Engine Out Over Europe

18 Oct 19 16 Comments

On the 23rd of August 2019, The Aviation Herald reported that a Smartwings Boeing 737-800 suffered an engine failure in flight.

The Boeing 737, registration OK-TVO, departed Samos, Greece at 06:27 UTC for scheduled passenger flight QS-1125 to Prague in the Czech Republic with 170 passengers on board.

Smartwings aircraft on departure, taken by Aktron / Wikimedia Commons

About 15 minutes into the flight, flying over the Aegean Sea, they had just reached their cruise altitude of FL360 (36,000 feet) when the left-hand engine shut down spontaneously. At this point, they were about 100 nautical miles northeast of Athens, Greece.

The crew descended to FL240 (24,000 feet) and followed the checklists. They first attempted to restart the engine by windmilling: the airflow from the high airspeed rotates the compressor, in hopes that the compressor will reach a high enough rotational speed to ignite. When this failed, they attempted to restart the engine from the working right-hand engine, using a method known as cross bleed.

The left-hand engine did not restart.

The instructions in the Quick Reference Handbook (QRH) for an in-flight shut down event is to plan to land at the nearest suitable airport. Of course, the term “suitable” leaves some room for manoeuvre. At this point, they were about 30 minutes from Athens.

The flight crew decided that there was sufficient fuel on board to continue the flight on one engine, so rather than divert, they carried on at FL240, overflying North Macedonia, Kosovo, Serbia, Hungary and Austria on route to their final destination of Prague, leading to a flight time of an additional two hours and twenty minutes.

The controller at Budapest ACC (Area Control Centre) who dealt with the flight as they overflew Hungary posted in the comments the same day, complaining that the flight crew mentioned a technical issue, which he presumed was an air conditioning fault. At no point did the flight declare a MAYDAY or PAN. He had no idea they were flying on one engine.

I work at Budapest ACC and I was in contact with this A/C when they overflew Hungary. Not a word did they mention about engine failure we were informed about a “technical issue”. That is in most cases an air con failure for the 737 to fly at 240 or 250. So they came in at FL240 as they reached the Austrian border we sent them to Vienna Approach.

10 minutes later they called us back to inquire us why we haven’t told them about the engine failure. It turned out that the failure was announced over Prague, Vienna then called us back but we didn’t know either. Serbians were also unaware and I also asked my colleague who was working at KFOR (Kosovo) airspace – it is also operated from Budapest – but he didn’t know either.

Czech media confirmed that air traffic controllers in the other regions were also not informed, with the exception of Czech controllers at the final stage of the flight.

There was some discussion on PPRuNe on whether the controllers had a right to be upset that they were not notified.

FAA regulations CFR 121.565 are clear that this is a requirement:

The pilot-in-command must report each engine shutdown in flight to the appropriate communication facility as soon as practicable and must keep that facility fully informed of the progress of the flight.

However, the EASA regulations do not seem to have the same requirement, stating only that any flameout, shutdown or malfunction of any engine must be reported but it is not clear, as it is in FAA regulations, whether the flight crew must report the occurrence.

They began their descent at 08:49 and the Boeing 737-800 landed safely at Prague. Maintenance staff were able to diagnose the problem as a faulty unit closing the engine valve, so no fuel was being supplied to the engine. The aircraft was repaired and tested and returned into service about seventeen hours after landing. There is no record of the fuel they had available upon landing, although the fuel burn at the lower altitude would have been higher than estimated for the flight, there’s no evidence that they landed on low fuel.

The CVR was not removed, thus the discussion in the cockpit during the flight was overwritten.

Smartwings Airline confirmed the incident the following day, saying:

The crew proceeded in accordance with the safety and operational procedures for these cases and the aircraft landed safely. The commander is one of the most experienced in the company, the crew was in control of the situation and certainly would not underestimate anything.

The commander referred to here was the captain of the flight and also the Director of Flight Operations at Smartwings.

On the 26th of August, three days after the incident, the Civial Aviation Authority of the Czech Republic stated that they were investigating the situation. According to the translation on Aviation 24:

We have decided to launch an investigation as the procedure followed doesn’t seem standard.

Smartwings responded win an internal memo confirming that gross violations of the company policy were being investigated and reminding staff that they were not to comment on social media networks and were not to publish internal information on the internet. The gross violations in question were the leaks to the Czech press.

Czech news site ZDopravy followed up with the CAA (English via Google Translate):

The Czech Civil Aviation Authority stated in a statement that “… there was no landing at the nearest suitable airport as it is in the interest of the safety of the crew and passengers on board. “ The Director of the CAA David Jágr said on Czech Television that they were shocked by the pilots’ progress and consider the case unprecedented.

On the 13th of September, Smartwings confirmed that the captain of the flight was no longer holding the position of Director of Flight Operations, although he continued to fly for the airline. Their internal investigation, they said, had identified pilot error as the cause of the incident.

This followed a leak that their own internal documentation showed that the pilot had not followed Smartwings procedure to land as soon as possible rather than continue to the destination and that the aircraft had departed Samos with “minimal” fuel. In addition, the fact that what was at the very least a dubious decision was made by the chief pilot of the company did not reflect well on Smartwings’ safety culture.

Meanwhile, the company memo does not seem to have achieved its goal. Last week, Aviation Herald reported that they had information that during the cruise on one engine, Smartwings ops contacted a Budapest-based maintenance provider to ask if they had the spare parts on hand to deal with a blocked fuel supply. Presumably the answer was no, as the decision was made not to land at Budapest but instead to continue.

The CAA has confirmed that they knew nothing about the incident until the Aviation Herald contacted them asking for comment. They have not released any further details and it is unclear whether they are planning proceedings against the crew, the carrier or both.


  • Quite apart from legalities, it seems to me that if you have an engine failure with unknown cause, you have no idea whether the cause might not be a common one (e.g. contaminated fuel) which could also knock out the second engine at any moment.

  • Roger,
    You “pipped me to the post”, but yes I am of the exact same opinion.
    Leaving aside the legal aspects, there is also the crew’s responsibility to the safety of the aircraft and all on board.
    Common sense should have prevailed: An engine flames out in mid-flight.
    The crew, this time, do take out the checklist and attempt a re-light. This is not successful.
    So there is no directly known cause for the failure of an engine. Okay, a modern airliner can fly quite comfortably on one engine. I do not wish to dispute that the captain was in full control of the situation, and that the fuel, the remaining endurance, was not an issue either at FL 240.
    BUT: Would common sense not have dictated that the crew divert to “the nearest suitable airport”? So: Secure the aircraft, using the checklist assess the situation, alert ATC – who were quite rightly upset – and, if feasible, inform the company and DIVERT.
    As Roger quite correctly mentioned: One engine failed, cause undetermined. An attempted re-light failed. More than enough reason to opt for safety.
    Good basic handling, poor airmanship.

  • And, BTW, something similar happened to me on a cargo flight from Spain to Paris CDG. We did not have a flame-out, but a loss of oil pressure on one engine which prompted us – checklists of course used correctly – to shut it down.
    At the time we were flying over Bordeaux. The reson for the shut-down was obviously not connected to a potential common fault somewhere else in the aircraft, there were no passengers on board, only myself and my F.O. But even if we had suficient fuel to continue to Paris, this simply did not occur to us. We were well trained, there was no dispute possible: LAND AT THE FIRST SUITABLE AIRPORT. Which was Bordeaux. No “ifs”, no “buts”.
    Had we decided to continue and the company had found out, that is where we would have been in trouble. I am in no doubt that I would have been told to hand in my uniform jacket and get the fourth bar removed: I would have been demoted to F.O.
    So I am still pondering: What were these pilots thinking of? Playing the heroes with the lives of 170 passengers, not to mention those of the crew including themselves?
    In most companies that I have been working, the captain would have been lucky if he would only have been demoted. The F.O.? Possibly a spell in training? With another check to confirm his suitability perhaps.
    Unbelievable !

  • The Aviation Herald reports: “While OK-TVO was still airborne south of Hungarian Airspace (after the engine shut down over the Aegean Sea), the office of Smartwings is said to have contacted a Budapest based maintenance provider about whether they would have spare parts readily available, the fuel supply of one of the engines was blocked. Subsequently it was decided the aircraft would not land in Budapest.”
    So that gives us a clue as to what the “head of flight operations” who was piloting the plane may have been thinking.

    From the comments on that AvHerald article, it appears that the plane landed with less than 30 minutes of fuel remaining, which is cutting into safety margins (as if flying on one engine wasn’t enough!). What was the weather like on September 13th?

  • Mendel,
    You contribute relevant information, but the story does not really change.
    The CAPTAIN is responsible for the safety of the aircraft, NOT the person who happens to be in the OPS office. The CAPTAIN was flying and made the (wrong!!) decision to continue.
    As Roger already, correctly, remarked: “you have no idea whether the cause might not be a common one…” Spot on the mark.
    An aircraft continuing flight on one engine AND arriving with only 30 minutes’ fuel in the tanks was in fact in an EMERGENCY situation and should have declared this, by using the “Mayday” call.
    Drastic? Yes, you bet, but he was in an emergency of his own making. And still “kept mum” until it came out.
    Nothing to redeem him. He dug a hole for himself and, when finding himself in that hole, kept digging.

    To demonstrate how seriously most companies take the rules: I will tell you how I was once demoted myself.
    I was working for a cargo airline; we operated different routes, one was out of an small Italian airport.
    The ATC controller used to give us a clearance to an altitude well above the intitial altitude as prescribed in the SID. It had been quietly understood that this cancelled the SID altitude restriction and we were clear to climb directly to this higher altitude (or level).
    My annual simulator check was due. The examiner was a retired KLM 747 skipper and took the regulations very seriously indeed. The instructor who operated the simulator was our chief pilot.
    Of course, the departure that was programmed in the simulator was from that very Italian airport.
    The instructor, acting as the ATC controller, gave us the usual clearance: an altitude well above the initial SID. And, in line with the actual accepted modus operandi, I climbed to that altitude.
    The examiner failed me: I should have stayed on the altitude as listed in the Standard Instrument Departure (SID) until the SID exit point.
    Which was, of course, correct but not in line with the daily (or better: nightly because most air cargo is shifted at night) routine at that very airport. And so I was demoted to F.O. until a few weeks later the same chief piliot who had led me into this trap repeated the check and re-instated me as captain.
    Of course, mindful of this episode, the very next flight as (re-instated captain out of this very same airport, but now of course in the airplane and not the simulator, I was again given an altitude above the SID altitude. This time I stuck to the lower (official, as prescribed on the Jeppesen SID chart) one, not having reached the SID exit point yet.
    Which of course prompted a surprise remark from ATC asking why I was not climbing to the cleared altitude.
    This narrative, as I mentioned, is intended not to demonstrate my stupidity, but to show how seriously operators – professional ones – take the rules and procedures.

    • Rudy, the point is that OPS (as in, the head of flight operations) was piloting the plane that night.
      It’s fairly obvious that economy and safety are conflicting concerns, and the same person should NOT be in charge of both!

  • Dear Silvia,

    I always enjoy reading your articles because they you one of the very few writers who enable non-aviation-specialists to understand facts behind the stories. That‘s why I believe it‘s worth to correct some small details of your report about the Smartwings incident.

    First of all the term „suitable“ has a clear definition in aviation language, it‘s the nearest airport where a safe landing can be made according to the technical trouble. Translated into simple words, as soon as the runway is long enough, the weather is good enough and the airplane can easily deal with performance requirements (e.g. obstacles in case of a single engine go-around, etc.) such an airport is suitable. Suitability never considers economical factors as well as one of the most important human factors: „I want to sleep at home tonight“ ;-) So actually there is not much room for interpretation. Additionally it was daylight and the weather was good everywhere. Returning to Athens would have been the best option for many reasons: two runways, both well in excess of 3500m, easy access for spare parts delivery, „first aid“-maintenance facility, etc., etc.

    Some other small remarks to your report: a fully loaded 737-800 is hardly able to reach FL 360 within 15 minutes, it will rather take between 20 and 25 minutes in case a continuous climb has been provided by ATC (I have never experienced this during the past 25 years in Greek airspace).

    And in fact it takes also about 30 minutes to reach a position 100nm NE of Athens coming from Samos.

    Next remark is about flight time and fuel burn. The flight time was not prolonged by additional 02:20hrs, this was the total flight time. The „usual“ single engine cruising level for jet aircraft is – depending on weight and thrust – between FL 220 and 250, in most cases 240 and 250. The reason for this are several factors, e.g. at this altitude you can fly the highest indicated airspeed (usually between 300 and 340 KIAS) while also achieving the highest possible TAS (True Airspeed), usually in the range between 430 to 460 KTAS – details always depending on several other factors, too.

    In turn this means that the airplane is not flying much slower than in normal cruise altitudes. It can easily fly its maximum certified airspeed which is also easy to achieve on one engine only. This is one of the fundamental physical or aerodynamic facts to cover ETOPS-requirements where you have to gain as much distance as possible within a given time frame (usually but not exclusively the well known 120 and/or 180 minutes diversion time).

    And since all these physics are related to weight, speed and drag, surprisingly the fuel flow remains almost the same over the travelling time/distance comparing single engine cruise in FL240 to two engine cruise at FL360. There is only a relatively small increase in fuel burn, mainly because the airplane is not flying in perfect aerodynamic symmetry.

    Consequently even using „minimum fuel policy“ there is a good chance to make it to your final destination, even in one engine out condition.

    As you probably know a standard fuel calculation also requires a dedicated „route reserve“ for unforeseen circumstances like weather deviations or flying at lower altitudes the originally planned due to ATC requirement. Typically this is 5% of the calculated trip fuel, under certain circumstances this can be reduced to 3%. So this extra fuel covers already some of this „Smartwings-requirement“ to fly back home. Furthermore for taxi out usually for jet airplanes a taxi fuel of 200kg is considered. On airports like Samos it is easily possible to get airborne before you have used just 60 or 80 kgs of fuel due to very short taxi times.

    For further inflight re-plannings you can also change your alternate to a closer one than originally planned or – in case of very good weather conditions – you are allowed to consider the second runway of your destination airport as alternate, provided you have extra fuel for additionally 30 minutes (+ the required 30 minutes for final reserve).

    Technically for me it is easily imaginable that their remaining fuel was well within limits.

    Although the decision to continue to the final destination was a clear breach of all regulations and safety requirements I don‘t see this as dramatic as the press and other commentators do. Taking reference to the above mentioned ETOPS regulations even specially certified B737s are approved to fly single engine for 120 or 130 minutes according ETOPS regulations but for sure the Smartwings-737 did not have such an approval which requires very special maintenance procedures.

    The only true concern I would have had in this situation: why did the engine stop working? There was no clear sign, no noise, no vibration, nothing. It just stopped operating. An not knowing why and engine shut down without any obvious reason would make a bit more worried than a failure because of damage, etc. Why did it stop without any sign, with sufficient fuel in the tanks?

    Exactly in this moment these two incidents would have popped up in my mind: one A330 ran out of fuel in the middle of the Atlantic, luckily gliding into one of the airports of the Azores. All of this just because a faulty O-ring has been installed somewhere in the fuel system, finally leading to total fuel loss.

    And even a much better example: the BA B777 loosing both engines on short final into Heathrow due to fuel contamination with ice crystals as far as I believe.

    Dual engine failures are extremely rare and due to their extremely low probability they don‘t find their way into every day‘s risk calculation. But they happen. That‘s one of the reasons why I would consider „suitable“ a little bit more conservative than my colleague from Smartwings.

    …….just my 2 cents (or pennys ;-)

    Kind regards,

    • A quick note regarding timing… my notes of the FR24 data is
      06:27 departure
      06:46 reaches cruising altitude FL360
      06:49 start of descent (so the engine is already out by this point)
      07:05 FL240
      08:49 start of descent for approach

      I overlayed this onto the avherald article which said that the aircraft landed around 2:20 hours after the initial descent to 240. Based on this, I figured avherald’s estimate of 140 minutes looked feasible (06:49 to 09:10) but I don’t have the actual touchdown time to verify it– does anyone happen to know?

    • A commenter on the AVHerald story relates: “As per latest information, during CAA investigation, it showed up that they landed with only 900kg fuel remaining on board. Apparently commanders licence ua suspended.” I haven’t been able to find an independent source for this.
      30 minutes translates to 1250kg of fuel, apparently, making this landing below minimums.

    • “Although the decision to continue to the final destination was a clear breach of all regulations and safety requirements I don‘t see this as dramatic as the press and other commentators do.”

      I am not a pilot or other professional, but I do fly as a passenger quite a lot. This incident is legitimately “dramatic” because there was a substantial chance that the aircraft and its passengers and crew would end up spread all over the place in the Aegean Sea. It didn’t happen that way, they got away with it.

      I certainly agree with your interpretation of “suitable” while we’re looking for a place to land in this situation. Air travel has become so safe that perhaps some people are insufficiently frightened. A healthy fear would have suggested a prompt return to Athens, in the spirit of “let’s not push our luck”.

  • Almost all incidents where human error is involved have common traits. Loss of perspective is one of them. Eastern, Flight 401 was a perfect example as the flight crew became so distracted by a burned out landing gear indicator light that no one was monitoring the flight and they pancaked into the Everglades.

    It sounds as if this flight had a variation of the same sort of thing. An engine flamed out. Apparently the flight crew had in inkling that it was fuel system related, probably from onboard diagnostics. There is a logic to their actions, but this logic is flawed by lack of perspective. Apparently, the pilot became so obsessed with this problem that he forgot that the first priority was the safe operation of the aircraft.

    Just for the purpose of argument, let’s assume that they knew this to be a hardware problem and not related to fuel contamination. The left engine is flamed out and will not restart. The right engine still has the same possibility of failure as it always has, but now the consequences of such a failure have risen dramatically.

    The PIC’s first responsibility is the safe conduct of the flight. A drunk passenger, a mechanical problem, unexpected severe weather or any number of other possibilities are the responsibility of the PIC. In most cases, whenever there is a problem, the obvious solution is to get on the ground and deal with the problem there.

    We occasionally read about diversions because of one unruly passenger. A diversion for an engine failure should require no explanation; it’s prudent to analyze with it from the safety of the ground. It’s far better to ruin the short-term plans of 170 passengers than it is to risk their lives pressing on.

  • I can agree with George on most, but not on all fundamentals.
    A jet engine has a much, a VERY much increased fuel burn at lower altitudes (flight levels if you prefer).
    This all has to do with the fact that a jet engine operates most economically when the airspeed matches the outflow of the exhaust gas. In a pure jet, this means at a high mach number, preferably in the supersonic range. And high altitude because of the lesser drag.
    The introduction of fan engines had two benefits: the mantle of bypass air from the fan screened the high velocity air which meant that the turbulence caused by the interaction of hot air core and the cooler, slower air from the fan was reduced. This turbulence was the root cause of the high noise cuased by earlier jet engines. The lower-speed bypass air muffled the noise, resulting in much quieter engines.
    The other befefit was that the air thrown by the fan is moving at a lower speed, which does not need the aircraft to fly at the higher – and not always practicle – speed required to be economical.
    Moreover, the air is less dense at high altitudes and the air displaced by the fan becomes the main driving force, in contrast to lower levels where a larger proportion of the thrust comes from the hot air.
    The very large fans of modern jets have increasingly compensated for this effect and modern jets can cruise more efficiently at lower levels.
    Still, I used to do regular flights from Ireland to Lviv (formerly Lvov) in the Ukraine. We could only do this non-stop if we cruised at FL 430.
    This allowed us to climb above the so-called RVSM airspace.
    But even so, if cruising much lower than FL 390 it would not have enabled us to get to Lviv without a refuelling stop.
    Before the introduction of RVSM (Reduced Vertical Separation Minima), we used to cruise aircraft like the Corvette or Citation 500 at levels between FL 350 and 390. This enabled us to fly non-stop from Dublin to e.g. Nice, Barcelona, even Rome. When RVSM had become a reality it was very difficult in Europe for private operators to obtain RVSM approval, even if the aircraft had been retrofitted with the required equipment and the crew had undergone training. It required a mountain of paperwork that did not seem to be demanded by USA-based operators.
    The result was that we became restricted to maximum FL 280 and we were no longer able to cover those distances without a refuelling stop.
    The C550 Bravo and the C500 Stallion conversion were well able to climb straight to FL 430 and be above RVSM airspace, provided that ATC would clear us to climb through this airspace.
    The 20-series Learjet was guzzling fuel at an alarming rate unless operating at or above FL 390. We usuallly cruised at FL 430, 450, even 470 for range. It could burn as much fuel when taxiing as a 737. That because it was of course still at low (ground) level. I have seen a burn of 500 lbs if we had to wait at a busy holding point.
    I do admit that I retired before the introduction of aircraft with very high bypass ratios, maybe because of an increased ratio of low-speed air they are less affected by the need to fly high for range. But if I may venture a guess: Any aircraft designed to fly at high cruising speed will be operating most efficiently at high levels.
    Turboprop engines displace a proportionally much higher volume of air at lower speed and therefore can cruise efficiently at lower levels.

  • Oh, by the way, to comment on what Mendel mentioned:
    The head of the flight ops department is NOT EVER in charge of an aircraft in flight – unless, of course (s)he happens to be the pilot in command.
    Of course, there are operators that put their captains under pressure to accept circumstances, defects, deviations from procedures, weather conditions or minimum rst periods that are outside of the limits.
    I have actually overheard the chief pilot of a small air taxi company tell one of his pilots say: “I cannot force you to accept this flight (meaning: breaking the law), but if you don’t there will be ‘a parting of the ways'”.
    This was in a period when there was virtually no demand for commercial pilots, things have changed since.
    Mark summarised the situation correctly. George mentioned that a failure of both engines is extremely rare.
    So is winning the Euromillions, but I would prefer to be the rare winner of those millions than the equally rare pilot who is in an aircraft with a flame-out of both engines. Not everyone has the skills and presence of mind as Capt. Sully or Capt, Moody, not even all exerienced captains do. Would I have performed equally well under the same circumstances? I would hope so, I once lost both airspeed indications and near flame-outs on both engines. I was new to the aircraft and there were many conflicting indications, even the experienced engineer was baffled. I could do no more than keepmy eyes glued to the angle of attack whilst my copilot was assigned the task of calling ATC for vectors to the nearest airport.

    • Rudy, I’m confused by your initial statement and I’m not sure I’m following your point here.

      The captain of the flight has been confirmed to have been the Director of Flight Operations at the time of the flight. So I think Mendel’s point is that anything ops did at that time was related to him. I agree with you that as captain, he absolutely had the responsibility of the decision. But are you saying that in this situation, he should not consider himself ops (which I agree with) or that his decision shouldn’t be taken as related to that of ops (which, given his position, is a good ideal but hard to implement in real life)?

      • That is indeed my point. It looks like the captain asked the office to inquire for the spare parts, and made his decision to continue based on that answer, because he was also head of flight operations and thinking in both roles at the same time.

  • Sylvia and Mendel,
    Sorry it seems that I did not pay attention to the fact that the captain was also the director of operations.
    But it re-affirms my assessment, rather than contradicting it.
    The Director of Operations had no business, no right even, to put on his “Operations Hat” when confronted with this situation. Especially in smaller, cash-strapped outfits senior members of staff have overlapping positions that can cause a blinkered attitude, like in this case where the captain was first and foremost concerned with the operational consequences of a diversion: He would end up at some airport where no immediate repair could have been effected. Spare parts? How did he even know if any would needed and what kind of spares would be required? It may have been a matter of clearing a filter, it might have been checking a fuel system for contamination, it could, in fact, have had different causes. It would indeed have incurrred cost and another consideration might have been bis the limitations to his own ability to take charge as he would not have had access to his offfice and potentially limited access to communications with the office.
    So his decision was ultimately decided by his need to “run the show”, to take charge of the operations, rather than his first and foremost duty: to ensure the safety of the aircraft and all on board.

Post a comment:

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.