How will we stop the engine, then? Kigali Crash into VIP lounge
You may recall these photographs from the mainstream media in 2009, when an aircraft crashed into an airport building, killing one passenger.
The CRJ-100 had just taken off from Kigali International Airport when the flight crew declared an emergency after developing a fault with the throttle. They returned to Kigali and landed safely.
At the time, all we knew was that the as the ground crew were placing chocks in front of the wheels of the stopped aircraft, the aircraft suddenly accelerated. It veered to the right and proceeded straight through numerous blast fences before crashing into the VIP lounge.
The preliminary information released at the time didn’t offer much explanation. Here’s the basics of what was known initially.
The business jet was a Bombardier Challenger, specifically a CRJ-100 in the Bombardier CL-600-2B19 series. The Challenger was owned and operated by Jetlink, who had acquired the aircraft in June 2007, two years earlier. The Challenger 600 had two Avco Lycoming ALF 502 geared turbofan engines.
The flight was RwandAir WB205, a scheduled international flight from Kigali International Airport to Entebbe in Uganda. Jetlink Express Limited was operating the flight on behalf of RwandAir. On that day, the 12th of November 2009, there were ten passengers and five crew on board, including one flight engineer.
The aircraft took off normally at 12:54. During the initial take-off climb, the flight crew found that the thrust lever for the number one engine (on the left side) had jammed. They could not reduce the engine throttle at all, thus the left-hand engine was still producing take-off power. The captain notified air traffic control and entered a holding pattern. Meanwhile, the first officer requested the help of a company flight engineer who was travelling on the flight, and the two of them attempted to reduce the power to the engine.
After it became clear that they could not unjam the throttle, the captain notified Kigali that they were returning. After one go-around, the Challenger successfully landed on runway 28 and taxied to the parking bay.
There, the aircraft, which had come to a complete halt, suddenly accelerated again while the ground crew were placing the wheel chocks. It turned right and travelled about 500 metres before ramming into the VIP lounge.
At the time, the acting chief executive of RwandAir said they had no idea how this could have happened. He said that their best guess was that the aircraft auto-accelerated, whatever that might mean.
“The captain could not control it. The plane did not get airborne again, it taxied into the building. The captain was taken to the hospital with a broken leg. He has not been able to give us any information so far.”
The Director-General of the Rwandan CAA did not have any better explanation.
“For some unexplained reason, the plane, from the parking spot, took off again at full power and … took a right turn, unexplained, into the technical building,”
A week later, Rwandair made a statement but it was a repetition of the above facts with the addition that the flight crew had been tested for substance abuse. There was still with no hint as to how an aircraft could simply run away like that. Yes, the throttle was still jammed but having successfully landed and taxied to the parking bay, it made no sense as to how the aircraft then suddenly accelerated from a standing stop.
The flight crew recounted the sequence of events later, but they show a similar lack of understanding as to how it could have happened.
…after getting airborne, I asked my copilot to conduct the climb and after takeoff checks. She had a problem with retarding the left thrust lever. I called the engineer to help the copilot retard the left throttle, but it was not possible. I asked the tower controller for permission to land. I landed with one engine on maximum power and landed normally though heavy braking, the tires deflated and parked the aircraft and shortly the plane started rolling downwards toward the barrier and Air Traffic Control Tower building. I had no control over the plane as I even tried to steer it clear of the building.
The first officer:
…after takeoff, I tried to set climb thrust and noticed the left thrust lever could not adjust. I then informed the captain that the throttle was stuck. He tried to adjust it too but it was stuck. We then called for the engineer, we coordinated together and the captain focused on flying the plane safely while I communicated with ATC, did the checklists and we combined efforts with the engineer to try and adjust the left thrust lever. We landed safely and parked but the left thrust lever could not be adjusted still. As we were trying to retard it and shut it down while holding on brakes, the plane started rolling again. Efforts to stop it from rolling failed but the captain managed to control it away from the other traffic. We then hit a wall as the plane could not stop and the thrust lever was still stuck forward.
How could this happen?
Eight years later, the final report has been released by the Republic of Rwanda Ministry of Infrastructure. It’s finally crystal clear what happened.
First, let’s go over the exact sequence of events.
The playback of the Cockpit Voice Recorder makes it clear that the flight crew initially noticed that the thrust lever was sticking during the take-off roll. They decided to continue the take-off as they discussed the problem during the initial climb. The aircraft took off normally with the number one engine at 94% power and the number two engine at 91% power. As they climbed away, the number 2 engine decreased to 69% while the number 1 engine remained at 94%, which fits with the first officer’s description of attempting to reduce the power to climb thrust. The engineer came forward as requested and looked at the left throttle but was unable to rectify the problem. The captain informed the passengers that there wasw a fault and the flight was returning to Kigali.
It’s at this point that the number 1 engine power increased to 97%. The Automatic Power Reserve Command (APR Cmd) was shifted from ‘Not Active’ to ‘Active’, where it remained until the end of the FDR recording.
The aircraft began its descent. Power for the number 1 engine reduced to 96% and the number 2 engine reduced to 32%. Over the next few minutes, the power of the number 2 engine varied between 88% and 27% while the number 1 engine remained steady at around 95%.
The flight crew circled to lose altitude and the captain set up the aircraft for the approach and landing, while the first officer and the engineer continued to try to reduce the power to the number 1 engine. The captain commented that he was concerned that the engine might overheat.
What no one mentioned was any references as to how to deal with the problem. They had access to the quick reference handbook, the flight crew operating manual and the aircraft’s flight manual. At no point did anyone seem to consider that they should look up the problem to discover what the correct procedure was. Nor did they ever think to contact operations on the ground and ask.
The aircraft landed safely on runway 28. The number 1 engine was at 95% power and the number 2 engine at 27%.
The crew continued to discuss the issue as they taxied to the stand. There’s no reference to the full transcript, unfortunately, but an exerpt makes it clear that they had no idea what to do.
First Officer: How will we stop the engine, then?
Captain: We’ll just think it over. It’s a problem.
We’ll just think it over. The concept of looking it up or asking ground personnel for advice just never seemed to occur to them.
At that moment, a warning sounded through the cockpit and the first officer called out, correctly, brake overheat.
The flight crew stopped safely at the stand and shut down the right engine. They called the Senior Cabin Crew member to the flight deck and explained that as the left-hand engine could not be shut down, they would need to exit the passengers on the right side, through the galley door. The captain told the cabin crew not to open the door. “Relax until I give you the green light.”
The senior cabin crew member returned to the cabin and asked the passengers to unfasten their seat belts.
That was when the flight noticed that the aircraft was moving. The ground speed increased to 23 knots as they called out for the engineer.
The senior cabin crew member went to the right side of the aircraft to prepare for the exit. He said that the aircraft began moving at fast speed. Realising that they were likely to collide with something, he and the ground engineer both moved to the front seats and braced behind the bulkhead.
The captain shouted for the chocks to be placed. Then the cockpit voice recorder has only screams and the sounds of the crash.
As the Challenger came to a stop, the cabin crew opened the overwing emergency exits. The passengers were disembarked over the right-hand wing, as the left-hand engine was still running. One passenger was found trapped unconcious underneath the galley area. She was carried out and taken to hospital but died there shortly after.
The captain and the first officer were seated in the cockpit as the flight instrument panel collapsed on top of them. The captain was able to pull himself out from under, but the first officer was trapped and could not free herself.
A second aircraft which was close to the crash site and had just finished boarding, evacuated their passengers safely.
A witness who is not identified in the report saw the aircraft strike the wind barriers and then head uncontrolled towards the building. The fire trucks arrived a few minutes later and sprayed water into the engine intake, hoping to shut down the engine. He said that, ten minutes after the crash, the engine was still running. He forced his way into the cockpit where he found the first officer trapped under the collapsed instrument panel. The witness was searching for the fire shut-off valves when he noticed that the left and right throttles were in the maximum and full power positions. He retarded the throttle to the shutdown position. The engine shut down.
Rescue personnel were able to pull out the first officer about three hours after the crash.
The ambulance which was taking the captain, the engineer and two other passengers to the hospital were involved in a traffic accident and a pedestrian was killed.
So the first issue is to clarify what went wrong with the left-hand engine.
On the 1st of November, 11 days before the accident, an aircraft engineer from Europe was hired and assigned as the supervisor for the routine maintenance of the accident aircraft as it underwent 3500-hour/500-hour/24-month/16,000-cycle routine maintenance in hangar.
During the maintenance, personnel accessed the number 1 engine, opening the cowling.
The engineer who supervised the maintenance confirmed that he had personally closed the cowl door. He was familiar with stowing and security the cowl strut. It’s not clear if he had done so correctly and he was unable to confirm whether the engine cowl doors might have been opened again after the routine maintenance.
During the ground test of the left engine, a rod came loose on the fuel control unit. The engine became uncontrollable. The same engineer had not secured the rod before starting the ground tests and he was only able to shut down the engine by using the fuel shut off valve.
Two days before the accident, on the 10th of November 2009, the aircraft had completed the scheduled routine maintenance at the maintenance base at Nairobi. On the first flight after the maintenance, the aircraft turned back because of a generator problem on the number 1 engine. Maintenance staff resolved the issue and the aircraft was returned to service the same day.
After the accident, engine number 1 was examined. There was no evidence of any mechanical malfunction. However, it had clearly not been stowed correctly.
The upper core cowl support strut was not in its stowed position. The safety pin hung from its lanyard. The position of the strut blocked the fuel control unit actuating arm, so that it would not be possible to reduce the power of engine 1 to less than 93%.
The Challenger had flown 4.6 hours over six flight legs since the last maintenance so it seems pretty likely that the maintenance crew didn’t stow the cowl strut correctly. The vibrations from the engine caused the strut to move and from there it was only a matter of time before it obstructed the engine throttle mechanism.
This was not a new problem, actually.
Bombardier had issued eight service bulletins between August 2000 and January 2009 regarding the upper core cowl support strut.
Incorrect stowage of the Upper Cowl Door Support Rod may result in the rod working itself free during flight. This can result in possible Engine malfunction.
If the support rod for the upper cowl door is not put away correctly, it can fall between the engine fuel – control unit and the throttle control gearbox. The condition can cause an engine malfunction.
The Hold-open Struts on the Core Cowl Doors can be difficult to engage. To facilitate simpler engagement and ensure correct engagement, new Hold-open Struts are introduced. The new Hold-open Struts have a positive latching mechanism. This Service Bulletin gives rework instructions to replace the struts and associated Hold-Open Bracketry on the Tailpipe, for both left and right installations.
Although Bombadier recommended modifications to reduce the risk, the risk of an in-flight throttle jam was considered low, as they had provided updated procedures for dealing with responding to a Throttle Lever Jammed event in the Aircraft Flight Manual and the Flight Crew Operations Manual. As a result, the service bulletins were categorised as ‘discretionary for compliance’. The operator had complied with the initial service bulletins but the one in effect at the time of the accident hadn’t been complied with. The maintenance personnel knew how to correctly stow the cowl but they were not aware of the service bulletins or of the risk of a jammed thrust lever.
So, it was a problem. In fact, Transport Canada published the details of eight incidents involving a thrust lever jam of the left engine for the Bombadier Challenger CL-600-2B19.
The basics of all eight incidents were the same, with one big difference. In every other case, the aircraft landed and parked without further incident. Only in this one case did the aircraft ‘auto-accelerate’ once on the ground.
So what was different?
You’ve guessed by now. It was the flight crew. They simply did not have the training or experience to deal with the issue.
The crew had not yet received training of abnormal procedures during their most recurrency training conducted at 6 months intervals, of which the subject procedure for Thrust Lever Jammed (TLJ) was included.
The [Pilot in Command] got hired by the operator beginning of 2009, obtained CRJ100 rating on 9 March 2009 and was released to PIC upgrade on 24 March 2009 before completing all abnormal procedures.
It’s not just that they didn’t know how to handle the jammed thrust lever, they literally did not know how to deal with a situation where they didn’t know what to do.
The Aircraft Flight Manual and the Flight Crew Operations Manual had been updated with clear instructions for the Throttle Lever Jammed procedure. Of course, that only helps if someone bothers to read it.
Having established that they didn’t know how to stop the engine, they never checked their operational manuals or even asked anyone else as to what to do.
This was only reason why they had to struggle through a landing with one engine at full power.
In every other jammed throttle lever incident documented, the flight crew pressed the FIRE/PUSH button to shut down the left-hand engine and then landed on the remaining one.
The flight crew in this case, however made circumstances for landing unnecessarily difficult. In fact, the final report remarks that considering how badly the flight crew handled the situation, the fact that they landed the aircraft safely under the circumstances was commendable.
Once they had landed, with the engine stuck on full power and the brakes overheating, they made another bad decision.
They really should have come to a halt immediately and evacuated the aircraft. Instead, they decided to taxi to the parking area and stop there.
To be fair, it’s not really a surprise that having managed to land the plane and taxi it to the stand, the flight crew might have felt that the worst of it was over. The captain became completely focused on the evacuation of the aircraft. His priority was to get the passengers out of the aircraft as quickly as possible, which led to them being told to unbuckle their safety belts shortly before the impact.
But the engine was still running and the brakes had overheated. The captain said later that he believed the left tyre deflated as they landed. Clearly, what he didn’t understand was the effects of this on the brakes which he was now relying on to hold them in position.
Skid marks on the runway show that the right tyres were braking after touch down. This backs up the captain’s testimony that the left tyre had deflated during the landing. The brake overheat warning also could have resulted in the deflation of the tyre. Either way, it seems clear that the tyre on the left side was deflated by the time it was parked on the stand.
It’s not clear whether the parking brake was set or not. The parking brake after the accident was in the engaged position, which was unlikely to have happened as a part of the crash, as the handle needs to be turned 90°. However, neither pilot mentioned setting it and the FDR data shows that it was not engaged.
It doesn’t actually matter. Once the left tyre was flat, there was no traction for the brake pressure. If the parking brake was set, then the left tyres would simply have been dragged across the ramp without the brakes doing anything at all.
So while they were discussing the evacuation, it’s likely that the tyre was still deflating or possibly ruptured as the wheel fuse plug melted from the high temperatures of the hot brakes.
Worse, having shut down the right-hand engine, the hydraulic pressure was reducing.
The net effect was that the brakes finally failed completely (or were not set) and the aircraft began to move again as the number 1 engine continued to run at 95% power.
The high power of the engine and the flat left inside tyre left the only braking on the right inside tyre, which explains why the aircraft turned hard to the right as it began to move. Now with nothing to stop it, the aircraft accelerated rapidly, crashing through the jet blast barrier and into the building.
The flight crew’s failure to identify corrective action and their lack of knowledge of applicable airplane and engine systems in response to a jammed thrust lever, which resulted in the number 1 engine operating at high power and the airplane configured in an unsafe condition that led to the need to apply heavy braking during landing.
Also causal was the flightcrew failure to recognize the safety hazard that existed from overheated brakes and the potential consequence on the braking action needed to park the airplane. Contributing factors included the possible failure by maintenance crew to correctly stow the upper core cowl support strut after maintenance, Flightcrew’s failure to follow standard operating procedures, the company’s failure to be availed to manufacturer safety literature on the subject, and the susceptibility of the cowl core support shaft to interfere with the throttle control mechanism when the core strut is not in its stowed.
The final report concludes that ‘the flight crew was certificated, and experienced in accordance with applicable air regulations, but was not fully competent as it.’ They recommend that States of Registry and operators should ensure that at least all abnormal and emergency procedures training be completed.
It’s hard to believe that the final error in the chain of events was so simple. They honestly did not know how to shut down the engine nor that they could look up the answer or ask operations for help.
I can’t help but feel that the root cause is the airline who allowed them to sit in the front of the plane without having the knowledge and skills they needed to handle an emergency.
You can download a PDF for the full report in English from the Ministry of Infrastructure (MININFRA) website.