Iran Accident

[Richard Taylor]

Passenger plane crashes in Iran mountains - BBC News (http://www.bbc.co.uk/news/world-middle-east-43103192)

Not much detail yet other than:

* was routing Tehran - Yasruj
* around 60 aboard
* came down in a mountainous area in Central Iran

[SFI145]

Iranian aircraft with 60 people on board crashed near the city of Semirom in the Isfahan Province of Iran, according to Tasnim news agency citing Iran's emergency centre. The aircraft went off radar midflight from Tehran to Yasuj.

[piperpa46]

According to TV2 Denmark quoting Tasnim and Fars it's an ATR 72 of Aseman Airlines

[Jonty]

According to the BBC bad weather is hampering rescue efforts

[c_coder]

https://www.mirror.co.uk/news/world-news/passenger-plane-carrying-60-people-12045621

"Witnesses told local media that it appeared the aircraft was trying to make an emergency landing on a pasture before it crashed (http://www.mirror.co.uk/all-about/plane-crash) in the Dena mountain range."

[Thaihawk]

ATR72-500 according to the latest on this on the BBC,

Passenger plane crashes in Iran mountains - BBC News (http://www.bbc.com/news/world-middle-east-43103192)

[A0283]

Aseman Airlines plane, en route from Tehran to the south-western city of Yasuj, came down in the Zagros mountains.The Red Crescent deployed search and rescue teams to the site near the city of Semirom in Isfahan province. Flight EP 3704 left Tehran at 08:00 local time (04:30 GMT) and disappeared from radar about an hour later. The aircraft crashed on Dena Mountain, 22km (14 miles) from Yasuj, news channel Irinn reported. The plane was a French-made twin-engine turboprop ATR 72-500. On board 60 passengers, 2 security guards, 2 flight attendants and the pilot and co-pilot.
No survivors.
+
The aircraft involved is a 24 year old ATR 72. Registration EP-ATS. MSN 391.According to FR24 logs flight took off at 04:33 UTC. Last signal was received at 05:55 UTC when flight was at 16,975 feet and descending. FR24 state they can only use MLAT for tracking as the six types of the airline carry old type xpdrs.

[Jump Complete]

If it’s 24 years old it can’t be a -500 (properly -212A) AFAIK, which came out I believe in 1997. Anyway, a sad day regardless of the minor details.

[Super VC-10]

https://en.wikipedia.org/wiki/Iran_Aseman_Airlines_Flight_3704

[ChickenHouse]

Mountains, snow, bad weather, quite "interesting" circling NDB approach and teardrop ... not a combination you want that often.

[aterpster]

Jeppesen has only two charts. Appears a VOR/DME was recently installed. Interesting there is no departure procedure.

[aterpster]

Per Av Herald the crash site is about 16 miles west, northwest of the airport.

[ThreeThreeMike]

The Associated Press is reporting that until a few months ago the aircraft had been out of service for seven years.

The aircraft only began flights months ago after it was “grounded” for seven years. The airline posted a photo of the plane on Instagram saying it would be up and running after repairs were done and safety checks and tests were conducted. It’s unclear what caused the aircraft to be grounded for an extended time.

[DaveReidUK]

Reportedly WFU due to shortage of spares.

[txl]

Video supposedly showing a couple of seconds of final approach into Yasuj on a ATR a few months ago:

https://twitter.com/ranjourr/status/965143539553325056

Iranian Aerospace Forum on Twitter (https://twitter.com/Aerospacetalk):


Crash site: Near #Semirom, 14NM N of Yasuj (YES/OISY) with high terrains
Last communication: Over IFN at 9:22 AM, crew requested descend to FL170
W/x: Poor condition w heavy wind & snow, OVC090.

One PAX reportedly a no show.

Aircraft overhauled after a few years in storage due to lack of spares, put back in service in November 2017:

https://twitter.com/Aerospacetalk/status/934748143366868992

This supposedly shows the weather conditions in the area:

https://twitter.com/MohsenApply/status/965264654065897472

[aussiepax]

Weird, there is such a long valley lining up with that runway in Yasuj, hard to imagine the need to be so low WRT those mountains in the northwest in that clip.

[Hotel Tango]

The Associated Press is reporting that until a few months ago the aircraft had been out of service for seven years.


We don't know yet of course but I would very much doubt that aircraft serviceability had any role in this. More likely a scenario relating to weather and procedures.

[aterpster]

A really lousy NDB approach. Wonder whether the Iranian aviation authority was/is working on a VOR approach? They spent the funds to install that VOR/DME recently.

[beamender99]

The wreckage of a plane that crashed in Iran on Sunday has been found, local media are reporting.

from the BBC

[jolihokistix]

Found? No, not confirmed yet.

[Carbon Bootprint]

Reportedly WFU due to shortage of spares.Can you name the source? I'm not meaning to sound doubtful and all, but that is a lot of time for bird to be on the ground if it wasn't properly looked after. Thanks.

[txl]

Re: Source

The Guardian reports, citing Iranian media: (https://www.theguardian.com/world/2018/feb/18/iran-plane-crashes-aseman-airlines-yasuj)

An Iranian news website said the plane had recently rejoined the air fleet after seven years of undergoing repairs. “The plane which crashed today faced technical problems midair during a recent flight a few weeks ago,” Roozarooz news reported.

An Instagram post from Aseman Airlines two months ago that announced the plane was back in service after seven years was deleted this morning after the plane crash, Roozarooz reported.

Also, see the tweet by Iranian Aerospace Talk Forum I cited above. This could be sourced from the allegedly deleted instagram post.

[RatherBeFlying]

The missed approach from MDA demands accurately tracking an NDB back bearing. Not every ADF installation does back bearings as well as front bearings.

How many crews today can do a good wind correction on an NDB bearing?

Was there any calibration flown on the approach?

Even more exciting would be flying a missed from as much as 4000' below the MDA. Was calibration for the missed flown from just above the runway?

[aterpster]

The missed approach from MDA demands accurately tracking an NDB back bearing. Not every ADF installation does back bearings as well as front bearings.

How many crews today can do a good wind correction on an NDB bearing?
The old salt Russians. Dual ADFs on RMIs.

Was there any calibration flown on the approach?
Only Iran knows that.

Even more exciting would be flying a missed from as much as 4000' below the MDA. Was calibration for the missed flown from just above the runway?
That might not be part of Iranian flight inspection specs. Or, many other countries for that matter.

I believe the recent installation of the VOR/DME is evidence that Iran was on the path to improvement at this difficult airport.

[ThreeThreeMike]

We don't know yet of course but I would very much doubt that aircraft serviceability had any role in this. More likely a scenario relating to weather and procedures.

I agree with you.

The economic sanctions placed on Iran years ago caused quite a few aircraft to be sidelined.

[Kulverstukas]

TEHRAN, Iran (AP) - Iranian search-and-rescue teams on Tuesday offered the first images of the site of an airplane crash in southern Iran that killed 65 people, with officials hoping to reach the aircraft's "black boxes" to learn exactly what downed the flight.

State television aired footage showing the plane crash site against the side of a snow-covered mountain near Yasuj, some 780 kilometers (485 miles) south of Tehran, Iran's capital from which the Aseman Airlines flight took off on Sunday.

A helicopter pilot interviewed by state television said the crash site appeared to be only 30 meters (100 feet) from a peak on Mount Dena in the Zagros Mountain range.

"Some large parts of the plane, which were labeled with the Aseman company logo, were seen," said the pilot, identified by state TV only as Capt. Soheili.

[A0283]

the wreckage had been found through drone images before the air force's helicopters were deployed.

From now on drone deployment appears to have become a standard tool in both SAR and investigative efforts. Similar use in the Antonov case. And especially helpful in difficult conditions ... Like the high snow in these two 2018 cases. In this case the helicopters could not land and people on foot are the only option. Drones make a search far more efficient and effective which is a great help for those in SAR and contributes to their safety.

[G-CPTN]

What size of drone?
Are we talking 'pilotless aircraft'? - or a portable toylike device?


Edited to add:-
The wreckage was finally spotted by a military drone (https://www.reuters.com/article/us-iran-airplane/iran-finds-wreckage-of-crashed-plane-on-top-of-mountain-idUSKCN1G40T3)

[twistedenginestarter]

Not every ADF installation does back bearings as well as front bearings. They are Non Directional Beacons, surely?

[aterpster]

Blockage by terrain could limit the use of an NDB. Flight inspection has to verify signal strength for both the approach and missed approach. A weak signal can cause a false bearing.

It will be interesting to learn whether the flight flew the procedure correctly to well below MDA, then did a missed approach.

[Timmy Tomkins]

It's hard not to feel for this crew. Who knows if they pushed limits but the captain was a veteran and presumably familiar with both airport and the limitations of the approach; as sad end to what was probably a distinguished career.


Not the type of approach facility I would want to deal with in those conditions; especially when they could have had a VOR DME approach there. Was there radar cover in the area or where they expected to let down using own nav? That could explain the end result if the aircraft only had basic kit.

[albatross]

The missed approach from MDA demands accurately tracking an NDB back bearing. Not every ADF installation does back bearings as well as front bearings.

How many crews today can do a good wind correction on an NDB bearing?

Was there any calibration flown on the approach?

Even more exciting would be flying a missed from as much as 4000' below the MDA. Was calibration for the missed flown from just above the runway?

An ADF That doesn't allow tracking outbound from the NDB?
How the heck would you do an NDB approach?

[JW411]

I think the poor chap is getting confused with a Back Course ILS. He obviously has never flown an NDB let down in his life (unlike me).

[Patanom]

Europian Datum 1950 is using in Iran. The height difference from WGS-84 is 30.24 meters.
Therefore, I think it started to decline about 1 km earlier. The distance along the axis of the runway differs by approximately 60 meters (2 arc seconds).
Sorry, the distance along the axis of the runway differs by approximately 30 meters ( abt 1 arc seconds).

[Plumb Bob]

Jeppesen has only two charts. [...]On that approach chart 16-1 (dated 20 Oct 17) the minimum altitude during the teardrop turn is given as 12,500 for category A and B aircraft, but 11,400 for the faster C and D categories. I would expect that to be the other way around, i.e. the higher altitude for the faster airplanes with their wider turning radius, as well as their outbound course that is aimed closer to the mountain range in the east.

[aterpster]

That occurred to me, too. It would be interesting to see state source; i.e., the official Iranian chart in their AIP. I doubt that is available on-line, though.

Jeppesen is careful, but not perfect.

In any case, the accident did not occur in the area of the base leg.

[galaxy flyer]

Actually, the Iran AIP is online, but I couldn’t access the charts on an iPad, requires desktop.

[ironbutt57]

They are Non Directional Beacons, surely?

not the onboard equipment

[Machinbird]

Everyone is discussing the interesting approach, but it is not clear to me that the aircraft even reached the Yasouj NDB to commence the approach.

Those are some serious Mountains north of Yasouj, and tall mountains, strong winds can create extreme problems in maintaining altitude.

So far, information on the route of the flight has been pretty sketchy. :confused:

[JW411]

ironbutt57:

Exactly so; the NDB is the bit on the ground and the ADF is the bit in the aircraft.

[old,not bold]

Exactly so; the NDB is the bit on the ground and the ADF is the bit in the aircraft.2 parts of the same system; the ADF will point to the NDB. And that's it, at least for us KISS adherents.

As I understand the remark about a "back-bearing", an ADF will not show exactly that. What the pilot must do is subtract 180 (using a calculator if necessary) and fly the reciprocal, not forgetting to adjust for drift, so that the bearing to the NDB remains constant as he flies away from it. Or something like that. It's been a long time...... as I recall it, some mental agility is needed when you are doing this in IMC as part of an NDB approach/cloudbreak, what with everything else to think about.

[wiggy]

Not that we do this (much) anymore but I'm a bit confused about reference to calculators and the like. Has the old RMI skill of "pulling the tail" to fly outbound from a VOR/NDB on a fixed radial simply died a death?

[The Fat Controller]

NDBs don't have radials !

Flying a TRACK away from an NDB is not an easy task, especially when there is a serious crosswind.

[aterpster]

Everyone is discussing the interesting approach, but it is not clear to me that the aircraft even reached the Yasouj NDB to commence the approach.

Those are some serious Mountains north of Yasouj, and tall mountains, strong winds can create extreme problems in maintaining altitude.

So far, information on the route of the flight has been pretty sketchy. :confused:

So far there are two possibilities based on that sketchy info:

1. the flight got off course along the airway and began descent prematurely (for whatever reason), or

2. The approach had been flown and a missed approach was in progress, albeit off course.

The recorders hold the answer because there surely is no ATC radar once in that valley or canyon.

[ironbutt57]

Flying a TRACK away from an NDB is not an easy task, especially when there is a serious crosswind.

"pull the tail"

[wiggy]

NDBs don't have radials !

Flying a TRACK away from an NDB is not an easy task, especially when there is a serious crosswind.

OK, OK, my bad, excuse the nomenclature, but I suspect/I hope you understood what method I was driving at with that post....there is a basic RMI technique that covers both VOR radial and NDB outbound course tracking and there's no real need to deploy calculators and do loads of pre-flight planning to do so.

Take a procedure which has an outbound track from an NDB..you can achieve and hold the required TRACK from your NDB by "flying the tail" or "dragging/pulling the tail" of the ADF needle so that the tail lies against the RMI compass rose on the outbound track demanded by the procedure (FWIW for any errors I was taught to use double that error as the steering correction) . Yes I know you need to take drift into account, but TBH having done this stuff in the past for NDB, VOR and TACAN procedures, and well before the days of GPS and overlays I personally wouldn't say following a track or radial outbound from a beacon was in the rocket science level of difficulty, it is (or at least it used to be) standard procedural IF "pilot stuff". I'd also agree it can be demanding on a bad weather day and very unforgiving of gross errors if there are terrain issues..


Call me old fashioned but I do agree to a certain extent with Arctic Circle's comment.

[twochai]

So far there are two possibilities based on that sketchy info:

1. the flight got off course along the airway and began descent prematurely (for whatever reason), or

2. The approach had been flown and a missed approach was in progress, albeit off course.

There is a third possibility: LOC - Loss of Control in severe icing.

[Plumb Bob]

Well, yes, and not a particularly super combination of aircraft type and environment.
Of course that has been addressed in the past, but, [speculation mode] think of car tires that become brittle with the passing of time if not flexed (massaged) and remaining stationary for years on end – how well can deicing boots withstand deterioration by the weather during a very long period unprotected without any use? That combination of external influences just could be a new factor here. [speculation mode off]

[aterpster]

There is a third possibility: LOC - Loss of Control in severe icing.

That's why I included "for whatever reason" in my option #1. Severe icing and being quite off-course are usually independent events.

[Timmy Tomkins]

My post #31 asked this question; does anyone know the extent of radar cover in the area?


Was there radar cover in the area or where they expected to let down using own nav? That could explain the end result if the aircraft only had basic kit.

[Teddy Robinson]

with all due respect, radar cover in an area with those MSA's is an oxymoron

[Airbubba]

A report that the Aseman ATR 72 fleet has been grounded:

Iran Civil Aviation Organization (CAO.IRI) has ordered Iran Aseman Airlines to ground its ATR 72 fleet until further notice, following the crash on February 18.

https://twitter.com/aviationirancom/status/966796258894667776

[RatherBeFlying]

The investigators, once they have the FDR data, will likely want to repeat the approach and miss in VMC with one of the grounded airframes with a similar ADF installation to see where the 310 back bearing takes them.

Even better would be in similar winds aloft. The FDR data should be sufficient to derive this.

[aterpster]

You're assuming they made the approach. We don't know that.

[AAKEE]

Timings/reported distance and probable time for crash and position assumes it wasnt possible to make it to the IAF and back to position for crash.

[RatherBeFlying]

The crash position seems to have some imprecise alignment with the missed approach segment which has raised my suspicions on the quality of the missed approach course guidance.

If they had not begun the approach, there's a whole bunch of other questions. I look forward to further information.

[twochai]

https://news.aviation-safety.net/2018/02/23/iran-suspends-operations-atr-72-200-500-aircraft-wake-fatal-accident/

[Timmy Tomkins]

with all due respect, radar cover in an area with those MSA's is an oxymoron

That's a rather patronising response to a sensible question. Old a/c, probably no GPS, and as others have said, we don't know that they were on or missing an approach.

They could have needed urgent decent (fleet grounded now so, icing/other failure?) and radar confirmation of position would have been very useful.

The terrain varies in height and MSA is based on the highest in the sector so being steered around the worst if descent unavoidable would have been invaluable - I speak from personal experience; so Oxymoron No definitely not

[mcdhu]

Also potentially relevant are altimeter temperature and wind error which, if not applied correctly could eat up a large proportion of the increment added to the highest ground.

Extreme cold and high winds, which were both reported, can have a a significant effect.

We wait to see

[Double Back]

ADF and a fixed card was a great tool for training ab initio pilots, it sure weeded out those who could not multi-task and paint a picture in their brains of the position relative to the ADF bcn or RWY. I did my share during training with intercepts from all kind of directions, entering holding patterns.. Using the RMI made life a lot better.

In actual airline flying, in 30 Years of airline flying, almost all of it on heavy jets, I might have flown 3 actual NDB approaches in not so good WX.
During the 17 Years as capt I started my non-precision crew briefing stating we had an emergency coming up :) That got everybody's attention.
Flying any non precision in a heavy jet is outright dangerous, enough terrible accidents have proven that. Just one wrong frequency and You got completely lost.

[aterpster]

Timmy Tomkins

My hunch is that they were on the OIFM ASR radar until 60 miles south of OIFM. But, that is still 80 miles north of OISY. I don't know whether Iran has long-range center radar covering the area between OIFM and OISY. If so, the minimum altitude of center radar coverage would likely be fairly high in the vicinity of OISY.

[Timmy Tomkins]

Thanks for the informative response; I suspect you are right.

[Patanom]

GPWS has geometric height input signal (no barometric). RA don't work in this range. I suppose they is used data abt height from GPS. In South Iran, I hope, can receive EGNOS from sattelite IOR 64 East. ATR-72 was below glideslope abt 30 metres and runway close abt 120 metres.

[Machinbird]

AvHerald has updated their info on this crash.
Crash: Iran Aseman AT72 near Semirom on Feb 18th 2018, impacted terrain (http://avherald.com/h?article=4b511c15&opt=0)
The following picture from AvHerald shows that they never reached the Yasuj NDB to begin the approach. I imagine the CVR will be able to provide some clues as to what they were thinking/seeing.

http://avherald.com/img/aseman_at72_ep-ats_semirom_180218_map4.jpg

[aterpster]

AvHerald has updated their info on this crash.
Crash: Iran Aseman AT72 near Semirom on Feb 18th 2018, impacted terrain (http://avherald.com/h?article=4b511c15&opt=0)
The following picture from AvHerald shows that they never reached the Yasuj NDB to begin the approach. I imagine the CVR will be able to provide some clues as to what they were thinking/seeing.

http://avherald.com/img/aseman_at72_ep-ats_semirom_180218_map4.jpg

An early poster who was flying in the area at the time, stated that TRW were around. Speculating: Perhaps the ADF was messed up by the TRW.

I hope Iran lets us all know in due course. As I recall, they do a very good job of accident investigation, with subsequent disclosure in accordance with ICAO protocol.

Perhaps TRW that were quite cold at the flight's altitude, which could result in a lot of ice in addition to turbulence.

RIP

[Patanom]

In the absence of visibility, they apparently went in to land on the GPS.
What are the "underwater" stones when entering GPS?
1. Is it known exactly which ellipsoid (datum) is used in the country?? (page 6 of manual GNS430)
2. Whether the reception of differential corrections (WAAS, EGNOS, etc.) is reliable?.

Maybe I'm wrong, but it turns out this picture:
- if the GPS is operating in diffrent mode (2 in field 6) sentence $GPGGA, the GPS will ignore
the incoming altitude from the barometer and will take the height of the GPS, as more accurate.
- What is the height of the GPS? This is the distance to the satellites.
- then "roll" the desired geoid / ellipsoid and get a height above the terrain.

A similar catastrophe was in Colombia on November 28, 2016 RJ-85, there in the airport area
the WAAS system operates.

[aterpster]

The crash site is about 4 miles from the airway centerline.

[Nil further]

Double Back

"Flying any non-precision in a heavy jet is outright dangerous"

What complete cobblers , cant speak for Mr Boeings but with Mrs Airbus , any NPA is a delight , very similar to an ILS , yes certain traps and gotchas but "downright dangerous , rubbish .

If you are briefing an emergency for a simple manoeuvre that is easy to brief and fly , you are in the wrong job .

[PEI_3721]

Re #64/67. The information appears to relate to Airbus aircraft which may use an alternative form of TAWS/GPWS.

Was this aircraft fitted with TAWS or EGPWS?
Did this aircraft have an external GPS (Nav). Was this connected to EGPWS, or did the aircraft only have a GPS receiver within EGPWS (no Nav), or none at all?
What is the resolution of the EGPWS terrain data base in this area, similarly for the airfield. If the airport was not fully covered then basic GPWS should still be available - subject to switching / selection.
Rad Alt will display altitude according to scale, normally at any time less than 2500ft agl. (E)GPWS will use this, and thus if serviceable there should at least have been a basic GPWS alert / warning.

IIRC ATR had a history with GPWS inhibit switching, it could be completely switched off to inhibit nuisance alerts (fatal accident, ... Taipei ??) . Is similar switching used for EGPWS, or is there the more conventional selection of terrain off - leaving the basic system active?

[Patanom]

Re #64/67. The information appears to relate to Airbus aircraft which may use an alternative form of TAWS/GPWS.

Was this aircraft fitted with TAWS or EGPWS?
Did this aircraft have an external GPS (Nav). Was this connected to EGPWS, or did the aircraft only have a GPS receiver within EGPWS (no Nav), or none at all?
What is the resolution of the EGPWS terrain data base in this area, similarly for the airfield. If the airport was not fully covered then basic GPWS should still be available - subject to switching / selection.
Rad Alt will display altitude according to scale, normally at any time less than 2500ft agl. (E)GPWS will use this, and thus if serviceable there should at least have been a basic GPWS alert / warning.

IIRC ATR had a history with GPWS inhibit switching, it could be completely switched off to inhibit nuisance alerts (fatal accident, ... Taipei ??) . Is similar switching used for EGPWS, or is there the more conventional selection of terrain off - leaving the basic system active?


I don't know, what kind of navigation equipment was installed on this airctaft. GPS, probably, is HT-1000.
RadioAltimeter can't work in this case - out of range ( abt 1500 m).

[Patanom]

HT-1000 GPS.

[PEI_3721]

Patanom, I disagree with your views of rad alt.
If the aircraft hit a hill then the rad alt was 0ft. Thus for flight preceding that, at some time the altitude would have been between 0ft and 2500ft.
RA input to the basic GPWS would have warned according to terrain rise and aircraft descent rate; whether the aircraft could be recovered would depend on the time available according to terrain profile and crew reaction.
If EGPWS was working then earlier alerts would have been given so that a recovery manoeuvre should not be in doubt ... except for ...

Local research suggest that the 212 version of the aircraft would have been fitted with EGPWS vice a Thales alternative, although the latter is now standard.

Previous EGPWS accident in ATR 42
Crash: Trigana AT42 enroute on Aug 16th 2015, aircraft collided with terrain (http://avherald.com/h?article=48ae1ad9&opt=0)

[FCeng84]

Double Back and Nil further,

My intent is not to fan a disagreement, but your exchange prompts me to ponder whether or not it is valuable for a flight deck crew to treat (at least within the cockpit) any situation that they rarely experience as if it were an emergency. It seems to me that there is high value in elevating the attention to details for everyone involved if the task at hand is one that is not common and has the potential to go south in a hurry if not done properly.

I am not suggesting that every time a crew encounters a situation that they have not seen in the last six months that they declare an emergency to ATC and make cabin announcements that would overly worry their passengers. What I am finding re-assuring as one who is in the airplane design end of the business and whose flying experience is about half as a test engineer in the back and half as SLF is that the folks at the front end take their safety job seriously and do their best to earn my trust and their pay. This is particularly the case when conditions lead to a combination of events not typically encountered.

Another part of my thinking on this is that with real emergencies so few and far between these days it seems valuable to me to have crews practice for the eventual bad day by treating a few less than perfect days as if they had the potential to become real "remember when" stories for pub visits deep into the future.

My deepest respect for all aircrews who spend their time up front anticipating and preparing for the day when they may earn a career's worth of respect over the course of just a few minutes or hours.

[autoflight]

Aircraft configuration changes alter GPWS responses. I recall A320 GPWS from many years ago where gear down at less than 1800ft/min rad alt closure rate to terrain would give "too low flap". This alert would be more typically found during a constant altitude flight during a circling approach over high ground outside the circling area with less than landing flap. Same result could be expected if other minimum altitudes or defined tracking are not respected with gear down.
I am confident that more than a few readers would find this safety warning limitation of GPWS surprising.
At some time the GPWS has to recognise that the intent is to land and this seems to be a transition from normal full flight capability to actual touchdown conditions. I accepted that there is a transition risk. Unless there have been other mode changes since, the reduced warning capacity needs consideration.

[Patanom]

PEI 3721, I agree, that EGPWS has to work accordance with this algoritm. But....
Two signals are sent at the input to the Geometric Height, - the radio altimeter and GPS.
I suppouse, that in this region there was a reception of differential corrections (WAAS or other)
through the geostationary satellite POR 140 degrees.
In this case GPS altitude will be used. In database of EGPWS is WGS-84, but in this region the Indonesian datum is used and the
height difference is 22.99 meters.
Those. the EGPWS system considered a different relief than the actual one.
I suppose that with the ATR-72 the same situation (the difference in altitude is 30 meters)

[interestedparty636]

From AvHerald:



On Mar 11th 2018 Iran's AIB released their preliminary report in Persian reporting that according to FDR and CVR the aircraft had been handed over to Yasuj Tower, the autopilot was set to 15,000 feet. Descending through 15,600 feet the crew activated the anti-ice systems. The aircraft levelled off at 15,000 feet on autopilot, the crew set the QNH to 1021 and maintained 15000 feet for about one minute. Then the engines were reduced to idle, the speed reduced to 200 KIAS with the angle of attack increasing, the engines get slightly accelerated. The speed continued to decrease and reached 129 KIAS (minimum maneouvering speed 132 KIAS), the pitch reaches 15 degrees nose up, the engines accelerate to 67% torque. The altitude target is set to 14,000 feet and the aircraft begins to descend at about 600fpm. The speed further reduces to 117 KIAS, a stall warning activates, the crew disengages the autopilot, the aircraft rolls 20 degrees to the left, the pitch reduces to about 9 degrees nose down. Descending through 14200 feet at 137 KIAS the autopilot gets re-engaged, the aircraft rolls right by 12 degrees, the pitch increases to 5 degrees nose down. A GPWS warning "TERRAIN AHEAD! PULL UP!" activates, the autopilot is disengaged, the GPWS warning continues for 12 seconds until impact.


The AIB continued that there was no technical malfunction of the aircraft, the engines operated in accordance to pilot inputs, all aircraft systems supplied the crew with valid data. Due to the cloud cover the crew remained unaware of the mountains ahead until 2 seconds before impact and rolled the aircraft sharply left in order to avoid the terrain.


The AIB stated that the crew should have maintained 17,000 feet in accordance with the flight plan, however, descended the aircraft to 15,000 feet followed by a target altitude of 14,000 feet on the autopilot contrary to flight rules. In addition, while the crew was permitted to conduct the flight with the weather data available at the time of departure, the latest weather information provided by Yasuj Tower indicating clouds up to 15,000 feet prohibited the approach to Yasuj according to company procedures due to cloud cover present at the aerodrome, the crew should have diverted to Shiraz or Isfahan planned as alternate aerodromes. Pilot discussions according to the CVR confirm the presence of cloud up to 15,000 feet confirming the accuracy of the weather report by Yasuj Tower. Although the aerodrome was still more than 10km away the crew appeared to be confident the area would be in visual meteorologic conditions. The AIB warns that all of this is first interpretation of first investigation results and is not to be taken as cause of the accident.

[Kulverstukas]

It turns out a pilot error, but why?

According to the report, it was allowed to reduce the altitude to 17,000 feet, but the pilot initially set the altitude of 15,000 feet on the AP, and then set the altitude to 14,000 feet.
Everything, including engines and anti-icing systems, worked fine, and a technical problem was not noticed.
The pilot at 9:31:14 received a warning from GPWS.

From another forum:

Interesting decision to re-engage the autopilot following a stall. Not necessarily a turning point in the sequence of events. I'm assuming the momentum of 67% from that point on? I believe around low 70's is a normal cruise power for those altitudes in that aircraft. Certainly reads like there was a major breakdown in situational awareness in the flight deck.

Ignoring a GPWS for 12 seconds is puzzling. Language barrier? Startle effect? Poor training?

Final (?) report, but not yet translated into English ...
http://www.mizanonline.com/files/fa/news/1396/12/21/1555079_507.pdf

[aterpster]

The minimum altitude on the airway is 17,000. But, they could have elected 15,500 when within the north MSA sector. However, they were almost 4 miles off course, which is a significant navigation deviation and an unsafe location from which to begin the instrument approach.