Murphy strikes: update on ATV-5 power problem

Update on the final activities to prepare ATV-5 for undocking and on the current issue affecting power, provided by Charlotte Beskow, from the ATV Programme team.

During the ATV-5 attached phase we have enjoyed a long period of ‘normal work’. A nice change from previous flights that were all quite ‘animated’. Make no mistake, ATV-5 has been a busy mission in terms of operations. There have been plenty of cargo operations, the standard activities of refuelling the ISS, oxygen discharge, air discharge as well as the occasional propulsive support. And on the whole, the vehicle has performed very well, (even the cabin air circulation fan is still running despite, our bets on it failing within 3 months).

The week was meant to be ‘quiet’

The first week of February was supposed to be the last ‘quiet’ week before undocking preparations, undocking, free flight and re-entry activities got under way: The ATV industrial design authority in Les Mureaux was busy preparing for the undocking and a planned shallow re-entry (a revised re-entry scenario which calls for some modifications to the usual procedures); the cargo people were busy ensuring that the trash loading stayed within the agreed CoM [centre of mass] envelope; and, the rest of us were reviewing our procedures and planning the logistics for the last few weeks of ATV (refreshing our memories of the undocking procedures, shift scheduling, meal orders, travel arrangements, post-flight data retrieval, material recovery, etc).

In ATV-CC Toulouse, the ground teams were preparing for ‘post-ATV’. This includes disposing of ground system hardware, removing equipment from the control rooms, etc. A sad task, but unavoidable at the end of a programme.

Then, on Tuesday afternoon this week, 3 February, things changed. ATV was again in visibility when the downloaded telemetry triggered a series of alarms, changing the quiet afternoon into a rather active one.

Quick analysis: ATV remains fully operational

A quick analysis of the data (which took longer than it takes me to write it) showed that one of the four power chains had stopped working. Some of the alarms were related to failure, others to the consequences of this shut down. ATV is programmed to react autonomously to a large series of failures and this is one of them. Extensive testing and validation was done back in 2005 – 2007.

Further analysis of the telemetry by the off-site team showed that the vessel’s reconfigurations had taken place exactly as expected. Redundant configurations had been activated and from a functional point of view, this failure is transparent, i.e. we can continue performing all planned ATV operations.

The optimists amongst us (99.9% of the teams) were quick to point out that Murphy had been ‘kind’ to us. He picked the ‘best’ possible power chain; neither the on-board computer nor the TDRS transmitters/receivers – which link us to data relay satellites used for ATV telecommanding – are affected. Also, it happened at the beginning of the week and during normal working hours (a rare treat; usually failures have occurred between Friday at 19:00 and Monday at 08:00…). There are no school holidays at the moment and all the teams are more or less in place for the final few, precious, weeks of ATV (yes, two individuals are out of circulation due to weekend skiing accidents – but nothing worse than a a broken wrist and a set of cracked ribs; and a few people are affected by severe colds – but that goes for half of Europe right now).

Consequently everyone got down to work. All the preparation done years ago, during the development phase, pays off (yet again!). There is a wealth of information available about various failures and their consequences. The ATV-1 testing had lasted many many months and the current design authority teams were heavily involved back then. Experts looked at the current telemetry and compares the ‘signatures’ to what was experienced, formulating hypothesis and elaborating a strategy.

It helps that we have, over the ATV years, continuously improved the tools and means available off site, allowing the experts to process ATV data in their offices (where they have access to their documentation and various analysis tools) almost in real time. A vast improvement on the means in place for ATV-1 back in 2008.

Testing hypotheses

By Wednesday afternoon, (24 hours after the failure) the first analysis results are on the table. It is not conclusive, i.e. a root cause has not been isolated yet, but there are a few hypothesis to work on. More investigations are necessary and some experts will come down to ATV-CC within the next two to three days to conduct them (including the one with the cracked ribs…).

In parallel, all teams (flight control, design authority, engineering support) are preparing for the case of undocking with only three out of the four power chains available. It can be done, no doubt about that, it is just a matter of double (and triple) checking that all the procedures and coordination run smoothly since the undocking sequence is extremely tight with respect to timing, and there is quite a lot of flawless coordination required between the control centres in Moscow and Toulouse and the ISS crew before the second set of mechanical hooks, holding ATV to ISS, can be opened.

Planned undocking remains 14 Feb, eight days from now. If we miss the timing in the tight sequence, then undocking has to be delayed. This has already happened previously, and we know that it is better to avoid it.

This morning, ATV-CC will run a simulation of this off-nominal undocking scenario. They ran one a few months ago ( the exact same power chain was ‘failed’ in that simulation, by coincidence) but it is wise to refresh the procedures for everyone.

We have busy days ahead of us!

–Charlotte