Satellite reliability is a key issue in long term space sustainability because it has direct influence on the replacement lifecycle and also because, for orbits above about 600km, the end-of-life disposal success rate depends to a large extent on satellites retaining their command chain and manoeuvre capability. Whilst there have been large improvements in the like-for-like satellite reliability over the last few decades the same could not be said for human appetite for risks, each innovation and increase in complexity brings its own risks.
A group of failures pertinent to sustainability are those causing a release of smaller debris – small external particle collision, blankets coming adrift, short circuits and battery rupture. Such events could be anywhere from benign to catastrophic and can be caused by internal failures, space weather phenomena such as protons and charging or micro-meteoroid/orbital debris impacts.
When such an event leads to an anomaly it is either because the design precautions did not work or because the event magnitude was outside the specification and accepted as a risk. By their nature it may be hard to determine which of these factors caused a specific environmental event. Furthermore, it can be difficult to distinguish even the basic cause. Some types of electro-static discharge from electron charging, especially the highly destructive sustained discharge, can have a similar signature to some regimes of micro-meteoroid/orbital debris event, where they each involve both attitude and electrical disturbances, often linked by the generation of a local plasma. In such circumstances the operator and manufacturer themselves may not be able to determine the actual cause.
A long standing problem is that most events go unreported in any detail and satellite manufacturers and operators may be reluctant to connect the few big, publicly-known, events with space weather or the man-made environment; manufacturers can also be tied by non-disclosure obligations towards their customers. This is just a part of the landscape and so the data available to the public for scientific review is clouded.
There have been over 250 fragmentation events with a modest number of well documented cases and a significant fraction of recurring events involving launch vehicle upper stages. Beyond these there are many events with no clear distinction as to which are caused by on-board events, induced by space weather or otherwise, or by impact from orbital debris or micro-meteoroid events. Thus despite a large number of events there are researchers in each of these fields unable to make conclusions because of the ambiguity in classifying each specific event, though improvements in space situational awareness monitoring would help.
Another aspect of sustainability is the chance of a satellite surviving long enough to be passivated and de-orbited, or re-orbited after its useful mission. In recent years guidelines for disposal have included an element of post-mission disposal probability, or reliability. Alongside the observation that only modest numbers of LEO satellites appear to achieve the required outcome there has been discussion over the recommended disposal probability level and the method used in its calculation.
The idea is that the estimation of the probability of post-mission disposal will translate into a proportional success rate for the disposal of satellites. Such success rates are used in long term simulations of the future debris environment. Reliability calculations, in most contexts, are usually estimated prior to launch based upon parts count or parts-stress analyses. This method fundamentally assumes that the design is correct and the rates of failures of components are considered individually rather than in a complex system. The designers can influence the predicted reliability through redundancy, temperature management and choice of part failure rates. However, design weaknesses leading to systematic errors are effectively bypassed because it is not possible to model in advance a failure mode that has not yet been identified. The issues of electro-static discharge are a good example that do not feature in the reliability analysis as it can often be believed that the precautions are sufficient for it to be discounted.
However, many practitioners in the industry are well used to the notion that many failures are caused by a weakness either in design or in the application of the design through manufacturing. Even the distinction between these two terms can be blurred when an assembly error is a result of insufficient communication of the design principles into manufacturing instructions.
The pre-launch predictions for post mission disposal obviously also cannot make any allowance for the many work-around solutions that are achieved through creative adaptation once a satellite starts having multiple problems in-orbit. How much would either of these problems affect the actual long-term outcome of post mission disposal? Clearly, it is difficult to answer today.
Analysis of anomalies is still hard even for the manufacturer and operator. Failure investigations usually result in a “most probable” cause though sometimes that is simply the least unlikely of several difficult choices. Coming back to the issue of using reliability predictions as a yardstick for post mission disposal, the position for regulators then is even more difficult because of the poor public data and they may never see clear data which provides evidence for the success, or otherwise, of their strategy.
Thus we see parts of the sometimes obscure nature of reliability, failure assessment and end-of-mission management. Just the fact that it is a difficult area suggests that it would be a good step forward to recognise the issues so that a more conscious allowance for uncertainties can be made to support sustainability in space.
Header image credit: NASA