NASA’s Exploration Ground Systems (EGS) program, prime launch processing contractor Jacobs, and their Artemis launch team partners have adapted their countdown procedures headed into the next attempt to send Artemis I on its way to the Moon. Liquid hydrogen (LH2) leaks seen while filling the SLS Core Stage on Launch Pad 39B at the Kennedy Space Center (KSC) in Florida stopped past launch attempts and rehearsals in the spring and summer.
After evaluating LH2 loading and leak resolution techniques during a demonstration test on Sept. 21, the launch team will give themselves more time to complete the process during the next countdown. The countdown timeline was also extended, and the team will take the LH2 loading process more slowly in hopes of finally making it all the way up to and through the terminal countdown.
Starting hydrogen loading earlier, starting more gently
The system for loading LH2 onto the SLS Core Stage and its Interim Cryogenic Propulsion Stage (ICPS) is pressure-fed, and the launch team is reducing the pressure used during the early part of the filling process when compared to earlier launch attempts in late August and early September.
“We lowered the storage tank pressure more, we extended the slow fill timeline, and we also ended up extending the fast fill timelines because the pressure was lower,” Tom Clark, Manager of Propulsion/Avionics Engineering with ERC, said in a Nov. 11 interview. He is the Core Stage Tank Propulsion Lead Engineer (TPRP) on the launch team.
The lower pressure, slower flow rate procedures were demonstrated in a tanking test conducted on Sept. 21. “Once we got into a fast fill configuration, we slowly raised the storage tank pressure,” Clark said. “That was the basic difference between launch attempt two [on Sept. 3] and our tanking test.”
A persistent leak was seen at the eight-inch diameter fill and drain quick disconnect (QD) in the LH2 Tail Service Mast Umbilical (TSMU) that connects to the Core Stage engine section, and the tanking test on Sept. 21 was used to better understand the sealing behavior of the QD and further develop loading techniques and troubleshooting procedures.
For the upcoming launch attempt targeting liftoff early on Nov. 16, another hour was added to the length of the launch countdown by extending the first of two built-in holds at T-6 hours and 40 minutes from two and a half hours in duration to three and a half hours. The Mission Management Team (MMT) will conduct a pre-tanking meeting starting at the beginning of the hold to review the status of the ground and vehicle systems and the weather forecast before deciding whether or not to fuel SLS and make a launch attempt.
Once a “go” for tanking has been given by the MMT, and then by Artemis I Launch Director Charlie Blackwell-Thompson, the liquid hydrogen chilldown process will now begin a couple of hours earlier than before, at the same time as liquid oxygen chilldown begins. “Part of that [extra time] is because by lowering the storage tank pressure, we’re lowering the flow rates, so it’s going to take longer to load,” Clark said.
“In addition to that, they want [more] contingency time in there in case we have to run our pre-planned contingency procedures, so if we encounter problems, they want a little bit more time to do that.” SLS has a maximum launch window on any given day of two hours for Artemis I, so lengthening the LH2 timeline and starting earlier should increase the chances that the systems and the launch team are ready to go at opening of the launch window.
“That wasn’t an option going into [the launch campaign] because we were borderline having the [storage] capacity of hydrogen that we need to load this newer, bigger vehicle, which is why they built a second LH2 storage tank [at the pad],” Clark explained. “But that won’t be available until Artemis II.”
“After we got in a few loadings, we realized we still have margin left, we still have enough liquid [hydrogen] left [in the storage tank] where we were able to extend the loading timeline.”
Both the SLS Core Stage and its ICPS in-space, second stage were fully fueled during the Sept. 21 tanking test. The test was conducted two and a half weeks after the second launch attempt was scrubbed on Sept. 3 after the leak at the hydrogen fill and drain quick disconnect could not be fixed.
Following the scrub, the LH2 tail service mast umbilical was again taken apart, this time at the launch pad. While the TSMU maintenance and retesting was in work at the pad, the launch team revised parts of its ground control software in preparation for the tanking test.
“We rewrote the procedure some to come up with a kinder, gentler loading,” Clark explained. “Basically, what we did was we lowered our storage tank pressure. We also slowed the chilldown phase, when we have the lower flow rates, and we tried to mimic what they were doing at Stennis [during the Core Stage Green Run]. We tried to do it per their timeline.”
Clark noted that the new quick disconnect seal also leaked at the beginning of LH2 loading on Sept. 21: “When we transitioned to fast fill, we saw the leak indication, and it did go up to about 7%,” he said. “That drove us to what we call a stop flow. We stop loading the vehicle and we safe everything at that point in time.”
At that point, the cryo loading team ran a pre-planned procedure in case they saw a leak. With the flow halted, they let the area around the QD warm up for a period of time and then started flowing LH2 again; during the Sept. 21 test, the hydrogen concentration remained acceptable after that warm up and resumption of LH2 flow.
Clark explained that the contingency procedure was developed based on testing done years ago at the Launch Equipment Test Facility (LETF) in the KSC Industrial Area. “They were seeing leakage down there with the initial testing, but they noticed when they picked back up again on subsequent runs after it warmed up some that the leak rates had improved,” he noted.
“So they actually ran specific testing for that at the LETF before we became operational. We tested specifically for that, we saw that same observation again, so we added that as a pre-planned contingency to our procedures.”
— Chris Bergin – NSF (@NASASpaceflight) March 9, 2017
The learning curve for hydrogen tanking of the Core Stage at KSC has been the focus of Wet Dress Rehearsal runs in April and June and then launch attempts in August and September. The configuration at Launch Pad 39B is different from the LETF site used for testing interfaces but not the fully integrated pad, Mobile Launcher, and vehicle complex.
The 39B complex is also different than the test stand at the Stennis Space Center where the Core Stage Green Run was conducted a few years ago. Stennis has more test controls, but it has a more compact fueling setup using docked barge tanks.
At Kennedy, the EGS and contractor launch team is still experimenting to dial in repeatable settings. “Cryo systems like predictability,” Brad McCain, Vice President and General Manager with Jacobs Space Operations Group, said in the Nov. 11 interview. “We like the same procedure over and over because that gives us usually predictable results, and so when you look at — even starting back in April and June, you know, we never loaded the same way twice.”
“We tried to adapt and mitigate the issues that we had each time, and so every time you change a variable or two you get different results. For the seal leakage, we know that thermal shocks are bad for any cryo system, so you want to minimize the thermal gradient across the seals.”
“With this particular configuration, the pressure shocks appear to correlate with the leakage that we’ve seen, so we want to minimize pressure, which is the kinder, gentler loading that [has been discussed recently],” McCain added.
The TSMU connection was disassembled, and the seals were replaced after issues in April, June, and September; the team is now focusing on evolving the loading process to something more repeatable. “I think we had to go in and look at the hardware seal,” Clark said. “But now that we’ve done that a couple of times, we’re more focused on it from an operational side of things… what can we do to be kinder to that QD seal.”
For loading liquid hydrogen at Launch Complex 39B into the SLS liquid stages, there are three sets of valves that help control the pressure and fluid flow. There is a valve complex at the LH2 storage sphere and then two on the Mobile Launcher: an LH2 “skid” for the Core Stage and one for the ICPS.
The valve complex for Core Stage LH2 on the Mobile Launcher supports a few set positions for filling up the stage. “It’s actually a three-position valve; [in addition to the close position] one is for a slow fill [and] one is for a fast fill,” Clark explained.
“So once we get into fast fill, it’s kind of wide open if we’re in that configuration both in the storage area and on the ML.” For fine control of pressure and flow at the higher, fast fill flow rate into the Core Stage, Clark further explained that the storage tank valves were used. “We lowered that storage tank pressure and then we bring it up gradually so we’re slowly increasing that flow,” he said.
The pre-planned hydrogen leak contingency procedure was run during previous tankings with different results; in the case of the Sept. 21 test, in an effort to better understand the sealing behavior of the QD, the launch team experimented with running the contingency procedure using the already lower storage tank pressure and by allowing the hydrogen concentration level from the leak to rise as high as 10% for as long as five minutes to see if the hydrogen levels from the leak would stabilize and/or drop.
A hazard gas detection system built into the Mobile Launcher is used to measure trace gases in enclosed areas like the umbilical plate cavities for all the connections from the Mobile Launcher to the rocket and spacecraft. The safety system provides awareness of a possible flammability risk from leaking propellant.
In the case of the LH2 TSMU, the detection system measures levels of hydrogen in the cavity between the mated ground and flight umbilical plates. Those areas are also purged with an inert gas like nitrogen or helium to further reduce flammability risk.
“They came across some testing that they did back in 2011, [and] I believe that testing proved that you actually have to get up to about 16% [hydrogen concentration] before you start getting into a flammable condition,” Clark said. “That was in an environment that kind of mimicked our disconnect area, those areas between the plates.”
“We use the 4% [limit] because that’s the very minimum that you would get into a flammable condition. We knew that was a very, very conservative number because to do that you almost have to be in a lab environment and under perfect conditions. In addition to that, we are purging those cavities with helium.”
As it happened during the Sept. 21 tanking test, the high hydrogen concentrations went away when the flow was stopped with the initial seven percent leak rate, and when the pressure and flow were more gradually raised as a part of the contingency procedure, the leak rate stayed under one percent for the rest of fast fill. NASA is looking at using the short-duration higher concentration limit if they have to resolve another leak during the next launch attempt.
Looking to put it all together
The launch team has covered all the ground-controlled phases of the countdown at different points in different Wet Dress Rehearsal (WDR) and launch attempts. Ground and vehicle issues scrubbed or abbreviated past attempts, but the Sept. 21 tanking test was the first time that all the Core Stage LH2 pre-launch loading phases were completed end-to-end.
After the leak in the fill and drain QD was resolved and fast fill of the Core Stage LH2 tank was established, thermal conditioning of the four RS-25 Core Stage engines was started. An LH2 engine bleed is established by flowing some of the propellant from the stage’s propellant tank through parts of all four engines and then the LH2 is dumped overboard through a four-inch bleed QD adjacent to the fill and drain on the TSMU.
A leak at the bleed QD during the June Wet Dress Rehearsal countdown forced that test to be abbreviated, which prevented the entire LH2 engine conditioning sequence from being demonstrated. Issues with Main Propulsion System instrumentation during the first launch attempt on Aug. 29 also scrubbed that countdown before all the tests could be completed.
During the Sept. 21 tanking test, the hydrogen concentration in the umbilical cavity went up again when the LH2 bleed flow was started. The launch team is taking as gentle an approach to the bleed QD, but with less options.
“There’s no real conditioning of the four-inch because we are operating to the engine [specifications], [but] it’s a lower flow rate, so from that aspect it’s still at the same level as we do with the eight-inch,” Clark noted. “The engine [people] want us to start with a basically warm engine, which is driving us to keep that QD warm but still going at it with relatively low pressure.”
During the longer, early phases of the LH2 engine bleed, the leak rate stayed within the standard limit. After the Core Stage hydrogen tank fueling was completed and the engines had been conditioned long enough, a “pre-press” test of the terminal countdown sequence for the Core Stage and RS-25 liquid hydrogen systems was conducted.
Due to the bleed QD leak in the June WDR, the terminal countdown was run on the Orion/SLS vehicle with the exception of the Core Stage liquid hydrogen systems. Although the Core Stage has gone through its own, standalone terminal countdown sequence twice during the Green Run campaign at Stennis, the pre-press test at KSC allowed the launch team to look at the Core Stage performance with the Pad 39B and Mobile Launcher systems and should help to dial in some of those parameters for the next terminal countdown sequence.
During the terminal countdown sequence, the Core Stage tanks are pressurized for flight and the LH2 engine bleed increases to a higher flow rate for essentially the last four and a half minutes before ignition to help get the engine inlet temperature and pressure into their required ranges for start.
“Shuttle had a [hydrogen recirculation] system that we would use during loading and condition the engines that way,” Clark noted. “It was a two-minute terminal count [for hydrogen on Shuttle], but they actually needed the longer bleed flow [for SLS] because they didn’t have recirc to condition the engines throughout replenish.”
Artemis I on the pad ahead of launch on Nov. 16. The 2-hour launch window opens at 1:04 am Eastern.
— Michael Baylor (@nextspaceflight) November 13, 2022
“We are still conditioning them, and we are still flowing through them; it’s a higher flow rate, but they needed that four and a half minutes minimum at flight pressure flowing through the engines.”
In the Sept. 21 test, the hydrogen leak rate went above the standard limit; that was allowed in the test, again for a short-duration, and the leak rate went down on its own.
On the next launch attempt, NASA will be looking to put everything together. They have run through all the phases of launch countdown to the point of handing off control from the ground to the SLS flight computers in bits and pieces, but to launch Artemis 1 they need to do it all together.
If they can apply all the lessons learned, then they’ll get back to pathfinding again. SLS would get control of the countdown again at KSC, and for the first time, the Autonomous Launch Sequencer (ALS) in the flight software would get to run the final 30 seconds of the countdown with a launch ready vehicle.
At present, teams are readying the Artemis I Orion spacecraft and SLS rocket for liftoff at the beginning of a two-hour launch window that opens at 1:04 AM EST (6:04 UTC) on Nov. 16.
(Lead image credit: Artemis I on the launch pad. Credit: Stephen Marr for NSF.)