On October 7, 2008, QF72 was flying south over Exmouth heading for Perth when chaos broke loose in one of the aircraft's navigation devices. Following speculation that the huge military transmitter at  the Harold E Holt naval communications base could have played a part, I asked officials from the Australian Transport Safety Bureau...

Rural and Regional Affairs and Transport Committee
Supplementary estimates Tuesday October 21, 2008

CHAIR—I welcome officers from the Australian Transport Safety Bureau. Senator Ludlam?

Senator LUDLAM—I have a couple of questions all around the same subject. I wanted to get your sense of the incident that occurred on 7 October with a Qantas Airbus flight over Exmouth being forced to land. Would you give us a brief overview as to what stage your investigation is at and what your overall feeling for that incident is so far?

Mr Bills—Our director, Julian Walsh, will give you most of the detail because he has been managing it on a day-to-day basis. I was meant to fly on that aircraft actually from Perth to Singapore, but obviously the accident meant that it was diverted to Learmonth, as you said. We have had several media conferences concerning the accident and we have done so in close cooperation with a number of colleagues internationally, including the French BEA, who are our investigation equivalent; Airbus, the manufacturer; our own CASA in Australia; and we will be enlisting the help of the US NTSB, our equivalent in the United States when it comes to testing a piece of the equipment, the ADIRU, that seems to be most directly involved in the accident. In terms of the detail, I will let Mr Walsh speak, but we do not want to tell you things you already know. We are happy to give you as much detail as you like.

Senator LUDLAM—I am happy for a certain amount of detail because most of what I ‘know’ has come through media reports so I am actually more interested in hearing directly from you.

Mr Bills—That is excellent, because some of those have been a bit off the mark. I will let Mr Walsh give you an update.

Senator WILLIAMS—We understand that.

Mr Walsh—In terms of process and in terms of how the ATSB responded, we were advised of the incident during the afternoon. We made arrangements at that time for some of our officers from our Perth office to fly to Learmonth. We chartered an aircraft and they were up in Learmonth a short number of hours after the aircraft landed. At the same time we had staff who flew from both Brisbane and from Canberra over to Learmonth and they arrived the following day. Our onsite activities in Learmonth continued from the Tuesday evening through to the Monday. As Mr Bills mentioned, during that time we had seven ATSB investigators in Learmonth. We had representatives from Qantas obviously, representatives from the French BEA, the Bureau Enquêtes-Accidents—excuse my French, it is not all that good—and we had representatives from the Civil Aviation Safety Authority as well. At the same time, obviously, there were parallel lines of inquiry going on with people back in the office here. We have also had a member from Airbus attend our offices here in Canberra. He arrived a couple of days later and worked closely with our flight recorders and people and provided us with assistance as well. Again, as Mr Bills alluded to, we held a number of media conferences throughout that period to try to keep people informed of the progress of the investigation in terms of what we had found. Essentially, the aircraft was approximately 100 miles south of Exmouth. It was at 37,000 feet, the autopilot and the auto-thrust systems were engaged on the aircraft. When the crew received what they call an inertial reference system fault, which occurred on the No. 1 Air Data Inertial Reference Unit—you will hear us refer to an ADIRU, and that is this piece of equipment—this caused the autopilot to drop out, or disconnect, and essentially from that moment on the crew flew the aircraft by hand, although there was a very short period of time—a few seconds—when they re-engaged the autopilot. Essentially the aircraft climbed approximately 200 feet quite slowly and that was actually in response to just a very slight input by the pilot on the control stick. That was when the autopilot was re-engaged and the aircraft returned to 37,000 feet and then the crew was flying the aircraft manually again. It was then a number of minutes later that the crew received a range of other faults and warnings within the cockpit, and essentially the aircraft then nosed down and pitched down to a maximum pitch angle of about 8.5 degrees quite suddenly. The crew responded very quickly to react to that and to return the aircraft back to the normal flight. The aircraft descended about 650 feet during that process. From 37,000 feet it descended about 650 feet before it started to climb back up to 37,000 feet. That took somewhere about 20 seconds. It is not a phenomenal rate of descent. In fact, the rate of descent, if you put it out over a minute, is probably something in the order of 2,000 feet a minute, which is not really all that unheard of. I suppose the issue that has probably caused the problem is the fact that the pitch down has occurred very quickly, and that is what has caused the problems. Essentially, the aircraft returned back up to 37,000 feet and then a couple of minutes later there was once again another of these events, but it was not as severe and the aircraft only pitched down at the maximum pitch down angle of 3.5 degrees, and it descended about 400 feet before returning up to the cleared level. The crew then diverted the aircraft to Learmonth, and the remainder of the flight was essentially uneventful from an operational point of view in terms of the aircraft flying, and the aircraft landed without any problems at Learmonth.

CHAIR—I bet the passengers would not have felt that good if they had said that on the plane on the way home. Too late then, wasn’t?

Senator LUDLAM—Without prejudicing the investigation, which I believe is still afoot, what can you tell us to date about the cause of the incident?

Mr Walsh—What is apparent is that there does not appear to be anything wrong mechanically with the aircraft. The systems have checked out okay. What we have identified is a fault within this Air Data Inertial Reference Unit, the ADIRU, and in particular there was a range of frequent and high spike values in the angle of attack information that the ADIRU was sending out. The ADIRU basically has two components. It has an air data reference component and it has an inertial reference component. The air data reference component essentially receives information in relation to barometric altitude of the aircraft, angle of attack, temperature and one other parameter. I have forgotten what the other parameter is.

Senator LUDLAM—But it is a device that is telling the autopilot where the aircraft is and what—

Mr Walsh—Essentially, it takes that information and then the other part of it, which is the inertial reference part, computes positional information, and together that component feeds the information to a range of systems on the aircraft. It feeds that information to displays in the cockpit and also feeds that information to the flight management computers. Essentially what has happened in this situation is that this erroneous and high angle of attack reading within the ADIRU has been sent to the flight control primary computer and the flight control primary computer has responded to that thinking that the aircraft is in a high angle of attack. But of course the aircraft was still flying. The aircraft has tried to protect what it thought was a potential stall situation because of the high angle of attack and said, ‘This aircraft needs to pitch down to avoid that happening.’ So, essentially it is that erroneous message from the ADIRU to the flight computer that has caused this problem.

Senator LUDLAM—Are there more than one of these devices on these aircraft? Are there a number of redundant systems or is it just this one critical device?

Mr Walsh—There are three ADIRUs on the aircraft, and at the moment. That is obviously something that we are quite keen to understand. It is quite a complex system. Obviously, part of the investigation will be to understand the interaction between those units, what interactions there should have been and what protections there should have been perhaps within that ADIRU to isolate that erroneous data. They are certainly lines of inquiry that we are quite keen to pursue at the moment. Essentially, those ADIRUs have to go to the manufacturer’s facility in the United States for downloading of data and for more detailed examination. They are the only organisation that has the equipment and the facilities to do that. The normal protocol for doing that is for it to be done under our supervision, but because there are other interested parties, including Airbus, the BEA from France, the National Transportation Safety Board—our equivalent in the US—and US the component manufacturer, we will all be present during that testing. At the moment we are developing a testing protocol with all of those bodies to ensure that we are all agreed on exactly how this testing process should go ahead and to make sure that we do it carefully so that we do not lose any data while we are doing it. It is a matter of making sure that we have a very clear plan of how we do that before we physically start the work and, essentially, we all have to agree to that process.

Senator LUDLAM—I have discovered since starting this job that I am spending an appalling amount of time climbing on and off aircraft, and each time we are asked to switch off mobile phones and other radio transmitters, radios and so on. Is this device known to be sensitive to external sources of various frequencies of radiation?

Mr Walsh—No, I would not say it is. These ADIRUs are on a range of aircraft, including Airbus-type aircraft. There are also similar pieces of equipment, ADIRUs, in a number of Boeing aircraft—for example, the Boeing 777. They are commonly used throughout the world and this sort of reaction is certainly not something that is known as a result of that sort of thing.

Mr Bills—The fact that it appears that it is the ADIRU No.1 that has had these extreme values and the other two have not suggests that there is an issue with the ADIRU No. 1, but we do not know what it is yet. Clearly, we need to understand that and, as Mr Walsh said, we want to understand why it was that with one seemingly misbehaving in the way it did the other two did not take over in some way. Both things are very important.

Senator LUDLAM—That was where I was going before. As to the line of questioning about radio transmitters and so on, the aircraft at the time was 100 miles from the most powerful transmission station in the southern hemisphere, the Harold Holt ground station. Are you taking that seriously as a potential factor in the cause of the incident?

Mr Walsh—We do not close our minds to anything in an investigation of this nature. We really do need to be open to evaluate all of the facts and all of the circumstances surrounding an occurrence such as this. The aircraft was about 100 miles from that transmitter. We have done some preliminary calculations based on some indicative values as to the sorts of power outputs and wavelengths that we expect from that antenna. The amount of energy at the aircraft 100 miles away from that antenna is quite small, and we think it is very unlikely that that was a factor. Having said that, we will not dismiss it. We have already been in touch with the facility and we will get information from them in relation to the time of operation of that antenna, the power output, frequencies and so on. We will gather that information and we will look at it, but of course if we look at the ADIRU in the US and we find a definitive problem, of course it would be very nice if there was something quite clear that we find when we are able to do that testing.

Senator LUDLAM—Is that base operated by United States contractors or Australian personnel?

Mr Walsh—I do not know, but I think it is run by Australian personnel now. I have not personally been in touch with them, but the people who have been in communication have indicated that they are going to cooperate fully to provide us with the information that we need.

Senator McGAURAN—What base are we talking about?

Mr Walsh—It is the Harold Holt very low frequency transmitting station on the North West Cape.

Senator LUDLAM—Have there been other incidents of a similar nature in that airspace in the past?

Mr Walsh—No. There has been media speculation, again, about a Boeing 777 that had an occurrence back in 2005.

Senator LUDLAM—Was that in the vicinity of that transmitter or was it elsewhere?

Mr Walsh—No, that incident was investigated by us and the report is available on the website. I have a copy here as well. That happened about 240 kilometres north-west of Perth, so it is about 900 to 1,000 kilometres away from Exmouth. The nature of the failure in that one was determined. What had happened in the case of the Boeing 777 was that there was an accelerometer within the ADIRU that failed in 2005. It is a fault tolerant unit so it is actually able to accept those failures and it switches to another component and isolates the faulty component. What had happened a number of years later was that the accelerometer that was being used failed and the software within the ADIRU allowed the ADIRU to revert back to the one that had failed previously. There were really two faults, to some degree. There were two failures of the accelerometers and there was a software change done within the ADIRU that actually allowed the ADIRU to reference that faulty data. We had quite a clear picture of what happened in that particular case and safety action was put in place with that one to upgrade the software and to eradicate that problem. It was also an ADIRU of a different manufacturer, so it was not manufactured by the same manufacturer of this one here.

Senator LUDLAM—It would seem to be drawing a long bow to relate those two incidents in terms of the proximity to that base certainly, but are there other incidents that you are aware of in proximity to that transmitter, or ones like it around the world, interfering with navigation systems?

Mr Walsh—No, not that we are aware of.

Senator LUDLAM—Is that something that you have gone back and you have researched?

Mr Walsh—That is something that we will continue to research. As I said, we will not be closed to it, but we are not aware of anything at the moment. We will certainly try to gather any information of that nature.

Senator LUDLAM—It occurred to me on the basis of the media reports that, if my mobile phone is powerful enough to mess these systems around, then flying over a transmitter that is able to communicate with nuclear submarines halfway around the world would have to be a factor that you would consider.

Mr Walsh—It certainly is one that we are considering. We think it is unlikely at the moment, but we will cover that off and we will make sure that, if there is potential for that to have been an issue, we will certainly look at it.

Senator LUDLAM—When are you expecting to publish your findings and will those findings be made public or are they internal documents?

Mr Walsh—All ATSB reports are public documents and they go on to the website and are distributed broadly for the safety message to make sure people are aware of the circumstances and are aware of what safety action has been taken to address safety issues. To give you a definitive time frame for how long the investigation will take is very difficult. We endeavour to put out a preliminary factual report to at least detail the facts of the accident within about 30 days of the accident. At any time if we identify any critical safety issues, we are not going to wait for the investigation to be completed to make sure that those things are acted upon. We would bring those things to the attention of the relevant authorities to act on straightaway and we would probably put out some sort of interim report to make sure that that was communicated publicly so people were aware of progress and what had been done to address any issues that do arise.

Mr Bills—We are pleased that Airbus has indeed taken some safety action already to mitigate the possibility of this sort of thing happening again.

Senator LUDLAM—How can they do that when they do not know what the cause of the glitch was?

Mr Walsh—To clarify that, they have put out an operations engineering bulletin that essentially provides guidance to crews on what to do in the event of an inertial reference system fault so that they can isolate it straightaway. It is actually like a mitigation strategy to minimise the risk. In the unlikely event that something like this does happen again, there is a procedure and a checklist and a set of instructions to crews on how to respond to an inertial reference system fault. Essentially that is in place pending finding out what has actually caused this.

Senator LUDLAM—Thank you very much for your answers tonight.