day
by Wayne Eleazer
|Further adding to the “pucker factor” for the D-9 launch was the fact that a definite cause for the D-7 failure had not really been nailed down, at least in some personal opinions if not in formal documentation.|
Any launch produces some concern, but the launch of Titan 34D-9 planned for that day was considerably more intense in that respect than most. The previous launch from SLC-4, Titan 34D-7 in August of 1985, had failed. The loss was especially traumatic. A few of the “big” Titans equipped with solid rocket motor strap-on stages had failed at Cape Canaveral, but none had been lost at Vandenberg before the D-7 launch.
Further adding to the “pucker factor” for the D-9 launch was the fact that a definite cause for the D-7 failure had not really been nailed down, at least in some personal opinions if not in formal documentation. The success of earlier launches had encouraged the removal of some telemetry measurements from the vehicle in order to save weight and reduce cost. The Air Force mishap investigation board found that some very basic engine performance data was not available.
The failure occurred when the core vehicle liquid engines failed to start properly, leaving the vehicle on a ballistic trajectory after the solid strap-on motors burned out, but why that problem occurred was not all that clear. The investigators also complained that the Western Test Range’s long range tracking cameras were inadequate, although in reality the range would have had to provide telescopic X-ray cameras in order to determine what was occurring inside the engines, given the lack of engine measurements. The US Navy had even been enlisted to search for the remains of the rocket on the sea floor, but nothing very useful was found through those efforts.
Finally, the mishap board concluded there had been not just one but two propellant leaks, one of fuel and one of oxidizer. The board directed considerable suspicion at the clamps that held the propellant feed lines to the engine, although why the clamps, widely used and previously highly reliable, would fail was something of a mystery. A few people knowledgeable of the investigation even went so far to suggest that the neither the clamps nor the liquid engines themselves were the real cause. Instead, maybe the solid strap-on motors had spat out some flame sideways and damaged the engines.
Aside from the Titan woes, the loss of the Space Shuttle Challenger in January had added additional urgency to the situation. The shuttle failure made it almost certain that the new Titan IV program would be expanded far past its originally planned ten launches in order to substitute for the shuttle’s capabilities. The T34D was supposed to have been the last of the big Titans: booster production had been shut down almost four years earlier, with only a half dozen more launches planned before the Shuttle took over permanently. Titan IV originally was supposed to merely “complement” the Shuttle for a very limited number of payloads, not replace it. With the loss of the Challenger, the Titan IV program had a much brighter future, but only if it could be made a reliable system.
So it was a large and hopeful launch team—over 120 people—that crowded into the SLC-4 blockhouse for the D-9 launch. They had worked very hard to recover from the D-7 failure and to prepare the next vehicle. The D-9 launch was to be both the culmination of their efforts and the recovery of a degree of US heavy launch capability. Meanwhile, overhead, to help address the D-7 mishap board’s concerns on tracking camera capabilities, the US Navy had supplied a Cast Glance-equipped P-3 aircraft. The P-3 used special stabilized cameras to collect photographic data from the air; the airplane’s crew was about to get an eyeful.
|Unlike the Challenger failure of three months earlier, the T34D had not come apart due to a design flaw. While the motor design had been tested and qualified over 20 years before, as well as requalified for the longer T34D solid motor design change, the failed motor segment simply had not been manufactured correctly.|
Liftoff for the D-9 mission occurred at 10:45 am. It took but a few seconds for even casual observers to realize something was terribly wrong. The sound of explosions boomed across the base. A large portion of one of the massive solid motors arced through the air, looking like an enormous tumbling florescent tube. The vehicle impacted in the launch pad area, doing serious damage to not only the SLC-4E pad used for the T34D but also the nearby SLC-4W pad that launched T34B missions. Extraction of the blockhouse personnel from the badly damaged site proved to be a challenging effort, not only due to the large number of people involved and the highly toxic fumes from the burning wreckage, but also because of the numerous automobiles and trucks that were aflame in the area.
Everyone was shaken by the failure, but fortunately no one was hurt. The mishap investigation revealed that, at T+8.7 seconds, a burn-through occurred in one of the strap-on solid rocket motors. The motor ripped open, broke loose from the core vehicle, and due to failure of some destruct system components, a large portion fell intact in the launch pad area and exploded.
The cause of the burn through was improper fabrication of one of the motor segments, which allowed hot gases to reach the outside casing. Unlike the Challenger failure of three months earlier, the T34D had not come apart due to a design flaw. While the motor design had been tested and qualified over 20 years before, as well as requalified for the longer T34D solid motor design change, the failed motor segment simply had not been manufactured correctly.
The T34D suffered from the problem that all US expendable vehicles faced in the late 1970s: they were not going to be around much longer. T34D production was shut down in the early 1980s, and it was supposed to be the last of expendable booster developed before the shuttle took over. With those plans known well in advance, it should not have surprised anyone that workers on the program found other jobs with more of a future, and managers in multiple companies associated with the Titan program chose to redirect their resources to more profitable ends. Under this pressure, the strap-on solid motor manufacturer, Chemical Systems Division (CSD) of United Technologies, simply forgot how to make the motor segments properly. Furthermore, normal inspection methods at the factory were unable to detect such flaws.
The corrective actions required were drastic. CSD basically had to relearn what it had forgotten about making the Titan motor segments, even as the company prepared to resume production of a newer version of the motors for the Titan IV program. The other problem was how to detect the kind of mistake that caused the T34D-9 failure. Very large X-ray machines were installed at both Cape Canaveral Air Force Station and Vandenberg AFB. Each Titan solid motor segment was x-rayed and the results studied prior to the segment being stacked to assemble motors at the launch site.
Meanwhile, as a result of the national decision to return to use of expendable launch vehicles following the loss of the Shuttle Challenger, the Titan IV program was expanded enormously. Instead of the only ten launches of a single type of booster envisioned under the original Complementary Expendable Launch Vehicle (CELV) program, all which were to be launched from a single pad at Cape Canaveral, the program eventually grew to 39 vehicles with multiple different versions. Titan IV versions came to include not just those using Centaur upper stages as originally planned but also Inertial Upper Stage and no upper stage variants. Two Titan IV pads were employed at Cape Canaveral versus the original one, SLC-4 at Vandenberg was converted to Titan IV as well, and an additional new Titan IV Centaur pad was planned for Vandenberg.
The new demands on the Titan IV program required additional performance, and a competitive procurement for new solid strap-on motors was conducted by Lockheed Martin, the prime contractor. Unlike previous Titan procurement efforts—and, in fact, all previous government booster procurements—Lockheed Martin was directly responsible for booster hardware procurement rather than the Air Force. The objective of the new solid motor procurement was to attain both increased performance and reduced costs, hence the competition.
|The motor seemed to perform well—for a bit over one second. Then it exploded.|
The result of the competition was that a completely new design solid strap-on booster was to be developed by Alliant Technologies. The new strap-on motors that would fly on the Titan IVB version of the booster featured composite rather than the old steel cases, with much larger segments and considerably higher performance than the CSD motors that had been used. The older CSD solid motors would be phased out, just like they were supposed to be before the expanded Titan IV program began.
Then problems began to crop up. The first new Titan IVB motor segment came off the production line—and immediately was found to be delaminating. This imposed some delays while the problem was corrected.
Next, the delamination problem having been solved, the first new motor was being stacked at the rocket test stand at Edwards AFB, and disaster struck. The counterweight for the crane fell off, tragically killing one of the workers. The crane then collapsed, impacting and igniting the motor segment. This created a considerable delay as the test stand was repaired, the mishap investigated, and another test motor assembled on the site. And then came the day when the first new motor was test fired.
The motor seemed to perform well—for a bit over one second. Then it exploded.
An investigation revealed a need for a complete redesign of the motor. The original design had been manufactured correctly but had a major flaw, one inherent to the design of the motor and not seen before. That finding shocked the solid motor industry worldwide. There was another considerable delay to the Titan IVB program.
Meanwhile, Titan IVA launches still had to proceed. Plans were revised to move payloads planned for the Titan IVB version to the Titan IVA. Given the delays with the new motor, and after being told they had lost the competition, CSD was issued new contracts to keep producing the earlier version. And, of course, due to the wholly unexpected series of delays, the new orders were done in a halting manner. The effective message to CSD was, “What a minute, you’re still being replaced, but we need another few of your motors.”
There is an old saying that you don’t want to buy a car that was built on either a Friday or a Monday, those days in which workers had their minds either on the weekend or recovering from it, but that describes the unexpectedly ordered Titan IVA solids. The workers were not just looking forward to a weekend; they were once again looking forward to either being laid off or transferred to other programs.
On August 2, 1993, at 12:59 pm, Titan IVA K-11 lifted off from SLC-4W at Vandenberg AFB. This launch was different from past launches of classified payloads from Vandenberg; the Air Force invited the civilian press to watch—and once more the non-standard visitors got an eyeful.
The flight seemed to proceed normally during most of the solid motor burn, but at T+101 seconds the vehicle exploded.
|The irony of this failure was not just that it repeated a past failure, but that ways to avoid such a loss of attention and experience had been a subject of considerable discussion in the Pentagon.|
The mishap investigation revealed that one CSD rocket motor segment used in the Titan K-11 had an interesting history. The rocket motor inspection process occasionally revealed flaws in the motor segments and a procedure had been developed to correct some of them. The procedure was called a “pie cut,” although it more resembled cutting a slice out of an old style pound cake, the kind with the hole in the middle. The suspect portion of a flawed segment was cut out and new solid motor material poured into the resultant gap. The suspect segment used on K-11 turned out to be the most extreme case of this repair, with the majority of the originally poured material removed and replaced. This generated some controversy, but the concept of such an extensive repair was, in the end, accepted as being intrinsically no different in concept from the smaller repairs.
The repaired segment was X-rayed at Cape Canaveral and was rejected; it was then shipped to Vandenberg where it was accepted and assembled in a motor for Titan K-11.
However, the extensive repair by itself was not the problem. The real problem was that the workers who knew how to perform the repair had left CSD. Following the pouring and curing of the replacement “pie” section of propellant, the edges of the cut were supposed to be sealed to prevent the hot gases from reaching the outer case as had occurred on Titan 34D-9. But the sealing step was not done on the repaired segment for the K-11 mission. Lacking the necessary sealant, the hot gases did indeed reach the motor case, shortly before the end of the motor’s burn, and the vehicle exploded.
The irony of this failure was not just that it repeated a past failure, but that ways to avoid such a loss of attention and experience had been a subject of considerable discussion in the Pentagon. The analysts in the Defense Department’s comptroller office had argued that the Titan IV boosters should be built at a rate that was as economically efficient as possible. The completed boosters could then stored until launched and the production facilities either shut down or diverted to other uses, resulting in a far lower cost to the Titan IV program than building the vehicles slowly. In other words, the manufacturers’ plants should not be left nearly idle between vehicles that rolled off the production line, nor should the production speed be delayed except as required due to technical issues and funding limitations. The comptroller’s concern was that with a slow production rate, the Air Force had to pay the contractors to keep their plants operating even when they were not building hardware very fast, so it cost more per booster to build them at a slower rate.
Both the Air Force and the National Reconnaissance Office strongly disagreed with the comptroller. Building the boosters at a brisker pace would indeed mean they would sit in storage until launched, but during that time period once again the companies that built the hardware would lose interest in a completed program and divert their resources to future projects. The failures associated with the shuttle shutdown had shown that. The irony comes in when you realize they had inadvertently duplicated the same mistake—again—with the CSD solid motors as a result of the delays with the new Alliant motors.
The Titan IV program proved successful overall and, even more than that, it proved to be absolutely vital to the US recovery from the Space Shuttle’s problems. That said, nearly every basic assumption made at the start of the CELV program proved to be utterly false, or at least deeply flawed, and the Titan IV’s problems were not yet over—but that’s another story.