Inside Artemis 1’s complex launch windows and constraints

by Philip Sloss

When the flight hardware and ground systems are finally ready for the inaugural Artemis 1 launch to the Moon, NASA will also have to synchronize the timing of the flight with unique celestial mechanics. The Orion and Space Launch System (SLS) Programs in the Exploration Systems Development (ESD) division are working together to calculate when the Earth, Moon, and sometimes the Sun are all in the right positions to support the agency’s requirements for this first joint Orion-SLS test flight.

In general, NASA will have daily opportunities to launch this first Artemis mission to the Moon in an approximately “two weeks on, two weeks off” pattern. On a particular day that has a launch opportunity, the Exploration Ground Systems (EGS), Orion, and SLS programs will have a launch window that will vary from a few minutes to a few hours.

Launch opportunity factors

The Artemis 1 mission to the Moon has had a long history of planning. Formerly called Exploration Mission-1, this first combined test flight for the three programs in ESD plans to send an uncrewed Orion spacecraft on a cislunar Distant Retrograde Orbit (DRO) trajectory. Now re-branded as Artemis 1, the mission will see the Exploration Ground Systems (EGS) program launch Orion on top of the first Space Launch Systems (SLS) vehicle from Kennedy Space Center.

SLS will place Orion on a lunar transfer trajectory that passes within approximately 100 kilometers of the lunar surface. From there, Orion will begin a series of powered flyby and insertion burns to place itself into the cislunar DRO. This type of orbit was a strong candidate for an asteroid redirect mission, which was proposed during the Obama Administration.

NASA is now planning on staging its orbiting cislunar infrastructure in a Near-Rectilinear Halo Orbit (NRHO). On future lunar landing missions, Orion and its crew will rendezvous with SpaceX’s lunar lander variant of Starship or the Lunar Gateway – both of which are in active development.

Although it is an artifact going back several generations of space policy, the DRO mission provides a setting to accomplish dozens of flight test objectives and system checkouts on this first solo Orion mission.

Credit: NASA.

(Photo Caption: The initial configuration of the Lunar Gateway, consisting of the Power and Propulsion Element and Habitation and Logistics Outpost modules. NASA hopes to expand the Gateway over time, with additional U.S. and international modules in the works.)

This flight to a DRO has its own set of launch constraints. “There’s just a handful of key drivers that we [use to] determine whether or not there is a launch opportunity on a given day for Artemis 1,” Mike Sarafin, NASA’s Artemis 1 Mission Manager, said in a September interview.

“The further you get into the manifest, the constraints will change because we need to account for things like lighting for [lunar] landing and docking and some other things, and so I’m just going to speak [about] Artemis 1. I would expect some of these to be in play for later missions, but there actually may be additional constraints.”

For Artemis 1, the opportunities tend to repeat about once every two weeks, with a launch period of about 10-15 consecutive days. This is then followed by about two consecutive weeks without a daily launch window before the next launch period begins.

NASA is tentatively planning to hold the different rounds of flight readiness reviews ahead of Launch Period 18, which runs from February 12 to February 27, inclusive. The next period, Launch Period 19, would run from March 12 to March 27, inclusive.

Sarafin explained that there are four major factors that determine whether or not there is a launch window available for Artemis 1 on a particular day. This is primarily driven by the available performance of the SLS Block 1 vehicle.

“We’re specifically trying to hit our Trans-Lunar Injection point, and there are days where the Earth-Moon alignment do not support launching with the Interim [Cryogenic] Propulsion Stage [ICPS] or the Block 1 vehicle,” he said. “So performance is our number one constraint.”

Credit: NASA.

(Photo Caption: An excerpt from an Artemis 1 mission availability document produced by the Mission Analysis & Integrated Assessments (MAIA) group within NASA’s Exploration division. The excerpt was shared on the NSF forums by a poster who obtained it through a Freedom of Information Act (FOIA) request. The revision of the document supplied was produced from February 2021 data and reflects Eastern time zone dates and times for launch. Analysts are performing a “rolling” analysis of launch periods and windows, so although the actual times aren’t explicit here, they may change by minutes or seconds.)

“[The] number two constraint is the Orion spacecraft should not exceed greater than 90 minutes of eclipse during any point in the mission, and that’s for power production and thermal reasons. We’re using solar arrays to produce power, the batteries are sized based on that as well as a few other constraints, and then also the thermal system.”

Within the 10- to 15-day periods that have a daily opportunity to launch, there may be a day or two “cut out” by the 90-minute maximum eclipse time constraint for Orion. “Occasionally [you] have this odd seasonal constraint that comes along where the Sun is behind the Earth during the outbound transit to the Moon or during the return transit that [eclipse time] exceeds 90 minutes. And just the angle of the solar eclipse from the spacecraft’s standpoint causes an odd day to be cut out every now and then,” Sarafin explained.

The third constraint is that Orion must splashdown in lighted conditions, which aids recovery operations. Artemis 1 will be Orion’s first re-entry from the Moon, providing a high-priority test objective for the Orion Program.

“For this uncrewed flight test, we’re trying to hit a lit landing, so that governs where we have at least a half an hour before sunset or we land at least an hour after sunrise, that drives some of our launch opportunities,” Sarafin said. “The last constraint is essentially the allowable range between entry interface and the splashdown target zone.”

“For a skip re-entry, we want to fly somewhere between 2,500 and 4,000 nautical miles [approximately 4,630 and 7,408 km] range from entry interface to splashdown, and we’re nominally targeting around 2,800 nautical miles [5,185 km]. If we get shorter than that, it’s going to exceed some of our test criteria, and if we get longer than that, again it’s outside of our test criteria as well as the vehicle design.”

“So performance, eclipse, lighting for landing, and then range from entry interface to splashdown, those are the four things that we screen against to determine whether or not we can launch on a given day for Artemis 1,” he said in summary.

(Photo Caption: A render of the Block 1 SLS on Pad 39B ahead of launch. The Artemis 1, 2, and 3 missions are expected to utilize a Block 1 SLS vehicle.)

The initial Block 1 configuration of the SLS launch vehicle utilizes the ICPS, a modified United Launch Alliance (ULA) Delta Cryogenic Second Stage (DCSS) used on their Delta IV rocket. The ICPS acts as an in-space second stage above the SLS’s five-segment Solid Rocket Boosters and Core Stage. The SLS boosters and Core Stage insert ICPS with Orion on top into a parking orbit that is unstable and highly elliptical, with a low end – or perigee – within Earth’s atmosphere but with a greater-than-usual high end – or apogee.

For Artemis 1, the SLS Core will be targeting an insertion orbit of 30 x 1,806 km.

“The ICPS does not have the capability to throw an Orion massed vehicle to lunar vicinity from a low Earth circular parking orbit, so the SLS Core Stage assists the ICPS/Orion stack with insertion into an elliptical parking orbit,” a NASA paper on SLS launch windows notes. “The apogee of that parking orbit is a dial that can be used to transfer performance from the Core Stage to the ICPS or vice versa.”

“Later, when we get the Exploration Upper Stage (EUS) [with] the Block 1B vehicle, we will not have that constraint to get to the Trans-Lunar Injection (TLI) point, and it’s because we have to loft the trajectory with the Interim Cryo Propulsion Stage to achieve the performance that we need to hit that specific point in space for Trans-Lunar Injection,” Sarafin noted. For future SLS launches using EUS, the currently in-development stage will insert into a circular parking orbit of about 185 km.

“The circular parking orbit allows the future configurations of SLS to send spacecraft to the Moon on virtually any day of the month,” the NASA SLS launch window paper notes. “This is necessary as other [future] mission constraints, like rendezvous in lunar vicinity, may restrict the launch availability.”

There are also other performance factors, such as the bulk temperature of the solid propellant in the SLS boosters, which are affected by local temperatures that vary from season to season. The boosters typically have better performance at higher ambient temperatures that occur around the summer months at the launch site in the Northern Hemisphere.

Credit: NASA.

(Photo Caption: From a NASA paper on Artemis 1 trajectory considerations when the mission was still called Exploration Mission-1, a graphic that shows one of the launch period constraints due to the SLS Block 1 configuration. The highly elliptical insertion/parking orbit used with the ICPS places some restrictions on when the mission can launch. Future SLS versions with the in-development Exploration Upper Stage can use a circular parking orbit that increases lunar transfer opportunities.)

In addition to the constraints on which days will have launch opportunities, the daylight splashdown/recovery requirement at the end of Artemis 1 will also dictate the duration of the test flight. If necessary, Orion can extend its stay in lunar orbit to ensure it splashes down in daylight.

“The mission does vary between what we generally term a ‘short-class mission’ which is about a 25, 26, 28 days or what we’ve generically termed a ‘long-class mission’ which is somewhere between 38 and 42 days,” Sarafin explained. “Within a launch period, we will switch based on the day that we launch from long class mission to short class mission, so it’s really dependent on the day that we go whether we’re short class or long class. And it all has to do with that three-body problem and the alignment [of the Sun, Earth, and Moon].”

The orbital period of the DRO is about 12 days, which dictates the length of Orion’s stay at the Moon. The overall mission duration will be between four and six weeks. Notably, both mission durations are longer than the spacecraft’s maximum limit with a full crew of four onboard – 21 days. Orion’s systems are capable of operating up to around 210 days in an uncrewed configuration, such as on Artemis 1.

Day of launch windows

For windows of opportunity on a specific day, the length of the launch window can range from minutes to hours. “We have ‘day of’ [launch] performance constraints, and it’s a bit like a bell curve or a sinusoid,” Sarafin said. “So if you’re early in the launch period, you have a relatively short or long launch window, depending on your set up coming into it, and then it will oscillate the further you get into the window.”

“Based on the data that I’ve seen, [the windows vary] anywhere from a really short window, like six minutes in duration where we meet those four constraints that I gave you before, or they go up to a maximum of a two-hour launch window that we could set up for on any given day.” There are days where the launch window extends for even longer than two hours but there are other operational and vehicle considerations that, at least for Artemis 1, will limit launch attempts to a maximum of a two-hour window on a given day.

“That’s a self-imposed maximum duration for a couple of different considerations,” Sarafin explained. “If we’re not going to make that day, we need to detank, inert the tank, and then reconfigure the vehicle and set up for the next opportunity.”

Credit: NASA.

(Photo Caption: Another graphic from the NASA trajectory considerations paper which shows how the launch windows vary from season to season. NASA does not have a daylight launch requirement for Artemis 1 and there are few daylight launch opportunities at Kennedy Space Center in the winter months. All of the launch windows in the February 2022 launch period are in darkness.)

Liquid propellant loading (and unloading) for the SLS vehicle is one of the main considerations outside of celestial mechanics.

“Because this vehicle is so big and we use so much liquid hydrogen (LH2) and liquid oxygen (LOX) [in both the Core Stage and ICPS], it takes a while to do that and then reconfigure and set up for the next attempt. So we’re preserving the option for a next attempt after a certain amount of time by creating a maximum of a hundred-and-twenty-minute window.”

The SLS Core Stage strained the cryogenic loading capacity at the Stennis Space Center for its one-off Green Run test campaign. When combined with the ICPS, it comes especially close to the limit of the existing LH2 storage sphere at LC-39B.

In a November 2018 tour of pad 39B, NASA Pad Operations Manager Ken Ford said that the two LH2-LOX stages in the SLS Block 1 vehicle could use up to 730,000 gallons of liquid hydrogen on the day of a launch attempt, which is calculated to include the maximum two hours of launch window hold time.

Ford also noted in the 2018 tour that the current hydrogen sphere can hold up to 850,000 gallons of LH2, of which 765,000 gallons are available to support a launch. That’s enough for a single launch attempt – but not enough to support another launch attempt the following day.

While the vehicle is filled with liquid propellant, some of the LH2 is consumed for engine conditioning and some boils off inside the tanks. If a launch attempt was to be scrubbed after waiting through a whole two-hour launch window, and the propellants in the tanks were drained back into the spheres, they would end up being over 280,000 gallons short of the LH2 needed for another attempt.

Credit: Brady Kenniston for NASASpaceflight.

(Photo Caption: SLS Core Stage-1 performs a test fire on the B-2 test stand at Stennis Space Center in 2021. During the Green Run, Core Stage-1 was put through several levels of systems tests, culminating in a several-minute engine firing. Core Stage-1 will be flown on the Artemis 1 mission in 2022.)

That deficit would be smaller for attempts that are scrubbed before using up the full two-hour hold time, but in the case of a scrub on the first Artemis 1 attempt, a 48-hour turnaround would be required. “For Artemis 1, we have the old hydrogen and oxygen spheres out at 39B, and we know that we can have three attempts over a seven-day period, strictly based on commodities,” Sarafin said.

“That assumes that you tanked, that you scrubbed in your window while you were fully tanked, and then had to detank, and then we would actually have to do a little bit of hydrogen top off using tankers to achieve the next attempt 48 hours later. And then assuming that we tanked on that one and then scrubbed after we were fully tanked, we would have a third attempt after 72 hours because we would have to replenish both hydrogen and oxygen out at the pad. So we could get three attempts in seven days with [that] replenishment plan.”

NASA already uses tanker trucks to replenish boiloff in the spheres at the pad, bringing them in from their vendor. In order to expedite replenishing the LH2 sphere at pad 39B, the agency has contracted with ULA to tap some of the LH2 stored in the hydrogen storage sphere at Launch Complex-37B that currently supports Delta IV launch operations.

EGS recently constructed a second, larger liquid hydrogen storage sphere at 39B to augment the existing LH2 sphere. However, it is not yet activated, and software validation has only just started. Because of this, it will not be ready to support SLS launch attempts until Artemis 2.

Sarafin noted there are other considerations for longer daily launch windows, at least for this first launch. “We also [limit windows to two hours] for preflight analysis purposes because the longer the window, the more cases that we have to analyze,” he said. The Orion, SLS, and ESD programs are jointly working on a flight readiness analysis cycle (FRAC) to define all of the parameters and considerations for the Artemis 1 launch.

Similar to computationally-intensive analytical modeling, the analysis cycles have to look at varying parameter combinations for launches and launch trajectories from minute to minute. The timing of launch periods and launch windows is one of the end products of the analysis, which computes the trajectory of the launch vehicle throughout its flight.

Credit: NASA/Ben Smegelsky.

(Photo Caption: The large, Apollo-era liquid hydrogen storage sphere at Launch Complex 39B is seen during verification and validation activities with the Mobile Launcher in November 2019. The two cryogenic stages of the SLS – the Core Stage and the ICPS – come close to consuming the capacity of the sphere in a single launch attempt. Therefore, NASA will only be able to make three launch attempts in a seven-day period for Artemis 1. In the lower-right background at the base of the sphere, a much larger, second LH2 sphere can be seen under construction by NASA for use beginning with Artemis 2.)

In addition to nominal and optimal trajectories, the analysis also looks at areas that include flight envelopes, performance boundaries, the open-loop guidance flown while the vehicle is in the lower atmosphere, engine throttle buckets, and abort margins and capability. After a full run-through of the process called FRAC-0, the analysis group is running numbers for FRAC-1, which will continue until the vehicle is ready to launch.

“We’re continuing to work the flight readiness analysis cycle,” Sarafin said. “Because of the lead time required, we actually do what we call a rolling analysis cycle, which rolls from launch period to launch period. So we’re assessing launch periods months after when we plan to launch simply due to the lead time required to get through that analysis cycle.”

“Our team has a plan where the analysis will be ready when the vehicle is ready, so we have a plan right now that accommodates that and also accommodates the lead time for this flight readiness analysis cycle. It rolls from launch period to launch period, so the analysis will be ready when the vehicle is ready, and that’s our macro strategy.”

There can be days within the Artemis 1 launch periods where the launch window is significantly longer than two hours. This could provide an opportunity to consider in real-time whether there is a “best” two hours for weather or other factors. However, NASA decided not to try that for this first SLS launch.

“We decided against a ‘pick ’em’ criteria where we have [for example] 200 minutes and we could pick 120 out of the 200 minutes to launch in,” Sarafin said. “It actually created more work from a preflight analysis standpoint with marginal [advantages].”

“We had a pretty healthy debate on whether to have a ‘pick em’ for the 120 minutes, to skew based on weather, to skew based on a whole host of other constraints, and what we agreed on was not to do that. We would select the 120-minute constraint mostly based on optimizing our performance to accomplish the objectives that we have set out.”

Credit: NASA.

(Photo Caption: An overview of the Artemis 1 mission, in which SLS will place Orion on a lunar transfer trajectory that passes 100 km above the surface of the Moon. Beginning with the lunar flyby, Orion will make a series of engine burns to first place itself on the Distant Retrograde Orbit trajectory and then to return to Earth.)

On the short end of the launch windows, Sarafin said there were no reservations towards making an attempt to launch Artemis 1 with a window that might be only a few minutes long. “Any day where we can launch and have an opportunity is a good day to try to launch,” he said. “If we meet all of our constraints, we would try to fly that day, so we would take it if we could get it.”

“We do not have instantaneous windows, [but] we [could] have windows that are single-digit minutes or tens of minutes [long],” he added. “We’ll take the first opportunity where we meet our constraints and any day where we can meet our constraints, regardless of window duration, is a good day to launch.”

The February 2022 launch period begins with an approximately 21-minute window on the 12th. The window duration increases to the maximum 120-minute long period for several consecutive days before decreasing to an approximately 42-minute long window on the 27th. The February period is one of the only ones without any cutouts due to the Orion eclipse or other constraints.

During other periods, there are days where the launch window is as short as one minute. One of the ways to optimize performance and maximize the duration of launch windows on a given day is to vary the inclination of the initial parking or insertion orbit. Shortly after liftoff, the vehicle will roll to a heads-down attitude and orient itself to a launch heading – or azimuth – that targets a specific orbital inclination. Once it is beyond the lower atmosphere, the vehicle’s guidance program will continually steer to stay on that heading and inclination.

“The azimuth that we’ll launch at is determined based on the time within the window that we go,” Sarafin said. For the SLS Block 1 vehicle, adjusting the orbit inclination was chosen over other possibilities for increasing performance. “There are several ways to improve overall vehicle performance, allowing TLI burns closer to perigee, including varying the target orbit apogee, perigee, inclination, or RAAN (Right Ascension of the Ascending Node),” the NASA SLS launch window paper explained.

“Apogee is used primarily as a performance exchange from the SLS Core to ICPS, so it is eliminated immediately. Perigee and orbit insertion altitude are used to control the SLS Core disposal in the Pacific Ocean, and yaw-steering to target RAAN can be very expensive performance-wise, so they, too, are not used. This leaves adjusting the parking orbit inclination.”

On days with launch windows that extend to the maximum of 120 minutes, the launch azimuth can change significantly. “At the beginning of the window it’s a high azimuth, middle of the window it’s essentially due-east azimuth, and then at the end of the window it would be a more southerly azimuth,” Sarafin said. “The azimuth ranges 46 degrees; it’s actually a pretty wide swath.”

(Lead image: The ICPS propels Orion on its trans-Lunar injection burn on Artemis I. Credit: Mack Crawford for NASASpaceflight)

Related Articles