PARIS — Solar-array failures on three(SS/L) satellites, the latest in June, were caused by the same manufacturing defect that triggered explosions while the satellites were still attached to the rockets carrying them to orbit, according to industry officials familiar with an independent review of the problem.
The three failures all related to the satellite solar arrays’ inability to correctly vent air as their rockets traversed the atmosphere, resulting in unsustainable pressure buildup when the vehicles encountered the lower outside air pressure at high altitude.
The events occurred over a period of eight and one-half years and resulted in insurance claims totaling some $422 million. The satellites in question — the Estrela do Sul/Telstar 14 launched in 2004, the Estrela do Sul-2/Telstar 14R in 2011 and theIS-19 in June 2012 — remain operational but with reduced service lives.
The first two satellites suffered major solar-array damage in explosions that prevented one of their two arrays from being deployed. The IS-19 event was apparently less severe, and the array was ultimately deployed but with reduced power.
But all three were caused by a combination of two workmanship defects that affected three panels. The defects were overly pinching the ends of the solar panels so that their honeycomb cores could not properly vent air, and insufficiently bonding the layers of the panel, making it susceptible to explosive depressurization.
SS/L President John Celli said the two manufacturing lapses both needed to occur on a given panel for the explosion to occur. He said it is all but certain that these manufacturing errors did not occur on other SS/L satellites that were launched.
In a Jan. 4 interview, Celli said the Independent Oversight Board hired to investigate the IS-19 failure had at its disposal forensic tools that were not available in 2004. For the 2011 anomaly, he said, SS/L believed it had found a cause and did not think that after 600 hundred solar arrays launched during the years, the problem could lie there.
“I have asked myself many times: Why did it take three failures for us to resolve this?” Celli said. “You can only act on the data you have. In 2004 we did not have the same data as we did in 2012. We have never before seen a ground failure of our solar arrays. We just couldn’t believe that bonding or venting was a problem.”
The 2004 and 2012 incidents occurred when SS/L 1300-model satellites, the most popular commercial telecommunications satellite model in recent years, were launched on Sea Launch vehicles operated by Sea Launch AG of Bern, Switzerland. The 2011 anomaly occurred when the SS/L satellite was climbing to orbit aboard an( ) Proton vehicle.
Sea Launch and Palo Alto, Calif.-based SS/L never got to the bottom of the 2004 incident, a fact that caused grumbling among insurance underwriters but otherwise did not prevent either SS/L or Sea Launch from continuing operations. Industry scuttlebutt at the time was that given the number of satellites successfully launched by SS/L, the failure must have been due to Sea Launch.
Seven years later, after the May 2011 failure, SS/L satellites were grounded for two months and the satellite builder spent some $13 million on an intensive in-house investigation involving more than 200 engineers.
The conclusion was that a small nylon hook holding the solar array’s graphite cabling in place somehow came loose. The cable snapped under pressure and broke part of the solar panel. SS/L redesigned the hook and reinforced its attachment.
Celli said that analysis is still good — SS/L reproduced the phenomenon on the ground — but that it mistook an effect for a root cause. It was the pressure buildup in the solar array that caused the explosion and the uncontrolled release of the cabling.
Celli said the IS-19 investigation was also assisted by additional information from ILS of Reston, Va., which commercializes the Russian Proton rocket. Celli said ILS had provided the data requested of it after the 2011 failure, but that SS/L did not know enough until 2012 and the third incident to request specific raw telemetry about what happened under the ILS Proton fairing.
ILS Chief Engineer Jim Kramer said that neither SS/L nor ILS had thought in mid-2011 there was a link between what happened in 2004 and the second event, and that SS/L in any event appeared to have solved the problem in its investigation.
In a filing to the U.S. Securities and Exchange Commission in November, SS/L’s former parent company, Loral Space and Communications — which sold SS/L in early November — said SS/L had incurred a $22 million charge in the three months ending Sept. 30 as a result of the solar-array issue. This is in addition to the $13 million SS/L spent following the 2004 incident.
Luxembourg- and Washington-based Intelsat has filed a claim for about $84 million with its insurance underwriters, estimating that about one-third of IS-19’s power has been lost because of the problem.
The two companies initially pointed the finger at each other. SS/L, which in recent years has been the most successful manufacturer of commercial telecommunications satellites, said its hardware has had no such problem elsewhere — except during the 2004 launch, also aboard a Sea Launch rocket. But even that link was not clear. A half-dozen SS/L-built satellites were successfully orbited by Sea Launch vehicles in the intervening years with no problem.
Sea Launch, with less of a track record than Loral in the past few years, was put on the defensive but said its telemetry showed no unusual functioning of the rocket from the time of launch until it separated the IS-19.
With an initial investigation reaching no definitive conclusion — similar to what happened during the inquiry following the 2004 anomaly — SS/L and Sea Launch agreed to divide the cost of their own in-depth investigation.
The three-member Independent Oversight Board that performed the investigation was led by John Wormington, a former vice president of the Aerospace Corp. In a Dec. 19 statement, Wormington said SS/L and Sea Launch “were extremely cooperative in providing the data and analyses that we required, and worked well in solving this very elusive problem. We have very high confidence in the conclusion.”
Kjell Karlsen, president of Bern, Switzerland-based Sea Launch, said the board’s findings “validated that the launch vehicle did not cause or contribute to the observed anomaly. … Sea Launch is pleased to have had the opportunity to assist SS/L in finding a definitive cause.”
Thierry Guillemin, chief technical officer at Intelsat, said in a Dec. 21 interview that one lesson the industry might draw from these events is that failure inquiries must be driven to the bitter end and not leave as many open questions as did the post-2004 investigation.
“Unfortunately, the industry accepted this,” Guillemin said, adding that Intelsat applauds the extra effort, and expense, made by Sea Launch and SS/L to drive the issue to ground.
Several industry officials said the solar-array manufacturing glitches occurred among SS/L’s Japanese subcontractors. Celli declined to assign blame to anyone but SS/L. “It is our design, and we are the prime contractor,” Celli said. “We have worked the issue with all our suppliers.”