Interstellar Maintenance
In engineering, maintenance is an essential topic not only during operation, but also when planning it during the design process. Maintenance has to be considered as early as possible in order to allow for the correction of errors and for the repair of hardware as easily and as fast as possible. For complex technical devices on Earth, a detailed maintenance plan is typically developed to fulfil this aim. When talking about spacecraft, the approach to the maintenance plan is somewhat different. While for manned space missions the maintenance of a spacecraft can be carried out by astronauts, this is not possible for unmanned missions. Often the cost of maintenance and service for space missions is actually higher than the cost of losing the functionality of the subsystems or even the entire spacecraft. Regarding the rather short lifetimes of space vehicles, which is in most cases only a couple of years, maintenance simply is not required because the subcomponents have sufficiently long lifetimes. For a spacecraft such as Icarus which would have a flight time of about 60 years, the question is not if on-orbit-servicing and repair is required, but how it shall be implemented in the spacecraft.
The only known spacecrafts which had, and still has a significantly long lifetime without having to be repaired, are the Pioneer and Voyager probes launched in the 1970’s. The maximum lifetime of the Pioneer probes is 31 years (Pioneer 10) and 22 years (Pioneer 11). For the Voyager probes 1 and 2 it is currently 33 years and expected to become up to 48 years until the end of their functionality in the 2020’s. The limiting factor for these extremely long missions is not the fuel, as one could assume, but the electrical power supply. Because they are so far away from the Sun, these probes cannot use solar panels. To be independent from external sources, they draw electrical power from radioisotope thermoelectric generators (RTG) which obtain their power from radioactive decay.
Even though the lifetime of these early interstellar probes comes very close to the projected lifetime of Icarus, they cannot directly be compared. As Icarus aims to fly a distance of about six light-years, it will undergo far greater forces than the Voyager probes which, to date, have ‘only’ covered a distance of about 0.002 light-years (Voyager 1) [1]. For Icarus, these external forces would occur due to the higher acceleration of the vehicle and the unknown environment through which it would fly, e.g. radiation and particle bombardment. Additionally the thermal loads on the mechanical structure must not be ignored. Between the engine and the payload, huge temperature differences will appear which would need to be compensated for throughout the structure. Furthermore, all moving parts on the vehicle, such as steering mechanisms for the propulsion system as well as instruments on the payload side, would be sensitive to external forces. Even more critical than these static loads would be the dynamic forces due to vibrations from the engine. Summing up these factors, the life of Icarus would be much harder than for the interstellar Voyager probes and therefore a completely new challenge concerning its design and the way of maintaining the vehicle in an operational state over a lifetime of about 60 years.
Typically, systems with the projected complexity of the Icarus spacecraft receive regular maintenance and repair services where the most stressed components are checked and replaced. For such systems on Earth we have a comprehensive support of repair facilities for each type of hardware. If we look for comparable analogues to spacecrafts such as Icarus on Earth we find that aircrafts have very similar level of complexity. The current operational lifetime of civil aircrafts is about 15-20 years, but only if they are regularly maintained. The mean time between minor checks and maintenance for civil aircrafts is about 300-600 flight hours (A-Check), which is about 1-2 months. Additionally they are checked daily after each cycle of take off and landing (Trip Check) [2]. Of course one has to consider that these systems are manned, opposite to the unmanned spacecraft Icarus, which necessarily leads to more stringent safety regulations. However, the challenge to design a spacecraft with a lifetime of 60 years without access to human repair services, compared to designing an aircraft with a lifetime of 20 years including monthly checks is obvious.
Concluding these thoughts, it seems inevitable to implement on-board-servicing devices on Icarus. But not only the repair devices themselves have to be considered, but also the spare parts which have to be transported additionally. As a rule of thumb it can be assumed that one-sixth of the total components should be capable of being replaced (spareable items) and one-sixth of the spareable items should be repairable off-line (repairable items). These numbers depend on the equipment which has to be removed from service and the depth of repair which has to be performed. By modifying the rule of thumb for the application to Icarus, we can assume one-tenth of the items being spareable and, as before, one-sixth of the spareable items being repairable. Because these items can not be removed from service on Icarus, they have to be repairable in-situ [3].
To carry 10% spare parts on an interstellar mission is a huge additional effort, due to the problems of storing these items, and the additional mass of the spacecraft which further implies higher fuel requirements. The spacecraft would have to carry its own “repair shop” including a large “warehouse” of spare parts. The “workers” then have to be considered separately: In order to be available at any position of the vehicle and to transport the spare parts from the “warehouse” to the respective repair site, mobile service robots are required. Partly this task could be fulfilled using robotic arms mounted on the outside of the spacecraft as already done on the International Space Station and on the Space Shuttle. They could be installed near critical components, such as the payload section or the propulsion system of Icarus. With the help of rails these robotic arms could slide along the structure of the spacecraft and reach various spots of the vehicle. Making the entire vehicle accessible for these arms would require a complex system of rails and arms which leads to the need for autonomous mobile repair vehicles. These could be small service robots which have their own propulsion and power system and carry a universal adapter, e.g. a robotic arm, which connects to the respective interfaces at the repair site. In this scenario the spacecraft would then contain an entire self-service and self-repair system where any defective device could be fixed automatically with on-board resources. To be capable of repairing a large variety of components, the design of the spacecraft subsystems should therefore be modular and generic where mechanical and electrical components are designed using common basic structures which could easily be replaced. This modularity is also required for the pre-flight assembly of the spacecraft in orbit, so that this design aspect pays off twice.
In order to cover all features of on-orbit-servicing, the service robots would have to perform the following tasks: Remote and on-site inspection, docking, maintenance and repair. For the remote and on-site inspection, accurate manoeuvring and therefore a sophisticated attitude control system would be required. For the next step of servicing, an additional technology would have to be used by the robots: Rendezvous and Docking. Docking is necessary wherever remote inspection is not sufficient and where the robot locally has to provide the spacecraft with resources or repair its components. For this case, universal mechanical and electrical interfaces have to be implemented, where the robot can dock, receive diagnostic data and provide resources such as power, lubricant or propellant. In addition to designing the interfaces universally, the easy external accessibility to potential repair sites has to be ensured. If, for instance, an insulation foil has to be cut first to access a repair site, the service robots would have to be capable of repairing the insulation foil additionally, which is not necessarily what they would have been designed for in the first place.
In the past and present space missions we already find a couple of such on-orbit-servicing satellites and robotic systems. Although this technology is still young and not yet really established in common space missions (Spacecrafts are rather thrown away or left dead instead of getting repaired or reactivated), there have already been some successful precursor missions on this field and more are to come within the next years. The following missions are noteworthy in this regard:
- SNAP-1
- DART
- ETS-VII
- ORBITAL EXPRESS
- FREND (SUMO)
- TECSAS & DEOS
- OLEV
Icarus can directly benefit from these missions, as they are important technology drivers in robotics, Rendezvous and Docking and Attitude Control. The on-orbit-servicing technology is a logical step in the development of future unmanned spaceflight and moreover it is inevitable for interstellar spaceflight and hence for Icarus.
References:
[1] Wikipedia, Voyager 1 – Current Status, http://de.wikipedia.org/wiki/Voyager_1
[2] K. Engmann, “Technologie des Flugzeugs”, Vogel Verlag, 3rd Edition, 2006.
[3] T. J. Grant, “Project Daedalus: The Need for On-Board Repair”, JBIS, 1978.
[4] W. Hallmann/W. Ley/K. Wittmann, “Handbuch der Raumfahrttechnik”, Hanser Fachbuch, 3rd revised Edition, 2007
Tags: Interstellar Maintenance
15 comments
Hi Phil
What do you think of using rapid prototypers/fabbers to make parts? Potentially the broken components can be recycled and refabbed. Depending on the power available a fusion-torch might be used to reduce broken parts to base elements for reconstitution.
[...] This post was mentioned on Twitter by Richard North. Richard North said: RT @icarus_project: Interstellar Maintenance http://www.icarusinterstellar.org/blog/interstellar-maintenance/ [...]
That’s a good thought! Manufactoring parts onboard is basicaly a great idea, but it has to be investigated what is more efficient: carrying the raw material or carrying the spare parts. This surely depends on which parts to reproduce and how they are designed. The fusion-torch technology in my opinion is a pretty wild idea, but nevertheless worth to be mentioned and to be investigated further. I will have a closer look at this, thanks for the comment!
I appreciate this post. Very practical.
Elsewhere, people have discussed this issue with a presumption that you would need “wardens” with artificial general intelligence. I am pleased that you made no mention of this. But would you explicitly state that maintenance an repair would not need AGI?
I’m still not entirely convinced that you would need a repair system. I appreciate that you acknowledge that we already have functioning craft (which has had no maintenance at all) at more than 50% of the lifespan of a 60 year interstellar mission. Granted, there are thermal, vibration, and radiation insults on an interstellar mission which are present in much smaller amounts in our current missions.
However, I think that your comparison to maintenance and repair requirements of aircraft may not be entirely fair. Aircraft have probably thousands of parts which move on every flight. Additionally, their environment changes considerably during each flight. I would think that an interstellar craft would tend to have a near constant environment in deep space. Unlike flying through the atmosphere, interstellar space is dry, cold, doesn’t have wind forces, and there won’t be thousands of landings.
Can’t components be put through an entire life cycle’s worth of thermal, vibration, mechanical, and radiation exposure in a brief period of time? How accurate would such an approach be to predicting long-term outcomes? If such testing reveals failures of components at an average time, can those components be redesigned until they survive without repair for the entire duration of the mission?
And if there are failures for which the hardware cannot be redesigned for, wouldn’t some of the components be designed to have in situ redundancy? Obviously, we do this all the time for electronics. Can the same be done for mechanical devices? Also, how much can proper design reduce the need for repair? e.g. Certainly, some components would use harder metals even if they weigh more.
Also, re: the warehouse, if part of the problem is the radiation flux of traveling through space at very high speeds, how could the warehouse be protected including from secondary radiation?
Also, what are you presuming in terms of the portion of the flight in which the rocket would be turned on? Might it be for months instead of decades?
Also, how do maintenance requirements change if the duration of the interstellar mission were much longer (e.g. 2,000 years)?
If we could design a craft which didn’t need robotic repair, then this would greatly reduce the complexity and the number of points of potential failure.
Thanks for your comment!
I would not negate the need for a certain level of artificial intelligence for these repair devices. Due to the lack of communication on those large distances and the wide range of applications concerning repair, it would be reasonable to implement AI in order to maintain their flexibility and autonomy.
Concerning the comparison with aircrafts I must say that of course you can not directly compare both systems. But you have to keep in mind that Icarus will go on a journey with an extremely long duration and moreover with a mostly unknown environment. And you only have one shot to do it right. Just imagine what happens if this spacecraft would fail on its flight. Space is an uncomfortable environment. It is extremely cold, you’re right, but it can also extremely hot! There is radiation we don’t know and there might be objects crossing Icarus’ way which can very simply hit the vehicle. These are things you don’t have to take into account for aircrafts and Icarus will probably only have one flight within its entire lifetime!
Testing components for Icarus is an absolutely inevitable topic. Materials and mechanisms will be tested before flight – that is essential and a standard procedure. Of course the aim of the spacecraft design is to avoid onboard repair, using an appropriate level of redundancy. However, I think that there is no perfect design. Something unexpected will happen, and then it’s good to have the possibility to act and help a spacecraft which is lightyears away. But you’re right, the design of the spacecraft must aim for a perfect working system! Having onboard repair devices must not belie the fact that sophisticated design is much more important.
I agree with your opinion that if we didn’t have robotic repair, the system would be less complex. But there clearly is a trade off between designing your system with simple or multiple redundancy or with one additional subsystem which is capable of performing repair.
And there is another point: Within 60 years, a lot of things can change. Technology and as well the knowledge about our universe will develop, so wouldn’t it be good to have “somebody” up there flying with Icarus who could react felxibly on changes, perform our commands and maybe modify subsystems?
[...] the maintenance during interstellar missions The Project Icarus Blog addresses the problem of maintenance. This is another fundamental issue for an interstellar [...]
Hello, thank you for this interesting post. I have the feeling that maintenance is often tackled after the more prestigious sections like propulsion, avionics… It is good that you seem to integrate the maintenance strategy in the design loop itself.
But I have a question. The choice of the maintenance strategy depends on the reliability of the components, on the available technologies to heal or replace them and on mission parameters (performance, safety, reliability). How do you intend to make the trade-off since most technologies are still in development? In other words, how will you choose, for your reference case, between a semi-automated maintenance system based on, say, a robotic arm, and a fully intelligent maintenance system using an army of nanorobots interacting with the spacecraft at the microscopic level (to give two opposite examples)?
> a certain level of artificial intelligence
The craft is a finite system. A programmed expert system should be able to cover the bases, IMHO.
> Icarus will go on a journey with an extremely long duration and moreover with a mostly unknown environment
I think that interstellar space is a fairly known environment. Known in that we know that it is mostly very empty and that it is so all the way until it enters the Alpha Centauri system. We have some idea as to the density of hydrogen in interstellar space and, I believe, we have some idea as to the density of dust and micrometeorites. Certainly, if there were a whole lot out there then we would be able to tell this as we look through interstellar space towards distant galaxies.
> Space is an uncomfortable environment. It is extremely cold, you’re right, but it can also extremely hot!
When you get far enough away from the sun space is extremely cold but not extremely hot. The extreme heat of our inner solar system has been overcome routinely by engineers of all satellites.
> There is radiation we don’t know
I believe that we understand how much of the cosmic radiation comes from the sun an non-sun sources. If we know the density of interstellar hydrogen then I believe we would have an idea as to the amount of radiation flux there is when traveling at very high speeds. I believe that the space environment of the inner solar system is more hazardous than deep space. There is more radiation and micrometeorites within the solar system than outside of it. I could be wrong but I think that there is generally a misconception that interstellar space is more dangerous than the solar system space that we are accustomed to. But I believe that interstellar space is more empty of both objects and radiation than space near our sun.
> Something unexpected will happen
I don’t know that this is an absolute given. We have launched many interplanetary missions. Now, exclude those missions that did something which Icarus will not do. So, exclude those that attempted to enter into orbit around another planet or land on their surface. Also, exclude those by the Soviets and the earliest launches as we are now more technologically advanced. What you are left with are missions which have a fairly low rate of failure. Besides, if there are sensors on board the first Icarus and there is a failure and we are able to detect which component failed then that can be redesigned before Icarus 2 is launched. This works best if the failure occurs during the acceleration phase of Icarus 1 (i.e. within the first few years of Icarus 1).
> And there is another point: Within 60 years, a lot of things can change. Technology and as well the knowledge about our universe will develop, so wouldn’t it be good to have “somebody” up there flying with Icarus who could react felxibly on changes, perform our commands and maybe modify subsystems?
This is a very good point. At a minimum we should include field-programmable gate arrays which could be programmed by radio signal from Earth 4.3 years before Icarus encounters the star system. This could give Icarus the ability to use its existing cameras and other sensors in the best possible way to maximize science collection. I suppose this could be combined with a rapid prototyper to produce science equipment. But it might be tricky to anticipate what equipment and substances one would want in 55 years.
However, if acceleration is completed within a few years then advancing technology would not come quickly enough to affect either acceleration or maintenance of parts affected by the vibration, heat, or radiation caused by the rocket. Granted, the terms of reference for Icarus calls for some deceleration to increase encounter time. So components used for deceleration could benefit. But deceleration may or may not be necessary depending upon how fast the science can be acquired. We should be careful about increasing the mass on the front end in order to buy some time on the back end.
I’d really like to know how long Icarus will be in acceleration mode. This greatly affects the present discussion since so much of the insults (known and unknown) will come from the propulsion unit.
You really didn’t address the question of whether putting the craft through the total life-cycle of vibration, mechanical, heat, and radiation exposure could significantly retire the risk of failure.
Please see if Icarus can be made largely maintenance free and without complex sensor, maintenance, and repair mechanisms. We have a long history of maintenance free interplanetary spacecraft which have functioned for many years. Hopefully the lessons learned can be applied to Icarus in order to make it less expensive and so earlier to launch.
@ destop:
Thanks for your comment!
Yes, indeed you can hardly estimate how the reliability of components will improve in the future and which new repair technologies will actually be available at the time Icarus will be launched.
But we know the following: Icarus is designed with current or near future technology, which has proven to be feasible and to be working with a certain reliability. That makes it possible to estimate the reliability of the entire system and subsequently the required kind of repair and servicing devices.
Another point is that the design of Icarus is still in the very beginning of its development. At this status serious analysis of the system reliability would imply a large uncertainty. Up to now I would say that we will need a sophisticated mixture of both types of maintenance systems. Therefore it’s important to investigate which solutions could basically be feasible and then decide on the basis of a more detailed spacecraft design which solution is reasonable.
@John Hunt
> I think that interstellar space is a fairly known environment
We know roughly about the content of the universe and we know that there is something in between the galaxies: The interstellar medium which consists of gas- and dust-clouds. We don’t really know what will happen when Icarus passes these clouds. It is to be expected that a flight through this gas will heat up the spacecraft to a certain level due to the bombardment of protons and electrons and it might additionally cause radiation. When flying through dust clouds you can image what happens. Of course we have these objects also in our solar system, but Icarus will probably travel with lower speed within this region. When leaving our solar system, Icarus will reach velocities about 10% of the speed of light, which increases the risk of colliding with interstellar medium.
> Besides, if there are sensors on board the first Icarus and there is a failure and we are able to detect which component failed then that can be redesigned before Icarus 2 is launched
Sure we will learn from the failures in a first Icarus mission and apply this knowledge to the second version. But we shouldn’t reduce “Icarus 1″ to a “crash test dummy” for the next generation.
> I’d really like to know how long Icarus will be in acceleration mode
For Icarus we don’t know this yet. In the Daedalus mission there were two boost phases, the first lasting for 2.05years and the second lasting for 1.76years. The distance covered after the first stage separation is 0.05ly and after the second stage separation 0.2ly.
> You really didn’t address the question of whether putting the craft through the total life-cycle of vibration, mechanical, heat, and radiation exposure could significantly retire the risk of failure
A total life-cycle test for the acceleration phase would last for 3.81years which is a huge effort if you want to test vibration, heat and radiation impacts. I think it’s not possible to test the entire spacecraft as you can do it with satellites, but if you choose to test its components for a full life cycle test – why not? It’s not impossible, depending on the subsystem you would have to decide if it’s reasonable or not. We also assume that there will be some precursor missions before the launch of Icarus, so that intensive testing will also be available onboard these spacecrafts in a “real” environment.
> Please see if Icarus can be made largely maintenance free and without complex sensor, maintenance, and repair mechanisms
You are absolutely right, it is not the goal to, expressed with some exaggeration, design a fragile spacecraft but a perfectly running servicing system. But you have to consider servicing and repair systems – that’s what we do.
The service robots you describe seem more complex than necessary to me, particularly with their sophisticated attitude control and station-keeping capabilities. I don’t know if the intention is to throw away or repair damaged service robots, but at the least they would certainly need some kind of fuel.
I think robotic arms could fill all of the roles you mentioned without needing dozens or hundreds of arms. I have seen concept videos of robot arms with gripping devices ar either end. With the simple provision of handholds over the surface and the probe, these arms could “walk” across the surface from handhold to handhold. I can image such arms having a gripper on each end and the actual repair equipment in the center. This way each arm needn’t carry two identical sets of repair equipment on each end, and whichever gripper is not currently gripping the probe’s surface becomes useable as an additional tool in the repairs.
I’m no expert here, but I tend to believe that an all-eletrical robotic arm consisting of motors and tools would last longer and cost less logistically than a free-flying robot with a sophisticated propulsion system. The only disadvantage I can think of at the moment is that such arms would be less useful during the initial orbital construction of the probe.
[...] colleague Philipp Reiss has recently published a paper on this blog discussing the motivation for automated repair and some of the challenges associated [...]
I think the need for maintenance is being overstated here. Present-day satellites and interplanetary spacecraft probably have harder lives than any interstellar probe. They experience the vibration and static loads of being launched from the Earth’s surface, through the atmosphere. Then they have to fly only 1AU from Sol, experiencing large, asymmetric, time-varying thermal loads – especially bad for those that spend their lives orbiting the Earth every 90 minutes. If Icarus can survive those insults, the interstellar part of the mission should be fairly gentle by comparison. No acceleration above 1g, no thermal loads except for those generated within the vehicle itself. Once the first few years’ acceleration is done, it’s not much different from being in a deep cave on Earth: a benign and (most importantly) unchanging environment. Stuff can last a very long time under these conditions. The long lives of the Voyagers should be taken as an indication of what to expect.
It does make some sense to carry spare parts, but it makes no sense at all to carry a robot to install them. Better to have them already installed – i.e. redundancy. If a part fails, installing its replacement is then a simple matter of switching it on.
The only known spacecrafts which had, and still has a significantly long lifetime without having to be repaired, are the Pioneer and Voyager probes launched in the 1970’s. The maximum lifetime of the Pioneer probes is 31 years (Pioneer 10) and 22 years (Pioneer 11).
If I’m not mistaken, I believe Pioneer 6 is the longest functioning satellite. Launched in 1965, telemetry was last acquired in 2000 for its 35th anniversary. It’s assumed to be still functional today, which would make it nearly 45 years old.
I recall reading an interview with one of the designers who judged that the reason for Pioneer 6′s long life was its simplicity. Engineers chose to use tested components with known high reliability rates instead of newer components which promised better performance. Only enough integrated circuitry was included to operate the satellite systems; actual computation, when needed, was done by sending data down the DSN to be crunched on Earth, then relayed back . There’s something intriguing about the fact that many microprocessor-based satellites since have suffered catastrophic errors and spontaneous reboots while warhorses like the Pioneer 6 — literally built out of Minuteman missile parts — continue on obliviously for decades with no attention. One lesson might be that the most straightforward way of reducing failures is to reduce the complexity of the spacecraft. In practical terms, that might mean that beam-powered propulsion, where the powerplant remains at home and conventionally repairable, is going to be more likely to succeed than including the powerplant onboard.
At any rate, the feasibility study of the Pioneer 6 design gave it an 82% likelihood of lasting 6 months, so there’s something to be said, too, for the gap between theory and practice.
[...] on from the recent blogs by Philipp Reiss and Robert Freeland II, I thought I would add a little of my own perspective and experience, as a [...]
Leave a Comment