It’s been some time since I last posted on this blog. There are some weak excuses (lots of work to get the TCAS RAs working correctly, busy running a business) and some strong excuses (three weeks holiday on a narrowboat, son’s wedding etc.). Anyhow, I’m back on the blogging trail and I can tell you about what I have been doing on the TCAS RA front. This has taken me far longer than I expected, because some things make TCAS RA monitoring more complex than it would, at first, appear.
Why not TA?
A sensible question is why am I not looking at TCAS TA (Traffic Advisory) alerts first, as these precede the Resolution Advisory commands. The answer is because TAs are not recorded on about half the world’s aircraft. If you think that Airbus and Boeing each manufacture about half the world’s aircraft, I leave you to flip a coin…
Inconsistent Data Sets
Now, as a regular reader of this blog you will be familiar with the fact that different aircraft record different parameters (and if you’re not a regular reader, where have you been? Welcome, and let me tell you that different aircraft record different parameters). This is particularly so in the case of TCAS where, currently, we have found 387 different parameters with TCAS in the name, including the obvious “TCAS RA” discrete but also the less obvious “TCAS Port to Channel 9 External Failure” or “TCAS Central Fault Display Input Active” and my personal favourite “TCAS XPDR Hijacking Mode (FO)”. Is it the First Officer who is doing the hijacking, I wonder?
Seriously, while many of these parameters are useful for TCAS system fault finding or maintenance, there are some which are useful for detecting operation of a resolution advisory, but inconsistent naming and different representations in the recorded data makes it difficult to use these parameters.
The most useful parameter, apart from the basic on/off discrete, is the Combined Control word which is available in almost all the frames. The parameter records a number of states including:
0 = No Advisory
1 = Clear of Conflict
2 = Drop Track
3 = Altitude Lost
4 = Up Advisory Corrective
5 = Down Advisory Corrective
6 = Preventive
7 = Not defined
Even here, the ARINC standard does not define the conditions 2 and 3, and not all aircraft record these states. However, we have seen aircraft record Drop Track and TCAS RA = True at the same time (and we know the pilot reacted, confirming this event), so we have to presume that a resolution advisory can be provided even when the system is not tracking the intruder aircraft.
As ever there is a trade-off of using enough parameters to correctly detect the event we are trying to monitor, and ideally which are available on the widest range of aircraft, while not using more than are necessary and adding unnecessary complexity to the algorithms.
Inconsistent Pilot Reaction
If pilots were all the same, life would be dull indeed.
The pilot reaction should be along the lines given in the FAA Introduction to TCAS II Version 7.1, thus:
Do not deviate from an assigned clearance based only on “TA” information.
To satisfy RAs, disconnect the autopilot and maneuver using prompt, positive control inputs in the direction and with the magnitude TCAS advises.
TCAS expects the initial vertical speed response to a “climb” or “descend” maneuver advisory using 1/4g acceleration within 5 seconds after issuance of the advisory.
Initial vertical speed response to an increase or reversal RA is expected by TCAS using 1/3g acceleration within 2.5 seconds after issuance of the advisory.
We can expect the pilot reaction to a TCAS RA alert to vary in speed of response and the strength of that response. Some of this relates to the type of alert, but here is a classic example of a response that is difficult to assess. The TCAS Combined Control parameter only shows “Drop Track” in this case, but we can see that the pilot made a substantial push forward on the controls to give a reduction in normal acceleration down to below 0.8g. Of more interest is the reduction in acceleration some seven or eight seconds before the combined control changed state. We can guess that a TCAS Traffic Advisory was shown (TA was not recorded on this aircraft) and the pilot assessed the situation and started making a precautionary reduction in the climb rate (from 2,000 fpm down to less than 1,000 fpm when the RA came on). He then took firm action to achieve a rate of descent of almost -1,000 fpm before settling back towards level flight.
Now, how do you reflect this in a few simple numbers? Well, after some trial and error we ended up with a set of measures which, while not perfect, can reflect the most common characteristics of a TCAS RA.
Before we start looking at TCAS RAs we have to avoid triggering in error, by being very careful to remove all the different failure cases we see in practice. The phase TCAS Operational embeds all these data validity checks:
The aircraft must be more than 360 ft above the airfield.
An RA period must not start or end below 360ft.
TCAS Combined Control must be active for less than 10% of the data.
TCAS Combined Control must not have values that are not defined in the ARINC standard (i.e. values 2, 3 or 7).
If a TCAS Status parameter is available, it must show “TCAS Active” or “Normal Operation”.
If a TCAS Valid parameter is available, it must show the valid state.
If a TCAS Failure is available, it must show no failure for at least 8 seconds (because these signals have been seen to alternate every 4 seconds).
TCAS RA Response KPVs
TCAS RA phases can only exist if the TCAS is Operational, as defined above, and must last for more than 2 seconds.
For a RA Alert, TCAS RA Direction will determine if the initial alert was Up = +1, Down = -1 or Preventive = 0. The length of time for which the alert was raised will be recorded as TCAS RA Warning Duration. Note that this will not include Clear Of Conflict time.
The pilot should disengage the autopilot following the TCAS RA, so the KPV TCAS RA AP Disengaged will be set to 1 if he has disengaged the autopilot and 0 if he did not. The additional KPV TCAS RA To AP Disengaged Duration will record the time between the onset of the RA and disengagement, which may be negative.
The time that the pilot took to identify the alert and initiate a response will be determined by monitoring the time from the start of the alert to a change in normal acceleration of more than TCAS THRESHOLD. TCASRA Reaction Delay will be recorded for the first change of acceleration in the correct sense.
We see RA Alerts where the pilot has started responding before the onset of the RA, and in these cases TCASRA Reaction Delay will be negative. To cater for these cases on a mix of aircraft with and without TA Alert signals, a simple assumption will be made, for aircraft without a recorded TA Alert, that a TA Alert started ten seconds before the RA Alert.
For the strength of response, we will use the peak g level as TCAS RA Acceleration.
If the initial pilot action to the RA is in the wrong sense, we will record TCAS RA Erroneous Acceleration with the peak acceleration as the measure of error. This will only be computed for the first RA advisory.
The TCAS RA Advisory can change during a single conflict. For aircraft with a useable TCAS Altitude Rate Advisory signal, changes in the altitude rate advisory will be used to detect changes, while for those which do not have this parameter recorded, a change in the combined control parameter will be used. The pilot’s response to the subsequent RA advisory will be timed at TCASRA Subsequent Reaction Delay with strength TCAS RA Subsequent Acceleration. Subsequent reaction delays will never be negative, and the automated analysis will not detect erroneous subsequent actions. Only the first subsequent change will be monitored.
To identify the overall impact of the manoeuvre, TCAS RA Change Of Vertical Speed will be computed over the entire RA period to summarize the effect on the flight path. TCAS RA Altitude STD, TCAS RA Altitude AAL and TCAS RA Heading will be computed for convenience when exporting data ready for further analysis (date, time, latitude and longitude are already available for all KPVs).
TCAS Failure Ratio
One system monitoring KPV will be included, namely TCAS Failure Ratio. This will record the proportion of the airborne time during which the TCAS system was Operational, as a percentage.
A picture speaks a thousand words, so here is one of the development diagrams showing most of the items described above:
Here the pilot took action before the RA was asserted. The RA Reaction Delay and TCAS RA To AP Disengaged Duration are both negative. TCAS RA AP Disengaged is 1 to show that he disengaged the autopilot. The black erroneous acceleration shows that the pilot started moving in the wrong sense, and the red correct acceleration will be the report TCAS RA Acceleration.
Pause for Breath
That’s enough for one day – in a later blog I will take a look at some more examples to show the variety of responses that we see.