Capabilities on the cusp: the impact of a responsive, flexible launch challenge with no winnerby Todd Master
|
| Many viable companies had no significant need for government development funding, as they already had private investments, and intended to retain all their intellectual property—often a point of contention in government-funded development. They sought government contracts, or any other form of government recognition, as it would aid them in attracting additional private funding. |
At the time, there were several dozen launch vehicle developers at various stages of technical and schedule readiness. These companies were promising to develop approaches consistent with what DARPA had been after for the better part of three decades: systems that could launch on short notice from anywhere on the planet, enabled by “clean pad” designs that required minimal or easily transportable launch infrastructure, and advances in autonomous flight termination systems. It was unclear who among these companies could and would deliver (see “Responsive launch is still not quite ready for prime time”, The Space Review, March 9, 2020.)
After extensive conversations in 2017, and bringing together industry and forward-thinking government stakeholders to discuss how to bring our goals to fruition, common threads emerged. Many viable companies had no significant need for government development funding, as they already had private investments, and intended to retain all their intellectual property—often a point of contention in government-funded development. They sought government contracts, or any other form of government recognition, as it would aid them in attracting additional private funding. They desired access to government test facilities and ranges at government pricing.
Multiple new launch vehicle developers assured us that they would launch their first missions by mid-2019, and that they would only need one to two days from setup through launch. We digested these inputs, and created something that would be beneficial to both industry and the Defense Department.
The genesis of a prize-based challenge to demonstrate flexible and responsive space launch
The feedback from our outreach led us to the conclusion that the conditions were ripe for a DARPA Challenge to address flexible and responsive space launch. The industry already had significant investment, a broad technical base, and a myriad of significantly different approaches to solve the same problem: delivering a payload to orbit in short order with minimal infrastructure and minimal pre-launch knowledge. An adaptation of their commercial product was directly applicable to near-term military goals, and as such, the winner of such a challenge would have a major leg up in future military procurements, giving companies the upside potential that could allow them to take the financial risks needed to participate in the Challenge.
DARPA leadership approved the prize competition with two critical stipulations. One was that DARPA would have no role in safety oversight or indemnification—we would rely on existing FAA regulations and commercial insurance. The other was that it would be completed in two years from the time that it was approved in early April 2018. The first of these stipulations was put in place with consideration of some hard lessons from the DARPA’s 2006 FALCON effort and work with SpaceX at Kwajalein. And with a fast turnaround, DARPA Launch Challenge now represented a new approach for space launch for the agency: a low-stakes effort that could result in a big payoff.
Challenge ethos: be safe, be fair
We created Challenge goals to be aggressive but achievable, and most importantly, relevant to military users. Responsive space could be a tactical enabler but only useful if capabilities could be delivered within a timely fashion.
| With a fast turnaround, DARPA Launch Challenge now represented a new approach for space launch for the agency: a low-stakes effort that could result in a big payoff. |
We debated internally what the right metrics should be. Should we require competitors to get to T-zero in a countdown test, or actually launch to orbit? How many times? How quickly? What was relevant to real-world tactical military use? What was achievable but impactful, given the current state of the art? What were the factors beyond our control that would drive schedule? What were the regulatory requirements? Could we literally launch from anywhere, and who would be the deciding authority? We considered all of these questions in our formulation. We took lessons from prior Challenges, incorporating two overarching principles into Launch Challenge goals from the start: be safe and be fair.
The scope of most challenges at DARPA required little interagency collaboration, as they were largely self-contained. This one was going to involve multiple government regulatory agencies: the FAA, the FCC, NOAA, and a slew of different safety organizations across the Navy, the Air Force, and NASA. We were biting off a lot, but we had the benefit of DARPA’s “brand” within the government on our side, and I cannot overstate the willingness of other agencies to bend and flex in order to support our mission of high-risk, high-payoff technology investment for breakthroughs in national defense.
We considered potential competitors’ stated capabilities, and balanced those against what would matter to military users. We considered factors that may be beyond their control during a launch campaign (e.g. range availability, air traffic, weather) and made accommodations in the Challenge timeline to allow for those. We examined the range and spaceport required timelines for integration, statutory requirements for government agency coordination and public notifications. We developed a set of guidelines—broad strokes of metrics and execution plans—that we later refined into a specific set of rules as we learned more about the regulatory flexibility we could get, capabilities of our potential sites, and timelines and technical characteristics of our competitors. We decided to require two launches in rapid fashion from two different locations, with as minimal pre-launch knowledge of technical parameters as possible.
Creating the right prize amount was critical to the Challenge. We had budget constraints to work within, and an intention to incentivize modification of commercial products and processes for our specific use case. We did not intend to duplicate industry funding, but desired to provide a compelling enough prize to get an interesting field of competitors. Typical prize competitions are not a cost recovery mechanism for competitors, but afford winners widespread recognition as top in their field, attracting both investment and customers.
In deciding what warranted a prize, we wanted the bulk of our funding to go to something that was directly relevant to a warfighter: capability on orbit, nothing less. We ultimately decided that one launch would meet a subset of our goals, but that a second launch was really required to show the full breadth of flexibility and responsiveness. We created a discrete qualification prize of $400,000, intended to offset some of the costs of applying for an FAA launch license. The real prize purse was allocated to successful launches, with a modest prize of $2 million for Launch 1, and the maximum extent of the Defense Department prize authority of $10 million for Launch 2. Our goal was, across all prizes, to cover opportunity costs for two launches for the median pricing in our initial expected field of competitors.
For the Challenge, we intentionally chose to use the commercial regulatory process governed by the FAA Office of Commercial Space Transportation, both for spaceports and launch vehicles to the maximum extent possible. In a future where space launch will be used as a logistics service, the Defense Department would be best served as procuring these services on a commercial basis, rather than providing in-depth technical oversight and management of those services. When we ship packages by UPS or FedEx, we don’t ask these service providers the details of their trucks and aircraft, their maintenance histories, or their parts and sparing plans—we simply pay for the delivery service to be timely and accurate and leave the rest to the shipper. That is how we would like to use space launch for on-orbit delivery.
Finally, we had to select sites spanning the range of our competitors’ approaches. As we had applicants considering horizontal launch systems—using an aircraft as its launch site or, for some, their first stage—we had to consider sites that could support both their approach as well as more traditional vertical launch vehicles. Within the United States, we considered all existing and projected commercial spaceports, military test ranges, and traditional space launch ranges. On the horizontal side, there were commercial spaceports in spades. On the vertical side, most were co-located on federal ranges, with the notable exception of Pacific Spaceport Complex - Alaska (aka “Kodiak”). This would ultimately have a big impact on our execution.
Too many cooks in the kitchen: who decides what’s safe?
We had set up a confounding problem for our competitors: they would not know which launch site they would be using until shortly before the Challenge began. This meant that they had to simultaneously obtain approvals to launch from any of our candidate sites. Our intention in setting up the Challenge was to use FAA licensing at all of our launch locations, something we had top-level agreements on from leadership at all of our sites. When those agreements were made, the sites assumed that the teams would have a few launches under their belt before the Challenge, conditions that changed through our course of execution. While true commercial FAA licensing encompasses flight and ground safety, in the cases of launch sites located on federal ranges, those ranges have responsibility and oversight for ground safety—a natural division of responsibility.
| Unfortunately, obtaining federal range approvals for agreed-upon waivers within the Challenge timeline became a bridge too far, primarily in consideration of the fact that we were working with never-before flown launch vehicles. |
US government space launch ranges (both military and civil) have traditionally launched rockets designed with either insight or oversight of the US government as one of their customers. Bringing a new launch vehicle on to their ranges that they had never seen before, with no government personnel involved in its development, represented new territory. As such, the federal ranges struggled with constraining their role solely to ground safety, first dipping their toes, and then diving in, to flight safety elements of the vehicles. With three federal ranges from three different federal agencies (NASA, Air Force, and Navy) with three different sets of rules, this quickly emerged as a potential problem.
We brought all three agencies together, along with representatives from the FAA, to review their collective requirements and come to common agreement on how they could be met. These agencies already collaborate in the Range Commanders Council (RCC), which sets top-level guidelines for safety implementations that are then trickled down into agency-specific policies. These agencies agreed to evaluate the participating launch system at the RCC level of documentation, not the agency-specific levels. This was a great collaboration and perhaps the first time there was a direct forcing function to do so. The collective group came up with a common understanding as to what portions of their guidelines were critical, what didn’t apply, and what could be waived.
Unfortunately, obtaining federal range approvals for agreed-upon waivers within the Challenge timeline became a bridge too far, primarily in consideration of the fact that we were working with never-before flown launch vehicles. As such, we effectively eliminated federal ranges from further participation in the Challenge.
We quickly looked at how we could still accomplish our goals with the remaining non-federal launch site candidate, the commercial spaceport in Kodiak. We ultimately concluded that we could stress our intended “hard problems”—short notice of payloads and orbits, minimal infrastructure, limited time to set up and launch—by using a separate launch location within the Kodiak spaceport. Moving 300 meters from our first launch location retained the same technical challenges as moving to a different launch site 5,000 kilometers away, minus a lot of planes, trains, and automobiles.
The great paring down: where rocket science meets timing
We started with 55 interested companies, pared back to 18 companies through our first round of information gathering, and then further narrowed to three competitors: Virgin Orbit, Vector, and Astra. All three passed Challenge qualification, satisfying DARPA that they had credible approaches that could be delivered in our timeframe and receiving license application acceptance from FAA/AST as “complete enough.” Based on their projections in mid-2018, these teams all expected to be launching to orbit within the next 12 months, meaning the challenge would be right at the beginning of their commercial operations.
The Challenge timing was crucial, providing incentives right at the critical juncture of their development where they could make minor course adjustments to stay both “flexible” and “responsive,” before they became too entrenched in a business model that just favored one or two fixed launch locations with a large customer backlog to fulfill.
| We’ve been asked if we would have another Launch Challenge. The answer is no, and for a good reason. |
Ultimately, development timelines pushed all of our competitors’ schedules to the right—not an uncommon reality in new launch vehicle development. As a result of this, by the time we got to Challenge execution, we had just one competitor remaining that was both ready and willing to launch in our timeline. Astra took a business risk, shifted their intended public engagement timeline, and opted to execute its first orbital launch attempt in the public spotlight of the DARPA Launch Challenge.
BYOI (Bring-your-own-infrastructure): creating a launch site from a concrete slab in days
In Launch Challenge rules, competitors were allocated four days to set up their launch system. This meant creating and checking out both a complete launch infrastructure and vehicle. We intended this time for arriving, unpacking, checking out internal company network connectivity, and completing FAA and range-mandated rehearsals and contingency response procedures. The rules then allocated 14 days for launch attempts, which included accounting for delays for things like “red” range days due to bad weather or communications issues, mandatory crew rest requirements, and technical issues to be resolved. We made assurances to our teams that if they arrived on time, they would be guaranteed no fewer than four “green” days within the span of the campaign.
We knew our teams had initially said they needed one to two days from launch, and we wanted to afford them chances to be successful, remaining mindful that a timeline that moved from days to weeks to months would be less and less relevant to military users. Getting 14 days of air traffic closure coordination in advance was no small feat, and required a lot of negotiation with FAA. The impacts of the current way we enact closures for space launch to commercial business for both air and maritime are significant, and we flexed to the maximum extent realizable. We had some distinct advantages here at Kodiak, which had fewer air traffic routes to work around as well as the added safety benefit of sparsely populated areas around our launch corridor.
We informed our remaining competitor, Astra, that our desired launch date was February 17. This meant that they could begin setup as early as February 13, with range availability for launch four days later. Astra was completing its final testing on their first orbital vehicle, “Rocket 3.0,” right up to (and through) the start of our window, ultimately shipping the vehicle by air transport on February 18. They had pre-shipped some of the ground infrastructure on February 14, and their team was busy preparing that for the rocket’s arrival. Our rules did not require them to start on our start date but did not afford them extra time if they started after our initial date.
Astra worked through launch vehicle and ground system checkout, erecting the rocket on the launch site within just four days of its arrival. After pre-launch preparations (and a quick turnaround securing of the vehicle for a blizzard), on February 29 (day 13 of 14), Astra proceeded into a launch countdown, which was ultimately scrubbed due to weather conditions. March 1 was a “red” weather day, and despite having been given their allocated four “green” weather days over the course of the 14-day span, we opted to give them an additional launch opportunity on March 2. Astra successfully completed the majority of their launch countdown, and after a guidance sensor glitch at T-minus 53 seconds, they scrubbed their launch attempt for further troubleshooting.
Flexibility and responsiveness on the cusp: What we did, and what we learned
Astra successfully demonstrated that they had a flexible approach to space launch that allowed them to set up and ready a liquid-fueled vehicle and its entire supporting infrastructure within a timescale of days. However, the hardest part of the Challenge was delivering to orbit—something that was not yet demonstrated.
We’ve been asked if we would have another Launch Challenge. The answer is no, and for a good reason. Our intention through the Challenge was to inspire developers to create flexible and responsive approaches to space launch. During Challenge execution, we saw one team, Astra, demonstrate an approach that met our intention, with a rocket that appears on the cusp of capability. We saw two other teams whose architectures were well suited for the problem, with Virgin Orbit seemingly on the verge of executing their first launch as well.
| We have lots of horizontal spaceports with few vehicles in development, and one true vertical commercial spaceport with many vehicles in development. |
These developments are encouraging to DARPA, and we believe we can take the combination of what we’ve learned on the process side and effectively combine it with these launch vehicle systems. Through our relationship with US Space Force and US Space Command, we have come to a shared vision of the importance of flexible, responsive launch, and we’re working to see how we can collaborate on future demonstrations in this mission area.
This requires a great deal of further discussion, but in the interest of brevity, here’s what we learned:
- The nation needs more commercial vertical spaceports that are not co-located on federal ranges. Launch companies need the certainty provided in FAA regulations, which allows the simplest path for transportability of licenses across spaceports. Federal ranges allow for flexibility through tailoring and negotiation, but that expends valuable time and resources that drive cost and schedule. We have lots of horizontal spaceports with few vehicles in development, and one true vertical commercial spaceport with many vehicles in development.
- We need to establish clearer lines of responsibility on launch approvals, and eliminate redundant functions between government agencies. Federal ranges that host commercial spaceports need to accept the limits of their responsibility, and avoid conflating safety with mission assurance.
- We need a real-time, data-driven approach to allowing integration of launch events with air and maritime traffic. We currently create broad hazard areas that impact airlines and maritime cargo and fishing operations for extended periods of time without regard to actual liftoff times or launch events, resulting in tangible financial impacts. These have been manageable with our current infrequent space launch cadence, but that time is rapidly coming to an end. These impacts can be reduced significantly with real-time data integration from spaceports and launch systems.
- We need further investments in autonomous flight termination approaches and space-based telemetry systems. Traditional approaches in widespread use today tether us to ranges and require time and resources to integrate, limiting both flexibility and responsiveness. Autonomous termination systems have been proven on a limited number of launch vehicles, but they must become the norm, not the exception. Space-based telemetry would allow further reduction in range requirements, promoting the ability to launch from truly anywhere and maintain situational awareness throughout flight.
- We need further investment and process changes in rapid mission planning. Current timelines for trajectory and dispersion analyses to be completed and reviewed are on the timescale of several days to weeks. These timelines will not support true responsive needs.
- We must consider how we can build satellites to favor robustness and ease of launch integration instead of dedicating every kilogram to maximizing performance. We used cubesat-class spacecraft on Launch Challenge, which did not require unique coupled loads analyses. This was a big enabler of our timeline and flexibility. Had we used larger, more complex spacecraft, the processes and timelines for coupled loads analyses would have quickly negated all of our plans. Shifting approaches on this will require new design ideas and a recognition that speed of launch integration to enable fast delivery to orbit is a unique and necessary quality.
The Launch Challenge provided critical information that allows DARPA to proceed with future programs based on knowledge of what is working within the space domain, level-setting industry’s abilities, and what we need to focus on next. We will work directly with our transition partners across the Defense Department space enterprise to understand the results that we achieved, and drive the nation to make near-term use of flexible and responsive space launch to retain and expand America’s military leadership in space.
Note: we are temporarily moderating all comments submitted to deal with a surge in spam.