18
$\begingroup$

On some instrument approaches, the autopilot-coupled approach is not authorized or not authorized below a certain minimum. For example,

  1. Autopilot-coupled NA (MRY: ILS OR LOC RWY 10R) enter image description here
  2. Autopilot-coupled NA below a minimum (LVK: ILS RWY 25R) enter image description here

Moreover, on some approaches, the DA is lower if we have a AP/FD/HUD.

I guess the lower DA is because we can give more attention outside the window when there is AP/FD/HUD's assistance. I didn't come up with any idea or context why autopilot-coupling won't work.

If it is turbulent and the autopilot doesn't respond quickly enough to keep the needle centre, we may not use autopilot at all. If the air is calm, the autopilot may well outperform hand-flying. However, it depends on the quality of the autopilot.

$\endgroup$
22
$\begingroup$

It's the fact the autopilot works more accurately than a pilot which is actually the cause of the restriction.

The decision to restrict the use of the autopilot usually comes from the certification agency after the ILS inspection flight detected glideslope erratic variations or reversals. The false signals are likely due to interference from the environment or the nearby activities with the runway equipment.

Flight inspection volume

The ILS volume surveyed by certification agencies using ICAO recommendations is located below a 5.25° angle. It means G/S indications are checked and expected to be correct only below this angle.

enter image description here
Source

This is why G/S are usually intercepted from below unless specific procedures are in place and extra care is given to validate G/S indications.

Anomalies in flight inspection volume

However there can be false and reversed guidance areas within this volume. They are discovered when the ILS is commissioned and flight tested by the certification agency. Example of glideslope reversal during an inspection on the nominal slope:

enter image description here
(Source, page 15-97)

In these areas, the ILS signal may vary quickly and if the autopilot were active it could react to the variations, while a pilot will likely not notice the interference, or will disregard it unless it follows a trend for a few seconds. A pilot with a Vertical Situation Display can confirm the ILS error visually.

A reversal, which is a guidance given in the wrong direction, can lead to a serious incident or an accident, due to autopilot quick reaction time. When such reversals are discovered during the inspection flight, autopilot approaches are not authorized, or forbidden during the final segment where the consequences can be the worst.

Consequences of a reversal during autopilot approach

Using the erroneous signal from the ILS may divert the aircraft from the approach path. Being below the glideslope close to the threshold is indeed not desirable, especially for a jet aircraft for which the thrust increase required to recover the lost altitude cannot be obtained instantaneously.

Suddenly jumping higher is also dangerous. It leads to a loss of airspeed, and possibly to an unrecoverable stall. Below is an example of autopilot reaction to a reversal. While it didn't happen in the certified volume of the ILS, the principle is exactly the one found in any reversal.

Some explanation about the context it happened

It was believed for a long time ILS radio signals for the glideslope contain false guidance at angles multiple of the nominal approach slope (e.g. 6° and 9° for a 3° G/S), but it was also taken for certain these false glideslopes were not reversals and would anyway guide the aircraft to the threshold, at a different descent rate (which would be enough to warn the crew about them following a wrong glide angle).

However G/S reversals, especially after 9° has been observed later (2011-2013), for some types of ILS antennas, without a full understanding of the reasons. Among these cases (source ICAO):

  • Schiphol Airport, The Netherlands, 2011, KLM, Embraer E190
  • Murcia Airport, Spain, 2011, Ryanair, Boeing 737-800
  • Charles de Gaulle Airport, France, 2012, Air France, A340 (BEA report)
  • Treviso Airport, Italy, 2013, Ryanair, Boeing 737-800
  • Eindhoven Airport, Netherlands, 2013, Ryanair, Boeing 737-800.

Eindhoven serious incident, May 2013

A B738 on autopilot was flying a non stabilized approach where the G/S had been captured from above. When reaching the 9° glideslope area, at 0.85 NM from the threshold, the aircraft suddenly pitched up to 24° and climbed from 400 to 1,000 ft. Airspeed plummeted to 100 kt, the stick shaker activated:

enter image description here
Source. Original source: Dutch Safety Board report

From the investigation done by the Dutch agency, it appears the actual guidance provided around 6° and 9° is completely reversed:

enter image description here
Source

As a partial mitigation measure, Boeing changed the Flight Control Computer software to limit the pitch up rate in G/S mode. The BEA also recommended manufacturers would prevent the capture of the G/S by an aircraft not on the published glidepath.

$\endgroup$
  • $\begingroup$ I expect the autopilot will be enable to filter out the erroneous noise and take the average. I think the AHRS does the signal noise reduction. $\endgroup$ – skyoasis Mar 16 '16 at 7:07
  • $\begingroup$ @skyoasis: The AHRS has no ILS input. $\endgroup$ – mins Mar 16 '16 at 7:13
  • $\begingroup$ I mean since AHRS can a lot of variation filtering due to engine vibration, turbulence to make heading, pitch reading stable. In analogous to AHRS, the NAV receivers should do the same. $\endgroup$ – skyoasis Mar 16 '16 at 7:22
  • 2
    $\begingroup$ @skyoasis: Probably to some degree (albeit it should be confirmed by reading the regulation for ILS receivers), however this doesn't prevent incidents to occur, like at O'Hare in July 2012. A pilot would look at the PAPI if one is available, or to the vertical path of the navigational display, if the aircraft is fitted with one, and detect something inconsistent with the ILS path. $\endgroup$ – mins Mar 16 '16 at 7:45
  • 1
    $\begingroup$ @rbp: I think it's in the Flight Inspection Handbook TI 8200.52, page 15-96 (Rate of Change/ Reversal in the Slope of the Glidepath): "The use of facilities which do not meet the change/reversal tolerance must be limited by a NOTAM (see Chapter 5, Paragraph 5.12) that withholds authorization for autopilot coupled approaches" $\endgroup$ – mins Mar 18 '16 at 22:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.