AINsight: Takeoff Performance System Errors

 - May 22, 2019, 6:49 AM

Automated takeoff performance systems provide for operators an optimized solution that is designed to ensure safety and improve economics. Before each takeoff, pilots input data into these systems to determine the optimal thrust and flap setting for a given runway and environmental condition. These solutions guarantee regulatory runway and climb performance while reducing engine wear and saving fuel. The problem, as with other automated systems, is that flight crews are susceptible to errors that can lead to a risk of a runway overrun due to a takeoff attempt with an improper thrust and/or misconfigured flap setting.

Recent runway overrun or “late takeoff” events highlight some of the risks associated with using automated takeoff performance systems. In August 2016, a Boeing 777-300ER had a late takeoff event at London Heathrow. The aircraft became airborne with only 318 feet of runway remaining and crossed the end of the runway at a height of less than 17 feet. Later investigation revealed that the crew selected the wrong runway/intersection in the Onboard Performance Tool (OPT), resulting in a lower-than-required takeoff thrust setting.

In this case, the first officer initially input the correct runway/intersection into the OPT. After the load/trim sheet arrived, both the captain and first officer re-ran the takeoff performance calculations in the OPT; the captain used the full runway length and the first officer used the intended runway/intersection selection. Before taxi, the crew compared their respective OPT outputs and elected to use the captain’s performance solution. The first officer changed her OPT inputs to match the captain's without any attempt to understand the difference. As a result, the thrust selected for takeoff was based on the full runway length, not the reduced-length intersection takeoff.

According to the report, during the takeoff the captain noticed that the runway centerline lights changed from all white to alternating red and white (900 meters/~3,000 feet remaining) as the aircraft approached Vr. The first officer commented, during subsequent interviews, that the rate of rotation “was a bit faster than normal.”

After the event, it was calculated that the distance required for takeoff (full length) was 10,987 feet and the distance available from the intersection used for takeoff was only 8,494 (a difference of 2,493 feet). Investigators noted that although the aircraft became airborne on the paved runway surface, the takeoff did not meet the regulatory requirements for an all-engine continued takeoff case, would not have been able to stop on the remaining runway in the event of an engine failure below V1, and would not have been able to continue the takeoff (meeting regulatory requirements) in the event of an engine failure after V1.

One year later, a Boeing 747-8 operated by a cargo airline overran the runway at Tokyo Narita. During this event the aircraft rotated 280 feet past the paved surface of the runway. According to reports, the airline reported a wrong thrust setting by the flight crew.

Later in 2017, a Boeing 737-800 crew in San Francisco selected the wrong runway in the automated takeoff performance system (TPS). Rather than selecting Runway 01L for departure, the pilots transposed the numbers and selected Runway 10L (a runway that is 4,220 feet longer). The TPS solution in this case resulted in a thrust and flap setting more commonly used for longer runways. This resulted in a takeoff with only 400 feet of runway remaining. As a result of this event, the FAA issued a Safety Alert for Operators (SAFO 18009) cautioning operators on risks associated with using these automated systems.

Low thrust and misconfigured-flap events are a serious threat to flight safety. Industry studies point to some common contributing factors that are either human- or technology-based. Data input or entry errors top the list of human errors. Common errors include entering the wrong runway (18L vs. 18R), runway transposition (13L vs. 31L), runway/intersection (18R/B vs. 18R), or wrong environmental or runway conditions. Other less common entry errors include the omission of a performance-limiting MEL item or entering incorrect weight and balance information.

Most takeoff performance systems will automatically upload the most recent Metar information. In the past, there have been some reports of wrong or corrupted information being uploaded into these systems. One “bug” in a system resulted in erroneous thrust settings due to inaccurate temperatures (the omission of a -/+) being uploaded.

From these events there are several valuable lessons learned. The first lesson from each of these events is that the takeoff performance system did not fail the crew (trust the system). In each case, had the crew input the correct information, the system would have provided enough performance to take off safely. If used correctly, these systems are reliable.

Another lesson is to remember the old adage: garbage in, garbage out. Always validate the information that is input into the takeoff performance system (i.e. correct runway, conditions, etc.). Likewise, when reviewing the returned-performance solution, ensure that the conditions match exactly what was requested.  

Lastly, the most critical step is to effectively communicate and coordinate with the other crew member when inputting data into the FMC and/or thrust management computer. Always verify and crosscheck that information with the correct performance solution and confirm it with the other pilot. Remember, these are automated systems, pilots should treat performance data entries with the same care as other flight deck automation such as FMCs or mode control panels.

Pilot, safety expert, consultant, and aviation journalist Kipp Lau writes about flight safety and airmanship for AIN. He can be reached at stuart.lau3@gmail.com