Factors In Ice-Induced Hard Landings, Part 2 

 Paris-Le Bourget Airport (LFPB) has a tangle of runways, with the shortest being runway 27, at 6,060 ft. The airport is historic. It opened in 1919 and was the landing site for Charles Lindbergh after he first crossed the Atlantic Ocean solo in 1927. Coincidentally, it is also the home of the BEA. Investigators had only a short walk across the field to begin their examination of the wreckage of the Phenom. Upon their arrival, BEA investigators immediately noticed a build-up of ice along the leading edges of the wings and stabilizer.
 
After the initial visit to the accident site, the BEA organized the investigation according ICAO Annex 13. They brought in representatives from Brazil as the state of manufacture and Malta as the state of the operator. Brazil, in turn, brought in technical advisors from Embraer and Malta brought in advisors from Luxwing, the operator.
 
The investigation began with a focus on the design and operation of the airplane’s de-icing and anti-icing systems, the stall warning system (SWPS) and the airplane’s performance in icing conditions. The Phenom has three de-icing/anti-icing systems - thermal, for the engine intakes, electric, for the probes and windshield, and pneumatic, for the wing and stabilizer leading edges. The pneumatic system inflates and deflates leading-edge boots--four on the wings and two on the stabilizer. Some Phenoms are also equipped with an ice detector, but the accident airplane, 9H-FAM, was not.
 
The Embraer Phenom flight manual says that the Wingstab de-icing system must be turned on as soon as the total outside air temperature is less than 5C in the presence of visible moisture, even when there are no signs of ice accretion. A warning in the manual says “The ice protection system must be kept on until the crew is certain all ice has been removed.” When the de-icing system is switched on, the stall warning system and the low-speed awareness tape in the airspeed indicators adjust upward. The difference is significant.
 
For the conditions of the accident flight, the Vref speed without the de-icing system on was calculated to be 96 kt., but with the Wingstab and engine anti-icing system on, the calculated Vref was 119 kt.
 
The calculated landing distance for the accident flight was only 4,252 ft., flying at the lower Vref speed and leaving the anti-icing and de-icing systems off. With the equipment on, the calculated landing distance was 7,549 ft., which exceeded the length of runway 27. In addition, the one-engine-inoperative climb gradient was negative, meaning the airplane could not conduct a missed approach with the de-icing system on.
 
A review of the FDR showed that 3.8 nm from the airport at 1,380 ft, the flight was maintaining 135 kt. It had slowed to 100 kt. by the time it was at 468 ft. An Embraer simulation showed that the stall warning came very close to activating three times during the approach before it activated at the end of the flight.
 
The BEA interviewed the pilot of a Piaggio P180 who flew an approach to the airport about 10 min. before the Phenom. He said his visual ice-accretion probe accumulated so much ice that he took a photo of it. He provided that photo to the investigation, and it does indeed show a massive ice buildup on the probe.
 
The 40-year-old Phenom captain had logged 3,625 total flight hours, including 2,961 on the EMB-500. He had worked at the charter company for 8 years. The co-pilot, age 25, had logged 625 flight hours, including 425 in the EMB-500. His commercial license and EMB-500 type rating were less than a year old at the time of the accident. Both pilots had completed training on de-icing/anti-icing procedures and systems in the last year.
 
The captain said in a statement that he knew that the aircraft’s landing performance would not permit landing at Le Bourget Airport if the icing conditions on approach required the continuous use of the de-icing systems until landing. He said that after coming out of the cloud layer at 2,000 ft. he saw no more ice on the wings, so he deactivated the de-icing and anti-icing. He also said he knew he would have to divert if he left the de-icing system on.
 
The captain’s statement prompted the BEA to talk to other Phenom pilots and to review online Phenom pilot forums. One Phenom pilot said he had been unofficially taught that he could deactivate the de-icing systems after 1,000 ft. if the leading edges of the wings were not contaminated by ice. Another pilot, speaking anonymously on a forum, said “Frequently, in the Northeast, accompanied by bad weather and icing. Phenom book Vref increases when hots re on. For some reason Cessna does not have the same requirements. Spoke with one pilot who did a lot of 100 flying in cold wx. Said that as soon as he cleared the clouds on approach, would turn off the hots so that he could approach and land at normal speeds. Seems reasonable, as long as you remember to turn them back on if you need to go missed.”
 
The flight’s operator, Luxwing, had a fleet of 21 business jets, including 7 Phenom 100’s. Its flight operations department was supposed to check that the landing performance of its airplanes was adequate for the forecast conditions at the destination airports. Apparently, they did not do that for the Paris flight.
 
Recommendations From Previous Accidents

In February of 2013, a Phenom pilot lost control of the airplane in the flare while attempting to land at Berlin-Schönefeld airport in Germany. The German safety investigation agency, the BFU, found that the crew flew the approach in known icing conditions and did not activate the wing and horizontal stabilizer de-icing system. A build-up of ice on the wings and the horizontal stabilizer and the flight’s slow approach speed caused the airplane to stall.
 
The agency thought the crew did not understand the connection between the de-icing system and the stall warning system and recommended additional training for pilots receiving EMB-500 type rating training.
 
In December 2014, another Phenom 100 crashed while on approach to Gaithersburg airport (GAI) in Maryland. The three people on board and three other people on the ground were fatally injured. The NTSB said the probable cause of the accident was “the pilot carrying out an approach at a landing speed below that recommended in the manufacturer's normal procedures in icing conditions and the non-activation of the wing and horizontal stabilizer de-ice system. The combination of these two factors led to a stall at an altitude which made recovery impossible.”
 
The pilot was flying at the appropriate speed for non-icing conditions. He was probably very concerned about stopping on GAI’s 4,202 ft long runway 14. He had a previous runway excursion at that airport in another type of airplane and would have been very aware of the landing distances required. If he had attempted to land with the de-icing system on, even if he performed a flawless approach and touchdown, he would have only had about 100 ft. margin to be able to stop. With the de-icing system off, he had about a 1,700 ft stopping margin. In addition, with the de-icing system on, he did not have a sufficient rate of climb on one engine to conduct the approach.
 
As with the BFU, the NTSB recommended better training. They asked the National Business Aviation Association, manufacturers and training providers to develop better pilot training for winter weather. They also recommended that FAA and the General Aviation Manufacturers Association develop automatic icing alert systems.
 
A problem with the training recommendations is that the pilots in both the Berlin accident and the Gaithersburg accident probably knew that they were operating in violation of Embraer’s flight manual warning. Lack of adherence to a warning or standard operating procedure (SOP) is not a training issue, it’s a compliance issue.
 
The automatic icing alert system recommendation seems like a better idea. The NTSB’s recommendation was directed at turbofan airplanes that require a type rating, are certified for single-pilot operation and flight in icing conditions. The concern is that solo pilots flying turbine-powered airplanes in bad weather are so busy that they may not notice when the temperature and moisture require de-icing equipment to be turned on, and an icing light would be a helpful reminder.
 
Unfortunately, the accident record shows that ice detectors are not foolproof. A case in the NTSB’s database illustrates the point.

Part 3: A link between de-icing related accidents.

Factors In Ice-Induced Hard Landings, Part 1: https://aviationweek.com/business-aviation/safety-ops-regulation/factor…