The crash was investigated by members of Safety Investigation Board (later changed into the Aircraft Accident Investigation Board). According to Pakistani newspaper
The Express Tribune, the crash of Flight 661 was the first ever crash investigation that would be conducted by an independent Safety Investigation Board. According to the decision that had been taken by authorities, the SIB would be separated from the control of CAA, which was operating under the Ministry of Defence. Instead, the SIB would report directly to the Aviation Division. Assistance would be provided by
French BEA and
ATR. The
NTSB of the
United States also had an active role in the investigation as the propellers were made by US-based
Woodward Propellers. Subsequently,
Canadian TSB was also involved in the investigation as the engines were manufactured by
Pratt & Whitney Canada. Investigators retrieved the aircraft's
flight recorder soon after the crash and it was sent to France for decoding. Chairman of PIA also claimed that the aircraft was "technically sound", while PIA spokesperson Danyal Gilani further accused the media of "baseless accusation", stating that "it defies common sense that pilots will fly aircraft with defects". Analysis of the flight recorder data continued through January 2017, with the reason why the aircraft was not able to safely land on the functioning engine still unknown. A one-page preliminary report of the Safety Investigation Board found a 'lapse' on the part of the PIA and a 'lack of oversight' by the PCAA.
Series of malfunctions Despite claims from PIA chairman that the aircraft was airworthy and had been checked properly by their employees, the findings from the investigation revealed that the aircraft in fact had three pre-existing technical problems; a fractured
turbine blade of the left engine, a fractured pin inside the flyweights of the overspeed governor (OSG) of the left engine, and debris inside the overspeed line of the propeller valve module (PVM). All three defects had directly contributed to the aircraft's unusual behaviour during the accident. One of the blades inside the turbine had fractured at some point before 7 December flight. During the flight, the fractured blade generated vibration that travelled through the turbine shaft. The vibration enabled physical contact (rub) between components, one of which was the No. 6 seal bearing. Contact between the turbine shaft and the bearing inflicted damage to the bearing, causing some metal flakes from the bearing to get in to the oil system of the engine. The engine finally degraded but it had not failed yet as the turbine was still able to produce power. If the engine failed to work properly, instead of being turned by engines, the propellers would be hit by the oncoming airflow from the front. When it occurred, the
blade pitch would decrease. Simultaneously with the decrease in blade pitch angle, the propeller speed would increase as the decrease of the propeller angle would cause the propeller to catch the wind more easily. If the propellers managed to get rotated by the airflow, these would endanger the flight as it would generate drag, causing the airspeed to decrease. The phenomenon is known as
windmilling. To prevent the blade pitch from reaching low angle and the propeller speed to go beyond the
overspeed value, the aircraft was equipped with a device called
the overspeed governor (OSG). The protection system of the overspeed governor will activate when the overspeed valve is opened. To open the overspeed valve, the aircraft is equipped with flyweights. In the case of an engine failure, these flyweights will open the valve by pushing a plunger upwards. The plunger sits directly above the flyweights and usually rotates in tandem with the propeller, along with the flyweights and the rotating part of the overspeed governor. To rotate simultaneously in a safe manner, the plunger is equipped with rotational pin to connect it with the flyweights and the rotating part of the overspeed governor. The pin inside the involved ATR-42 had broken off prior to Flight 661. Due to the broken pin, the plunger lost its connection with the propeller. The flyweights, however, were still able to rotate in tandem with the propeller. As a result, the corners of the plunger got caught with the flyweights. This put metal fatigue to the flyweights, one of which had broken off before the flight due to excessive stress. In Flight 661, the fractured blade caused metal flakes to enter the overspeed line, collecting itself near the overspeed valve. This put more stress to the flyweights as it hardened the rotation of the plunger. The excessive load pulled the flyweights upwards by a little, causing it to push the plunger upward by a bit, partially opening the OSG valve. The partial opening managed to decrease the propeller speed. At this point, the engine had not failed yet. The Propeller Electrical Control (PEC) sensed an anomaly on the propeller's speed and tried to change the blade pitch to increase the propeller speed. It failed to do so as the blade pitch was already controlled by the OSG, not the PEC. The PEC then sent messages regarding the failure to the crew. Meanwhile, the OSG valve became closed again as the last remaining flyweights began to break due to fatigue, causing the blade pitch to decrease and the propeller speed to accelerate. The sudden change in propeller speed generated sound that was loud enough for the crew to notice. As the flyweights were faced with an enormous resistance from the hardened rotation for another time, the flyweights were pulled up again, causing the plunger to go up as well and opening the OSG valve. The propeller speed started to decrease again. Minutes later, the torque of the left engine dropped to 0% in seconds and the captain declared that the left engine had failed. To prevent the propeller from windmilling, the crew had to feather the propeller. The crew then initiated the propeller feathering procedure, which involved the feather solenoid and the protection valve. In principle, the activation of the feather solenoid will provide protection on the feathering process, ensuring that the propellers will always be feathered and thus preserving the airspeed. To activate the protection system, the feather solenoid will retrieve oil from the opening of the protection valve. The valve has two modes; unprotected mode and protected mode. In the protected mode, the valve will open and oil from the overspeed line will pour in. As a result, the calculated pressure inside the OSG line will decrease, while a nearby chamber will record an increase in pressure. This difference in pressure will force the blade pitch to increase, thus reducing the speed of the propeller. In Flight 661, the line that connected the feather solenoid with the oil had been contaminated with debris of unknown origin. When the oil travelled through the line, it picked up the debris and caused it to clump together at a bottleneck. This restricted the oil flow to the feather solenoid, causing the protection valve to switch from the protected mode to the unprotected mode. Therefore, the propeller was not able to be feathered. Back in the OSG, the flyweights had failed as the plunger had pushed through them. As it was not positioned at the top of the flyweights, the plunger could not be pushed by the flyweights anymore and as a result the OSG valve was closed and could never be opened again, enabling the blade pitch to go beyond the lowest allowable angle and the propeller speed to accelerate. As all of the propeller feathering system had been compromised, the propeller could not be feathered properly and the airspeed continued to deplete. The unfeathered propeller on the left engine generated massive amount of drag. Additionally, the thrust was not symmetrical and thus the aircraft was pulled to the left. While the left propeller was windmilling, the propeller also rotated the turbine of the left engine to the same direction. As the blade pitch became nearer to zero, the engine managed to absorb the forces that rotated the propeller, causing the propeller speed to decrease gradually. The force of the hitting airflow eventually could not withstand the friction forces of the engine's turbine and the propeller suddenly stopped in its track. The sudden drop in propeller speed caught the crew by surprise, which caused the aircraft to barrel roll due to the excessive input that had been made by the crew. After the first plunge, the propeller speed managed to stabilize to a value below 5%. Despite this, the blade pitch had reached a value below zero, known as reverse angle. Such blade angle would have been prevented by the ATR-42 safety feature, the SLPS protection system. However, the system was overridden before by the contaminated overspeed line and thus it could not prevent the blade pitch from reaching low angle. The reverse angle, normally used for landing, generated massive drag with a value of around , seven times greater than the normal propeller drag in a single-engine operation. This further compromised the airspeed, causing it to continuously drop even though the lever of the right engine had been significantly advanced forward. By this point, it was impossible for the crew to maintain their altitude.
Cause of failures According to the manufacturer of the engine, Pratt and Whitney, there was a known problem with the blades of the
PW127 engines. If the turbine had exceeded the 10,000 hours lifespan, there would be fractures on the blades. The problem had been reported since 2007, nearly 10 years before the crash. The company then issued a service bulletin in October 2015 for design renewal on the blades and the service bulletin was transmitted to every operator of the engine. The service bulletin stated that operator of the engine should replace the blades within the very first available opportunity to change it. Pakistan International Airlines confirmed that they knew about the service bulletin, but chose not to follow it. Approximately 93 hours before the crash, the involved aircraft, AP-BHO, was brought into PIA maintenance facility for a repair. The engines were disassembled and the workers gained access to the turbine blades. According to the investigators, this was the very first opportunity for PIA to replace the turbine blades. The blades of the involved aircraft had accumulated a total of 10,004.1 hours, exceeding the threshold that had been set by the manufacturer, but PIA didn't replace them with the new blades. The aircraft was put back into service with the old blades. One of the blades eventually fractured, setting off the whole sequence of the crash. While PIA was responsible for the maintenance of the engines, they were not permitted to conduct maintenance on the OSG of the aircraft, which included the flyweights and the plunger. Under PIA, the components were listed as repair abroad item, in which said components should be sent abroad for maintenance. The OSG had been maintained at least three times and the last maintenance was conducted in 2015. At the time of the accident, the flyweights and plunger were not functioning normally as the rotational pin of the plunger had snapped. According to the manufacturer, Woodward Propellers, analysis regarding the aircraft's OSG components revealed that there was an improper assembly on the lower body of the OSG. According to the analysis, the components were forced to work altogether. The review further stated that the technique that was used during the improper assembly of the OSG had taken much more time and work than the normal technique. The maintenance hence was presumably done by someone who did not understand how to assemble the flyweights and the plunger, but did so anyway. The rotational pin of the plunger eventually snapped, damaging the flyweights. Unfortunately, investigators could not determine when or where this unauthorized maintenance took place. The company stated that a set of tests had always been conducted on whether the produced components from the company were at an acceptable level. According to the review, this was the first time that the company had ever received a faulty OSG as records from 1994 showed that there were virtually no reported similar complaints. Investigators concluded that it was unlikely, though not impossible, that the maintenance had been conducted by an untrained technician from Woodward Propellers. The debris inside the overspeed line was big enough that it could neither go through the filters nor a narrowing area inside the line. It was concluded that the debris had not originated from the metal flakes that had been produced by the failing engine. According to investigators, the debris was likely introduced into the overspeed line when the propeller
LRU was installed on the gearbox. However, investigators could not determine when or where such maintenance took place.
Oversight As the findings indicated that there were errors during maintenance of the aircraft, further investigation was needed regarding the role of structural factors into the crash. The aircraft was maintained by PIA in the airline's
Maintenance Repair and Overhaul (MRO) facility in Karachi. The report described the PIA maintenance facility as one of the pioneers in the region to achieve high-level certification, including from the
European Union Aviation Safety Agency (EASA). PIA's MRO facility in Karachi's
Jinnah International Airport had even provided services for countries in the
Middle East,
Central Asia,
Far East and
Southeast Asia. Operational review from investigators revealed that some improvements were required in several fields, including the assembly and disassembly of engines, bearing inspection and material handling. However, as further analysis regarding the airworthiness of PIA's fleet was needed, a review from Pratt and Whitney was requested to compare the reliability of the ATR fleet/PW127 engines that had been operated under PIA. The result showed that PIA's ATR/PW127 engines had very low reliability compared to the other ATR fleet worldwide. Even when investigators compared PIA with other airliners in near similar operating environment, the reliability of PIA's ATR/PW127 engines were still considerably lower. The logbook from 2008 to 2016 showed that there were several cases of inflight engine shutdown. Despite this, no proactive measures and further analysis had been taken by PIA or CAA. Investigators also noted that there were deviations from the procedure that had been issued by Pratt and Whitney. Such deviations were documented or registered by CAA Airworthiness system. Another problem regarding the issuance of dubious pilots’ licenses began to appear as well. Irregularities on physical pilot attendance during licensing exams and recorded number of participants during a specific period of time gave rise to the suspicion. The names of Captain Janjua and First Officer Akram initially were among the list of pilots who were suspected of having such dubious licenses, but the issue eventually became irrelevant to the cause of the crash as the AAIB concluded that the actions of the crew during the flight had been commensurate with their respective training and experience. The finding, however, was still concerning. Added with the fact that PIA had purposely not complied with the service bulletin that had been issued by the manufacturer, Pakistan's CAA, which was responsible for the oversight of the nation's aviation safety, was questioned on their oversight. The annual audits of PIA from 2014 to 2018 revealed that there were gaps and loopholes in the monitoring and evaluation of aircraft airworthiness and safety spectrum and CAA was unable to identify the issues. The CAA was described by investigators as being unable to demonstrate proportionate conclusions, identify the trends and undertake proactive interventions. The report concluded that the oversight mechanism that had been established by PIA and CAA was inadequate or ineffective to identify weak areas within the system's scope.
Conclusion The AAIB released the final report on 18 November 2020. The cause of the accident was a fracture of a turbine blade in the number one (left side) turboprop engine as a result of improper maintenance. This led to the initial engine failure. A fractured pin in the overspeed governor allowed the propeller to reach rotational speeds in excess of 120%. The highly variable propeller speeds resulted in rapidly changing aerodynamic characteristics. The propeller eventually settled into a very high-drag configuration. The ATR-42's behaviour was different from the 'typical' loss of a single engine and level flight became impossible. The report notes, "It was exceptionally difficult for the pilots to understand the situation and hence possibly control the aircraft." It was also noted that crew resource management was ineffective, but did not contribute to the accident. The AAIB issued several recommendations to the involved parties. The first part was consisted of two urgent recommendations, which were issued during the progress of the investigation. The first recommendation was issued in January 2019, in which PIA was asked to inspect their entire ATR fleet to change the blades that had met the criteria for replacement. The second recommendation was issued in August 2019, following request by the NTSB and Collins Aerospace. The recommendation stated that PIA should inspect all of the OSGs inside their ATR fleet and the OSGs should be sent to United States after said inspection, following discovery of an improper assembly. The second part of the recommendations were given after the completion of the investigation. Among those, PIA was asked to ensure strict compliance with the issued service bulletin, to conduct improvements on weak areas in its MRO facility and to identify critical performance indicators within its airworthiness and flight safety operation. The cockpit resource management training system of PIA and the country was asked to be revamped. Improvements were also ordered for CAA relating to its oversight system. According to the final report, the manufacturer of the aircraft, ATR, were to consider including a specific procedure in the future aircraft training programme, in case of possible similar aerodynamic features to that encountered by the crew of Flight 661.
FAA and
Collins Aerospace stated that they were considering a system review and also a possible improvements of the filtration inside the oil system to prevent it from being clogged by unwanted debris. ==Aftermath==