Testimony of Sven Grahn: Senate Science, Technology, and Space Hearing: International Space Exploration Program
Given at a Science, Technology, and Space Hearing: International Space Exploration Program
Tuesday, April 27 2004 – 3:30 PM – SR – 253
The Testimony of Mr. Sven Grahn , Vice President Engineering & Corporate Communications, Swedish Space Corporation
Distinguished members of the subcommittee, It is a great honour to describe SMART-1, Europe’s first space probe to the Moon which has been developed by the Swedish Space Corporation on behalf of the European Space Agency (ESA). The spacecraft weighed 367 kg (810 lbs) when it was launched by an Ariane-5 rocket on 27 September 2003. It is expected to reach an orbit around the Moon perhaps as early as in November of this year. In my statement I intend to concentrate on those aspects of the project where my own organization has contributed. The account is made from the perspective of the supplier, a quite small company in a small member country of ESA’s. What methods were used to permit us, a company of about 300 employees, to develop a sophisticated lunar probe of brand-new design in 39 months? That is the main question that I will address.
The Mission and its background SMART-1 is the first of ESA’s Small Missions for Advanced Research and Technology (SMART). Their purpose is to test new technologies that will eventually be used on bigger projects. The main mission objective of SMART-1 is the flight demonstration of electric propulsion for deep space missions. In early studies of SMART-1 a mission to an asteroid was considered. However, the piggyback launch opportunity selected put a strict upper limit on total mass and propellant mass. Also, a mission to an asteroid would require the use of busy and expensive ground tracking facilities because of the long distances involved. Therefore a flight to the Moon provided a solution to both these concerns. When the decision to fly to the Moon had been taken it was natural to include as much scientific instruments as possible. The tight mass limit provided an incentive for miniaturized instrument design – a bonus for later missions into the solar system. Thus the spacecraft uses a 68 mN stationary plasma thruster (PPS-1350 developed by the French company SNECMA and provided as a Customer-Furnished-Item by ESA) which consumes 82 kg of Xenon propellant to provide about 3.5 km/s of increased velocity that will bring SMART-1 from a geostationary transfer orbit to lunar orbit. The travel time will be in the order of 16 months. The final lunar orbit after capture is intended to be polar, between 300 km and 10000 km in altitude with the lowest point close to the lunar south pole. The Lunar observation phase will last for at least six months. In lunar orbit, the spacecraft will be pointed with one axis at the lunar surface for carrying out a complete programme of scientific observations from lunar orbit. The spacecraft carries a scientific payload weighing 19 kg which contains miniaturized instruments such as an imager for visible light and near-infrared light, an infrared spectrometer, an X-ray spectrometer and instruments to measure the effect of the electric thruster on the space plasma environment. Important science objectives of SMART-1 are to conduct lunar crust studies in order to test the current theories of the formation of the Moon, and to establish whether the large hydrogen deposits detected near the South Lunar Pole by the US Probes Clementine and Lunar Prospector, is indeed water. During the cruise phase to the Moon, experiments related to autonomous spacecraft navigation will be carried out using images from the star trackers and the miniature imager. ESA’s official cost figure for the SMART-1 project is 100 million euros at 2001 economic conditions (including spacecraft, launch, operations and part of the payload).
The spacecraft The spacecraft is designed with regard to the power needed for the electric propulsion, the severe radiation environment that is a consequence of the slow earth escape trajectory and the need for on-board autonomy. The design life of the spacecraft is two years. The spacecraft looks like a one cubic meter (35 cu ft) cube equipped with solar panels with a 14-meter (45 ft) span. Power is provided by a large solar array with almost 2 kW of initial power using highly efficient triple-junction cells and a 220 Ah Li-ion battery. The spacecraft’s attitude control uses reaction wheels and hydrazine thrusters for steering based on inputs from very compact star trackers and gyros. The spacecraft platform contains several new technology elements in addition to the electric propulsion. These elements are both part of the mission objectives and part of the answer to the question how a small company in a small country can build such a capable spacecraft. Autonomy was a major design driver for the spacecraft so that the long cruise to the Moon would not tie up expensive ground station time and operations staff. Therefore the avionics was entirely new and its architecture was designed so that on-board software could autonomously manage fault detection isolation and recovery.
The development task The project to develop the spacecraft lasted 6 years from the first contact in March 1997 between ESA and the Swedish Space Corporation until launch. After initial assessment, feasibility and definition studies the development contract was signed in December 1999 and the spacecraft was formally delivered to the customer after 39 months. The spacecraft was stored for a few months awaiting the Ariane-5 piggyback launch opportunity. The prime contractor team that managed and carried out the development of the spacecraft and several of its subsystems expended 280000 working hours to complete the spacecraft. In addition the team procured other subsystems and equipment under fixed price contracts with vendors. The prime contractor staff reached a maximum size of about 75 persons, including on-site outside consultants for specific development tasks
Previous experience of SSC as a prime contractor The Swedish Space Corporation designs, launches and operates space systems. We design and build small satellites, sounding rockets and subsystems for space vehicles. At our launch site in the far north of Sweden we launch sounding rockets and balloons and provide communications services to satellites with our extensive antenna facilities. The latest such support task was to NASA’s Gravity Probe B launched last week. The company was formed in 1972, has 300 employees, is owned by the government, and operates as a commercial corporation. The Swedish Space Corporation, at the time of its selection to develop SMART-1, had built and launched three successful spin-stabilized space physics satellites and was finishing the development of a small radio astronomy/aeronomy satellite with extremely accurate pointing capability (5 arc-seconds stability). This satellite, Odin, was launched in February 2001 and has performed brilliantly since then. The level of complexity of Odin is comparable to SMART-1. This made it possible for our company to at all contemplate taking on the development of SMART-1 when this task was offered to Sweden by ESA as part of a package for compensating our country for insufficient “industrial return” on its investment in Europe’s future in space. All our previous projects have been essentially multilateral projects under Swedish leadership. To develop these spacecraft SSC used a “skunk-works” approach in which a highly skilled small group of people is put to work with little outside daily monitoring and using only the documentation needed to build the product. “Peer” reviews of the technical work were used instead of formal reviews. Such an approach is often confused with the Faster, Better, Cheaper (FBC) paradigm. “Skunk-works” methods can be part of the FBC paradigm, but there is nothing in the “skunk-works” methodology that inherently assumes that higher risks will be accepted. For example; although we used military or commercial parts, tests and other measures were taken to convince us these parts would work, even if the analysis and test methods were unconventional.
Sometimes rigorous computer analysis was replaced by simpler “back-of-the envelope” analysis, but extensive testing on all levels was never cut back – rather the opposite. The first small satellite we developed actually was tested, almost fully integrated, daily for almost a year. In these projects low cost was emphasized as the driving parameter. In the FBC paradigm, as I understand it, higher risk is explicitly accepted. This was not so in our earlier projects. Instead schedule or performance could be used as “free” parameters. For example, the Freja magnetospheric research satellite launched in 1992 from China had a very flexible requirements specification which permitted costs and schedule goals to be met. For Odin, the sophisticated radio telescope-carrying satellite, schedule was not critical, but performance and cost was. By using a relatively small team (12 persons), the long development time did not cause excessive cost increases. Thus, our experience tends to confirm that you cannot get all the letters of FBC if you want to limit risk and have the assurance of low risk – you only get two out of the three – unless you add a new ingredient! The new ingredient to possibly resolve the FBC dilemma is smart technology and smart industrial methods. This is what we proposed to ESA for the SMART-1 project and which was in line with the Agency’s ambitions for the project. Thus, when ESA presented the spacecraft to the press in April 2003, Dr David Southwood, ESA’s director of science described the development paradigm for SMART-1 as “faster, better, smarter”. The outline of a truly industrial approach to spacecraft development Thus, SMART-1 was not developed within the “skunk works” paradigm but rather a “light” version of traditional system management methodology pioneered in the development of ICBM’s in the United States. Tight customer oversight of the supplier was used to provide a measure of assurance of low risk. However, ESA kept a comparatively (to other space science projects within ESA) lean staff of approximately 8 persons for day-to-day monitoring of us as the supplier. The monitoring staff consisted mainly of highly skilled technical specialists but also experts on management, project control and contractual aspects. For major project reviews the Customer used its normal level of resources with about 40 specialists spending 4-6 weeks examining every technical aspect of the project.
The contract type, cost-plus-incentive-fee (even with a negative fee!), was a way of keeping cost low (the risk to the supplier of developing a brand-new spacecraft with much new technology was not slapped on the price), but it also required much more detailed reporting of man-hours and other expenditures than for a fixed price contract. The organization within the Swedish Space Corporation that developed SMART-1 was the Space Systems Division based at the company’s engineering center in Solna, a suburb of the capitol Stockholm. This division has a total staff of 75 persons so the development of SMART-1 was a major task and indeed a difficult one especially in the early parts of the project when our previous working style had to be changed. However, some choices of technology and methods were worked out with the Customer early in the project that helped considerably in meeting the schedule without losing the assurance of limited risk.
1. Without trying to flatter ESA, a superbly competent customer helps any supplier, and it certainly helped a small company like the Swedish Space Corporation.
2. Commercial-off-the-shelf items from non-aerospace industry were modified for space use or used as-is, i.e. the CAN data bus developed in the automotive industry and a commercial real-time operating system. These items have been developed for commercial use by injecting massive amounts of human resources that is hard to match in the space industry.
3. Since a very large fraction of the spacecraft cost, perhaps 25%, can be related to software development, the most efficient developments methods available had to be used. These can be found in such fields as the mobile telecom industry where the requirements for short “time-to-market” for new products are extreme due to the cut-throat competition. Although so-called automatic code generation is not entirely new to the space business, it had not been used systematically in ESA programs. In SMART-1 we used this in the development of the attitude control software, the fault detection isolation and recovery software, and high-fidelity simulators of the spacecraft.
4. Standard software building blocks for spacecraft basic functions that were developed previously under ESA leadership were used and removed the need not re-invent them. In this way and by using commercial software building blocks (such as operating system) software development could be concentrated on the tasks specific to the SMART-1 mission.
5. Standardized logic circuit designs for implementing the international telemetry and telecommand standard is available through the efforts of ESA, both as ready-made circuits and as code for programming so-called gate arrays – chips that can be programmed to a certain task, for example to be a microprocessor.
6. The modern IT and telecom industry has created an extremely competent cadre of free-lance software engineers used to working in an environment where “time-to-market” and an industrial working style are primary values. This talent pool was tapped for SMART-1.
In five of the examples above one can see the outline of a truly industrial approach to spacecraft development, i.e. the widespread use of standard, well-tested building blocks permitting the developer to concentrate on product-specific work. ESA’s role in providing standard building blocks is reminiscent of the role of NACA in early U.S. aeronautics when this organization provided basic design standards such as airfoil profiles to the budding aeronautical industry. This is no revolutionary thought, but it needs to be applied systematically. In SMART-1 we tried to do this and we intend to continue along this approach. For a small company that cannot re-invent everything, this is the only way forward. One might say that space technology needs to “spin on” terrestrial and non-aerospace technology in order to be able to provide more “spin off”, i.e. technology that is spurred to perfection by the forbidding design environment that a space mission provides.
Concluding remarks We are indeed proud of our product, excited about working with ESA in advancing the state-of-the-art of astronautics and very flattered by the opportunity to share our experience with this distinguished deliberating body.
