Pen Air Saab Overrun Unilaska with Injuries

[Lake1952]

A Pen Air Saab Scandia 2000 off the runway in Unilaska with some injuries. Flight had a schools swim team aboard and was marketed as an Alaska Air flight.

https://www.ktuu.com/content/news/Plane-off-runway-at-Unalaska-Airport-563346322.html

https://www.cbsnews.com/news/alaska-plane-crash-peninsula-airways-flight-crashes-unalaska-2-critically-injured-today-2019-10-18/
​​​​​​

[Pilot DAR]

New item here:

https://www.adn.com/alaska-news/2019/10/18/small-plane-crash-reported-at-unalaska-airport/

It looks from the photo that the plane departed the runway to the left before reaching the end.

[DaveReidUK]

Looks like the Saab ran over the runway end chevrons and went over the edge between the Stop sign and the barrier on the right:

https://cimg2.ibsrv.net/gimg/pprune.org-vbulletin/687x434/unalasaka_285793fef0b7ac7381c3fa03fdf2080a963d7642.jpg

[DaveReidUK]

Avherald reports only minor injuries, whereas the Anchorage Daily News, Newsweek and CBS report two critically injured pax - can anyone confirm which account is correct ?

[Lake1952]

NTSB sending a team of 9 to the site.

AS Blog Statements (Flight operated as an AS flight):Incident involving PenAir Flight 3296 marketed by Alaska Airlines

Posted: 2:45 a.m. on Oct. 18, 2019

On Thursday at 5:40 p.m. local time, PenAir Flight 3296 was involved in an incident while landing at Dutch Harbor, Alaska.Right now, our highest priority is the safety and care of everyone who was onboard the flight. Our hearts are with them, their families and employees at PenAir.While the aircraft is operated by PenAir and the crew are PenAir employees, this flight was marketed by Alaska Airlines and the passengers are our guests. At this point, we understand 39 passengers and three crew members were onboard. We’re still gathering information about the welfare of those on the flight and the incident itself.We have established a toll-free hotline for anyone who believes they may have a family member or friend onboard. The telephone number, accessible from the United States and Canada, is 1-888-283-2153. A telephone number that’s accessible from Mexico is 001-877-542-6973.We will continue to work closely with PenAir, who is leading the response. This page will be updated as we learn more information. Posted: 10:15 p.m. on Oct.17, 2019 We have established a toll-free hotline for anyone who believes they may have a family member or friend onboard. The telephone number, accessible from the United States and Canada, is 1-888-283-2153. A telephone number that’s accessible from Mexico is 001-877-542-6973. Posted: 9:30 p.m. on Oct. 17, 2019PenAir Flight 3296 marketed by Alaska Airlines from Anchorage to Dutch Harbor, Alaska was involved in an incident during landing at 5:40 p.m. local time today.There are 39 passengers and three crew members onboard. We’re still gathering information about the welfare of those onboard and the incident itself.PenAir operates the Anchorage-Dutch Harbor service for Alaska with a Saab 2000 aircraft.

Flight 3296 departed Anchorage at 3:15 p.m. local time.

[OldnGrounded]

Avherald reports only minor injuries, whereas the Anchorage Daily News, Newsweek and CBS report two critically injured pax - can anyone confirm which account is correct ?

The local community radio station, KUCB, has been reporting two critical, ten total, injuries, all pax.

https://www.kucb.org/post/ntsb-send-9-investigators-unalaska-after-penair-plane-goes-runway-critically-injuring-two

[Airbubba]

Looks like the Saab ran over the runway end chevrons and went over the edge between the Stop sign and the barrier on the right:

A couple more pictures from social media. I just noticed that these pictures are also in your posting but they are not displayed unless you scroll horizontally on the first picture on some devices.

https://cimg2.ibsrv.net/gimg/pprune.org-vbulletin/1200x900/ehjsg4cucaa98b0_e255e7e3b364bfe279cb4f0607905a8867076e44.jpg
Photo by David Howard

https://cimg3.ibsrv.net/gimg/pprune.org-vbulletin/1918x1080/plane_1_cropped_large__71d5a270e2569b944c6582be78c23d3b97c2c 151.jpg
Photo by Jennifer Wynn

[Airbubba]

Now a report of one fatality.

Press Release: Penair Crash at Unalaska Airport

FOR IMMEDIATE RELEASE

Contact:
Erin Reinders, City Manager City of Unalaska
(907) 581-1251 – telephone [email protected]

PENAIR CRASH AT UNALASKA AIRPORT

Unalaska, Alaska, October 18, 2019 – On Thursday, October 17, 2019 at 5:41 p.m. Unalaska Public Safety officers and Fire/EMS personnel responded to a report of a Saab 2000 aircraft going off the end of the runway while landing at the airport in
Unalaska.

Unalaska Police, Fire and EMS personnel and other emergency responders arrived within five minutes of the event. Unalaska units set up incident command with state assets to manage the incident and assess life and safety issues. Once all safety concerns were identified and mitigated, EMS personnel extracted one patient from the aircraft and completed the evacuation of the remaining passengers. All 39 passengers and crew were accounted for.

The Unalaska Fire Department transported seven patients to the Iliuliuk Clinic and four other patients were brought to the clinic by personal vehicle. Patient injuries ranged from minor to critical. Of the patients transported to the clinic, one was medevaced to Anchorage, and one died of traumatic injuries suffered in the crash. Next of kin has been notified.

Law enforcement has secured the scene pending arrival of NTSB investigators, which may arrive as early as today. The runway and vehicle traffic through the area is shut down until further notice.

Access to Ballyhoo Road is closed and transportation is being coordinated through Unalaska Public Safety. The aircraft may still pose safety risks so the public is asked to stay away from the area.

Updated information will be provided as it becomes available. Thank you for your continued support and patience.




Address/Location
Unalaska Dept. of Public Safety (https://local.nixle.com/unalaska-dept-of-public-safety/)
29 Safety Way
Unalaska, AK 99685

Contact
Emergency: 9-1-1
Non-emergencies: 907-581-1233

[Airbubba]

Weather sequence observed a few minutes after the mishap and posted on Av Herald suggests a tailwind for the landing attempt on runway 13.

PADU 180156Z 30021G27KT 5SM -RA BKN039 06/02 A2953 RMK AO2 PK WND 30027/0154 RAB41 SLP004 P0002 T00560017

[pattern_is_full]

I would not have thought this would end in a death - assuming everyone was seated and belted.

But - if that #1 prop penetrated the fuselage once it disintegrated on the rocks...... :(

Photos here from the left side do show significant fuselage damage above/at the window line, abeam the prop. Possibly a prop root still embedded after scything upwards through seat area.

https://abcnews.go.com/US/high-school-swim-team-members-shaken-safe-plane/story?id=66363658

[Lake1952]

Runway length of 4100 ft (1250m) is obviously marginal in any kind of tailwind or runway contamination. Hard to envision the regular use of 732 Combi aircraft into this station that AS carried out before 2004.

[mickjoebill]

In describing the fatality, the report uses the term “traumatic injuries”

[dash34]

There have been a few deaths from propellor intrusion into the cabin in the last while. Is it not possible to construct a carbon/kevlar panel strong enough to prevent this that would be light enough to install without affecting performance?

[Australopithecus]

Nope. They cannot make an engine containment ring strong enough for turbine blades, so imagine the energy in a much heavier object released from the hub. A friend was once torpedoed by a runaway towbar as they we’re taxiing onto the gate in a Convair 580. The engine was already running down when the towbar hit the prop (at 700 rpm). The blade released and ended up 200m away planted vertically in the ramp.

[Zeffy]

https://cimg6.ibsrv.net/gimg/pprune.org-vbulletin/1184x736/screen_shot_2019_10_19_at_5_49_55_am_931064ebcb42d1680db880e 5d676257cf6f57277.png
https://cimg8.ibsrv.net/gimg/pprune.org-vbulletin/1320x2000/screen_shot_2019_10_19_at_5_51_04_am_aba92b91b8250ea487b22a2 4ed2acc32df646c1d.png

[unworry]

https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/992x744/72339697_2926523784042776_1278481020919742464_n_hpmain_4x3_9 92_e1021d13a665631dde9241986c15df9e9ba1f383.jpg
propeller intrusion

[Capt Fathom]

What was the point of that?

[Zeffy]

The point of posting the approach plate and takeoff mins?

The point of attempting to land under those conditions?

Or the point of designing what amounts to little more than a "cloud-break" procedure?

From AvHerald:
Related NOTAMs:
!DUT 10/064 DUT RWY 13 RWY END ID LGT U/S 1910180706-1910232000EST
!DUT 10/063 DUT RWY 13 VASI U/S 1910180702-1910232000EST
!DUT 10/060 DUT AD AP CLSD 1910180209-1910190209
!DUT 10/062 DUT RWY 13/31 NW 4000FT CLSD EXC MEDEVAC 1910180307-1910190307
!DUT 10/061 DUT RWY 13/31 SE 500FT CLSD 1910180254-1910190254

Metars:
PADU 180356Z VRB04KT 10SM BKN032 BKN042 07/01 A2953 RMK AO2 SLP003 T00670006=
PADU 180256Z 25004KT 10SM BKN035 BKN044 BKN120 07/02 A2952 RMK AO2 RAE14 SLP002 ACSL NE P0002 60004 T00670017 53003=
PADU 180156Z 30021G27KT 5SM -RA BKN039 06/02 A2953 RMK AO2 PK WND 30027/0154 RAB41 SLP004 P0002 T00560017=
PADU 180056Z 31011KT 10SM FEW034 BKN047 BKN060 07/01 A2950 RMK AO2 SLP995 T00720011=
PADU 172356Z 31006KT 10SM BKN036 BKN044 07/01 A2951 RMK AO2 RAB12E25 SLP999 P0000 60000 T00670011 10083 20056 (tel:10083 20056) 53005=
PADU 172256Z 19008KT 10SM BKN037 BKN047 07/01 A2950 RMK AO2 WSHFT 2210 RAB07E18 SLP994 P0000 T00720011=
PADU 172230Z 28008KT 10SM FEW033 BKN060 07/01 A2950 RMK AO2 WSHFT 2210 RAB07E18 P0000=
PADU 172156Z 25008G22KT 10SM BKN055 08/M01 A2950 RMK AO2 SLP993 T00831006=
PADU 172056Z 29010KT 10SM FEW025 BKN034 BKN049 06/02 A2950 RMK AO2 PK WND 29028/2035 RAB34E51 SLP994 P0000 60000 T00610017 53007=

[Capn Bloggs]

Interesting, Zeffy. Not your easy peazy sunny approach, by the looks.

Fathom, back to your C152.

[Airbubba]

Runway length of 4100 ft (1250m) is obviously marginal in any kind of tailwind or runway contamination. Hard to envision the regular use of 732 Combi aircraft into this station that AS carried out before 2004.

The runway is now charted as 4500 feet long. Was it lengthened when it was paved? I remember seeing those Alaska B-737's with the gravel kit in ANC years ago. There are still a couple of B-734 combi's flying for the Department of Energy carrying nuke materials and commandos to guard them.

Would the Saab 2000 be category B for circling?

[Airbubba]

According to the NTSB they had already done a missed approach prior to the overrun landing attempt.

NTSB: Unalaska crash occurred on plane's 2nd landing attemptFriday, October 18th 2019, 11:55 PM AKDTBy: Dave LevalThe National Transportation Safety Board says Thursday night's deadly crash of PenAir Flight 3296 in Unalaska (https://www.ktva.com/story/41199016/commercial-aircraft-off-the-runway-at-unalaska) took place during the second landing attempt.
NTSB Region Chief Clint Johnson said preliminary witness interviews report the plane attempted to land once, then circled back around. The second time, the plane landed then ran off the runway, crossed a road and almost ended up in the water.There were 42 people on the flight — 39 passengers and three crew. Alaska State Troopers on Friday named 38-year-old David Oltman from Washington state as the passenger who was killed. He died from injuries sustained in the crash.
"Our entire team is devastated by this tragic accident," said PenAir President/COO Brian Whilden as he fought back tears in a video statement. "On behalf of PenAir, Ravn Air Group, and all of our employees with the company, I would like to extend our deepest sympathy and condolences to the family and loved ones of our passenger who passed away."
Eleven people, including Oltman, were transported for medical care. One person was medevaced to Anchorage, while others are receiving medical care in Unalaska, according to a release on the City of Unalaska website Friday.The NTSB will now work to find out what caused the crash. Johnson said investigators have plenty of clues.
"The flight data recorder, the cockpit voice recorder, lots and lots of data in addition to the physical evidence," he said. "The airplane itself as far as crush zones, damage to the airplane, all that is going to be studied."
It could take at least a year to find out what caused the crash.
Fuel that had been leaking into the water from the Saab 2000 aircraft has been contained and removed, city officials stated. Remaining fuel in the plane has been removed.
Alaska Department of Transportation & Public Facilities officials have determined the runway was not damaged during the crash, according to a release from the City of Unalaska. A later statement from the city said plans were underway to remove the wreckage on Saturday. The runway and surrounding roadways are expected to reopen once cleared.


https://www.ktva.com/story/41203736/new-information-sheds-light-on-final-moments-of-deadly-penair-crash

[tdracer]

Nope. They cannot make an engine containment ring strong enough for turbine blades, so imagine the energy in a much heavier object released from the hub. A friend was once torpedoed by a runaway towbar as they we’re taxiing onto the gate in a Convair 580. The engine was already running down when the towbar hit the prop (at 700 rpm). The blade released and ended up 200m away planted vertically in the ramp.


Actually, turbine blades are required to be contained (as are compressor blades - including the fan) as a cert requirement. They can come out the back (or front) as low energy debris, but not out the side. Turbine (and compressor/fan) discs are another story - the energy is so high that containment is simply not practical (we're talking inches of armor plate).
I'm not sure how practical it would be to shield passengers from propeller intrusion - the energy of a released prop is going to be very high, so it would take a considerable amount of Kevlar/carbon fiber/whatever to prevent intrusion, and due to the aerodynamic properties of a prop blade, the trajectory is rather unpredictable so the potential impact area of the fuselage would be rather large.
I did some work on another proposed installation of the engine used on the Saab many years ago, and as I recall the prop was certified as primary structure - i.e. it should never fail - because if they lost a prop at power the resultant imbalance could fail the wing structure.
Then again, I don't think the structural cert requirement for the props included impacting a rock wall...

[lomapaseo]

due to the aerodynamic properties of a prop blade, the trajectory is rather unpredictable

Like all blades, the only aerodynamic properties it has after release is drag. If the prop has no outer case to guide it or other blades to interact,, It will go aft

[Lake1952]

The runway is now charted as 4500 feet long. Was it lengthened when it was paved? I remember seeing those Alaska B-737's with the gravel kit in ANC years ago. There are still a couple of B-734 combi's flying for the Department of Energy carrying nuke materials and commandos to guard them.

Would the Saab 2000 be category B for circling?

Bubba, did you notice the displaced thresholds on either end?

[tdracer]

Like all blades, the only aerodynamic properties it has after release is drag. If the prop has no outer case to guide it or other blades to interact,, It will go aft

But it will have a significant forward force on it up to the point where it releases - so it will initially start moving forward. Combine that with the variable affect of forward speed (anywhere from static to the max forward speed of the aircraft) and where the prop might impact the fuselage a significant distance away and you're talking a rather large potential impact area.

[Airbubba]

https://cimg1.ibsrv.net/gimg/pprune.org-vbulletin/508x944/dut_fc209084cb38a3302fdd5e97bfe25ce6ff4dc134.jpg

Bubba, did you notice the displaced thresholds on either end?

I did. It looks to me like the runway length (TORA) is 4500 feet on the approach plate posted earlier. The LDA is charted as 3900 feet. Did the 4100 feet come from Wikipedia or an old approach plate perhaps?

[Lake1952]

Airport diagram in a several year old AOPA airport guide.

[lomapaseo]

But it will have a significant forward force on it up to the point where it releases - so it will initially start moving forward. Combine that with the variable affect of forward speed (anywhere from static to the max forward speed of the aircraft) and where the prop might impact the fuselage a significant distance away and you're talking a rather large potential impact area.

the forward force is "lift" and like a leaf that goes to zero in much less than a second as it stalls. The forward speed of the blade is then matched by the forward speed of the aircraft and that leaves the unbalanced force of fore and aft drag relative to the aircraft

[Airbubba]

Is this the second U.S. Part 121 passenger fatality in the decade since the Buffalo Colgan Air 3407 crash?

[tdracer]

the forward force is "lift" and like a leaf that goes to zero in much less than a second as it stalls. The forward speed of the blade is then matched by the forward speed of the aircraft and that leaves the unbalanced force of fore and aft drag relative to the aircraft

Yes, but the momentum of a propeller blade is significant, and what if you don't have a meaningful forward speed when it lets go (e.g. takeoff power setting)? A prop can move a long way in that fraction of a second before drag takes over. If you're going to try to protect the fuselage, you need to assume it might go forward from the plane of rotation.
For uncontained engine failures I think the assumed scatter trajectory was +10/-10 degrees for high energy debris (-30 degrees for low energy debris IIRC). For a released propeller blade think you'd need assume something like +10/-30 degrees - given the distance from the engine to the fuselage that's a pretty large chunk of fuselage you'd need to protect.

[capngrog]

And then, of course, it could be providing reverse thrust ... I don't know whether or not the SAAB 2000 has reverse thrust (beta) capability, but I believe the SAAB 340 does.

[tdracer]

I'm reasonably sure it has reverse beta capability - it certainly did on the installation I was involved with (the engine was an Allison at that time - which was bought by Rolls about the time I left the project).

[hans brinker]

PADU 180156Z 30021G27KT 5SM -RA BKN039 06/02 A2953 RMK AO2 PK WND 30027/0154 RAB41 SLP004 P0002 T00560017=
PADU 180056Z 31011KT 10SM FEW034 BKN047 BKN060 07/01 A2950 RMK AO2 SLP995 T00720011=

It looks from the photo they landed rwy 13. can anyone confirm....?

[megan]

the momentum of a propeller blade is significantPlenty of experience around, one example.Eastern Air Lines Flight 611 originated at Boston, MA, for Miami, FL, with stops scheduled at New York-LaGuardia Field, NY, and West Palm Beach, FL.
Takeoff from LaGuardia was at 10:09 and the flight climbed to the cruising altitude of 22.000 feet.
At about 13:09, the No.3 propeller failed and a portion of one blade was thrown through the fuselage. It entered the lower right side at the galley section, severing control cables, electrical wires and engine controls, came up through the floor, fatally injuring a purser and left through the upper left side. A momentary fogging of the cockpit resulted due to the sudden depressurization of the fuselage. Heavy vibration was felt and all of the flight and engine instruments became either inoperative or impossible to read. Power was reduced and a rapid descent was started. An attempt was made to feather No.3 engine and orders were given to prepare for ditching.
An estimated one or two minutes after the failure of the No.3 propeller the front portion of No.3 engine and some of its cowling fell free of the aircraft. Concurrently the heavy vibration stopped.
A fire followed in No.3 nacelle but quickly extinguished itself. The crew set course to the Florida coast. At 13:12, the aircraft could no longer transmit because of failure of electrical power.
At the 12,000-foot level the descent was stopped. It was then found that controlled power was available from Nos. 1 and 2 engines, that No. 4 engine was running, although it could not be controlled by its throttle and that No. 3 engine had stopped.
Near the coast low clouds prevailed and the aircraft was let dawn visually to about 1,000 feet altitude, as most of the flight instruments remained inoperative.
The airport at Bunnell. FL, was sighted and circled. All emergency doors and exits were opened on approach and as the aircraft passed over the boundary of the runway and landing seemed assured, the ignition switch of No. 4 engine was cut. Brakes were applied hard during the landing roll causing one of the left tires to blow out.
When the aircraft stopped, fires started in No.4 engine and in the right landing gear. Both were quickly extinguished..
All passengers were evacuated.

https://aviation-safety.net/database/events/dblist.php?Event=ACET

[DaveReidUK]

But it will have a significant forward force on it up to the point where it releases - so it will initially start moving forward. Combine that with the variable affect of forward speed (anywhere from static to the max forward speed of the aircraft) and where the prop might impact the fuselage a significant distance away and you're talking a rather large potential impact area.

i don't think that's necessarily the case. The same principle applies to ice being shed from a blade as does to the release of the blade itself (apart from the vastly different amounts of energy involved, obviously).

In the former case, airframe manufacturers reinforce a relatively small length of the fuselage in the plane of the prop, because that's the part of the cabin that's at risk from ice being shed. Exactly the same applies to the trajectory of a departing blade, except of course that's there's no hope of stopping it if it's heading for the fuselage (which is why it's not supposed to happen).

[physicus]

S2000 does have beta reverse and excellent short field performance. It's so well (over)powered the yaw damper is on the MEL...

[West Coast]

Is this the second U.S. Part 121 passenger fatality in the decade since the Buffalo Colgan Air 3407 crash?

The lady who was sucked out the window on SWA?

[Airbubba]

That's what I'm thinking is the only other Part 121 passenger fatality in the past decade. FedEx, UPS and Atlas have had fatal freighter crashes since the Colgan mishap but is this only the second passenger fatality in this time frame?

[Airbubba]

Here's a passenger's account of the Pen Air landing from a KUCB news article.

Unalaska Passenger Recounts PenAir Plane's Crash LandingBy Nathaniel Herz • Oct 18, 2019

Before Thursday night’s plane crash (https://www.kucb.org/post/ntsb-arrives-unalaska-city-officials-aim-move-damaged-plane-saturday), the PenAir pilot had missed his first attempt at landing flight 3296 in Unalaska. So he circled around for a second try.
Patrick Lee, 57, was on the plane returning home with his wife, daughter, and granddaughter.

He said the approach was bumpy. The engines reversed and the flaps went up. But the plane never slowed down.

“I looked out the window and I could see the terminal, and he was flying to the terminal,” said Lee on Friday morning. “I yelled to my wife, ‘He’s not going to stop — we’re going into the water!’ She yelled at my daughter to hunker down and hold on to the baby as tight as possible, which she did.”

As the Saab 2000 twin-engine turboprop, with 42 people aboard, hurtled toward the water at the end of the runway, the pilot jerked the plane to the right, said Lee. It slid across the road before teetering to a stop on a rocky bank just above the water.

In the process, something — Lee thinks it was debris — crashed through the side of the plane above where his daughter Cody was sitting, with her baby in her lap. The two weren’t severely injured, but a man sitting nearby was knocked unconscious.

Eleven passengers were ultimately taken to the Iliuliuk Family and Health Services clinic, with injuries ranging from “minor to critical,” said a statement from local officials. One passenger, David Oltman, 38, died from traumatic injuries suffered during the crash, and another was medevaced back to Anchorage for treatment.

Unalaska’s 4,500-foot runway sits below Mount Ballyhoo with water on both sides. Pilots often face strong winds, and they’re sometimes forced to divert to a longer jet runway in the community of Cold Bay, nearly 200 miles northeast.

Lee, who runs a heavy equipment shop and has lived in Unalaska for more than 30 years, compares the Unalaska runway to landing on an aircraft carrier. He and his family were returning from a trip to Idaho for medical appointments and seeing friends.

He said the PenAir flight was smooth until it ran into turbulence on its approach to the airport, as it descended below the clouds. Even so, he said, “We’ve landed in far worse weather.”

On the pilot’s first attempt at landing, Lee said, he came higher than normal. “I could look at the ground and knew he was going way too fast,” he said.

After circling around Mount Ballyhoo, the pilot lined up again. He hit the ground “a little hard,” though not unusually so, said Lee. But “it was almost like he didn’t have brakes.”

“It was either that, or he was maybe thinking he could pop back up again,” he said.

The plane ran off the end of the runway and the pilot swerved, which Lee said likely stopped it from sliding into the water.

“It was a pretty violent crash,” he said.

Lee said only one of the plane’s emergency hatches would open, and passengers focused on unloading the women and children onboard, including 11 students from Cordova’s swim team, which was visiting for a meet.

They also tried to move the unconscious man before emergency responders took over, he said.

Lee’s granddaughter had a little bump on her head, and his daughter Cody’s hand was smashed. She went to the clinic afterward to get it checked out, but it was so busy that the family left to allow staff to help more seriously injured passengers.

Lee credited the quick action of volunteers and other emergency responders.

“They were there fast,” said Lee. “The community really came together.”



https://www.kucb.org/post/unalaska-passenger-recounts-penair-planes-crash-landing

The plane has been removed from the scene and put on a barge.

https://cimg0.ibsrv.net/gimg/pprune.org-vbulletin/1200x900/img_4669_5cae9c67e6d92a078e7625e108de792566198942.jpg
Laura Kraegel/KUCB

A cynical but possibly accurate reader comment on one of the news sites:

They'll take a year and spend $30 million on the investigation and come up with the conclusion that the pilot should have landed into the wind.

[lomapaseo]

Yes, but the momentum of a propeller blade is significant, and what if you don't have a meaningful forward speed when it lets go (e.g. takeoff power setting)? A prop can move a long way in that fraction of a second before drag takes over. If you're going to try to protect the fuselage, you need to assume it might go forward from the plane of rotation.
For uncontained engine failures I think the assumed scatter trajectory was +10/-10 degrees for high energy debris (-30 degrees for low energy debris IIRC). For a released propeller blade think you'd need assume something like +10/-30 degrees - given the distance from the engine to the fuselage that's a pretty large chunk of fuselage you'd need to protect.

Your summary is fine, Not practical to protect so lower your window shade and don't think about it :)

Don't confuse momentum fore and aft with polar moment of inertia because the fore and aft considerations are only in regard to the fuselage which is also moving with the prop

[Airbubba]

As is often the case in the code-share era, it appears that the branding of the flight is somewhat convoluted. Isn't Peninsula Aviation Services currently doing business as PenAir as a successor to the original company using a Part 121 certificate?

From the Anchorage Daily News:

The flight, PenAir Flight 3296, was sold by Alaska Airlines, which markets flights to Dutch Harbor. But the actual operator is more complicated.

The carrier is owned by Ravn Air Group and not PenAir, an Alaskan flying institution started in 1955 that filed for bankruptcy in 2017. A company called Peninsula Aviation Services Inc., which is affiliated with Ravn, bought PenAir’s assets (https://www.adn.com/business-economy/2018/10/04/ravn-affiliate-wins-penair-bankruptcy-auction-with-12-3-million-bid/) last year in Chapter 11 proceedings.

The name on the federal certificate as the operator of the plane involved in the crash is Peninsula Aviation, a Ravn spokeswoman said. But information distributed after the crash by Ravn and Alaska Airlines repeatedly names PenAir as the operator of the flight.

Orin Seybert, who founded PenAir, said his company had nothing to do with the crash except for providing the leased Swedish Saab 2000 twin-engine turboprop that’s in the process of getting transferred over to the new owner.
PenAir did not provide or train the pilots, Seybert said.

"I really hate that it happened,” he said. “But this accident was not PenAir’s fault even though it was one of our airplanes. That was a total Ravn operation.”

Ravn operates several Alaska-based air services under its brand: Corvus Airlines (the former Era Aviation), Hageland Aviation Services and Frontier Flying Service. The company overhauled its management team (https://www.pprune.org/left=https://www.adn.com/alaska-news/aviation/2017/10/01/ravn-alaska-under-scrutiny-after-fatal-crashes-gets-new-management/) in 2017 following scrutiny for a 2016 crash in Southwest Alaska that killed two pilots and their passenger.


https://www.adn.com/alaska-news/2019/10/18/one-passenger-died-in-unalaska-crash-of-penair-plane-flying-from-anchorage-that-injured-10-others/

[Lake1952]

So the last two US passenger fatalities were both due to penetration of the hull by engine/prop shrapnel.

[capngrog]

i don't think that's necessarily the case. The same principle applies to ice being shed from a blade as does to the release of the blade itself (apart from the vastly different amounts of energy involved, obviously).

In the former case, airframe manufacturers reinforce a relatively small length of the fuselage in the plane of the prop, because that's the part of the cabin that's at risk from ice being shed. Exactly the same applies to the trajectory of a departing blade, except of course that's there's no hope of stopping it if it's heading for the fuselage (which is why it's not supposed to happen).

You make an excellent point; however, I'm not sure that the loads on ice accretions are the same as those on the propeller blades. While the blades themselves have significant fore and aft loads (depending on propeller pitch), ice accretions might not. Ice accretions usually occur initially at the propeller blade root and often extend spanwise (blade span) along the leading edge of the blade toward the tip. My theory (guess?) is that ice accretions, being on the leading edge of the blade, do not experience pressure differentials created by the blade pitch and are primarily affected by centrifugal force. If that is the case, then the ice could be expected to shed within the plane of propeller rotation. Due to the aerodynamic loads on the propeller blade, I would expect that it (they) would shed out of the plane of propeller rotation.

Below is a link to a study titled, "Propeller Icing Tunnel Test on a Full-Scale Turboprop Engine" conducted under the auspices of the FAA (U.S.). I must admit that nowhere in the study was I able to find any specific reference to aircraft propeller ice shedding patterns; consequently, my theory is no more than a somewhat insufficiently educated guess (speculation).

The link:

http://www.tc.faa.gov/its/worldpac/techrpt/ar0660.pdf

[GordonR_Cape]

That's what I'm thinking is the only other Part 121 passenger fatality in the past decade. FedEx, UPS and Atlas have had fatal freighter crashes since the Colgan mishap but is this only the second passenger fatality in this time frame?

For the record, I noticed that on aviation-safety.net database the 2018 Southwest Airlines flight 1380 incident is classified as Forced landing on runway, instead of Uncontained engine failure. See: https://aviation-safety.net/database/record.php?id=20180417-0

[OldnGrounded]

For the record, I noticed that on aviation-safety.net database the 2018 Southwest Airlines flight 1380 incident is classified as Forced landing on runway, instead of Uncontained engine failure. See: https://aviation-safety.net/database/record.php?id=20180417-0

Right, but the narrative for SWA 1380 begins:

Narrative:
Southwest Airlines flight 1380 diverted to Philadelphia Airport, Pennsylvania, USA, after suffering an in-flight uncontained engine failure.

We know it was an uncontained failure, because we've all seen the separated parts and the results of the impact on the fuselage (although the parts that hit the fuselage appear to have been pieces of cowling, not fan blades).

[DaveReidUK]

For the record, I noticed that on aviation-safety.net database the 2018 Southwest Airlines flight 1380 incident is classified as Forced landing on runway, instead of Uncontained engine failure. See: https://aviation-safety.net/database/record.php?id=20180417-0

I think you're reading too much into the ASN classifications, which aren't intended to be definitive.

In ICAO-speak, the SWA occurrence would be classed as SCF-PP (System/component failure or malfunction - powerplant).

See Aviation Occurrence Categories - Definitions and Usage Notes (http://www.intlaviationstandards.org/Documents/OccurrenceCategoryDefinitions.pdf)

[Pilot DAR]

I think that the FAA's perception of propeller blade shedding can be inferred from the design requirement:

Sec. 25.771
Pilot compartment.
............
(b) The primary controls listed in Sec. 25.779(a), excluding cables and control rods, must be located with respect to the propellers so that no member of the minimum flight crew (established under Sec. 25.1523), or part of the controls, lies in the region between the plane of rotation of any inboard propeller and the surface generated by a line passing through the center of the propeller hub making an angle of five degrees forward or aft of the plane of rotation of the propeller. .............

This is a design requirement for the pilots, but not the passengers.

Ice seems to come off a propeller in line with the disc. I have broken a propeller ice panel on the baggage door of a light twin with the prop shedding large chunks of ice in flight. The contact point was directly in the prop disc. A blade developing thrust would seem to have a range of 5 degrees if the FAA design requirement is an indicator of experience. It's worth noting that the instant that a blade might release from the hub, it stops developing thrust, so would be acted upon by centrifugal force primarily. That said, when the propeller damage is a result of a ground strike, I imaging the physics change.

[MechEngr]

The separated blade should retain its angular momentum and instantaneously spin like a boomerang about its CG at the same rate as it was rotating when attached. The path of the CG of the blade will instantaneously be tangential to the rotation as the force causing the blade path to curve around the axis of the powerplant will no longer exist.

Over long distances lift and drag will cause the blade to undergo a complex trajectory, but over a very short distance and time, the blade should maintain its orientation and convert some of that angular momentum into lift, much like a maple seed does, though the advance relative to the airframe may not be much in that interval. I'd say that 5 degrees forward is easily a possibility and that armoring a plane against a blade is unrealistic.

[pattern_is_full]

I notice that about 90% of pictures showing actual results of props shedding and hitting a fuselage show that the blades hit within a fraction of a degree of the prop plane. (Image-google "propeller blade separation").

And there is a good explanation for that - whatever the aerodynamic forces acting on the blade after separation, they likely have on the close order of 0.01 seconds to produce any relative motion fore or aft. Once released, the blade will be moving laterally at about 94m/s (339kph), and only has to travel about 1m to reach the fuselage (assuming it will hit it at all) on most transport turboprops.

(Math is: 2-meter blade, 1-meter half radius. At 1800 rpm the half-radius is travelling at pi x R x 1800 = 5654m/minute or 94m/sec average speed. The motion of a thrown blade is a pirouette around the blade's center of mass, tip over root, like a thrown boomerang).

Interestingly, in cases where the entire prop departs (gearbox failure and such) the prop almost always moves backwards - drag on a now-unpowered prop vastly outweighing any residual thrust. E.G. Marine C-130 crash, Mississippi, 2017 - #3 prop knocked loose by shock of #2 blade failure, went backwards over the top of the fuselage like the iceberg ripping the Titanic - dit-dit-dit.

As to "prop-proofing" the fuselage: at 94m/s, a prop blade will be travelling at only ~1/8th the velocity of an average firearm muzzle velocity (800m/s) - but it will mass about 333 times as much as the average bullet and have around 40 times the momentum (p, where p=mv). Kevlar won't do much - would require steel armor. Assumes bullet masses 30g and prop blade masses 10kg.

[dash34]

I think that the FAA's perception of propeller blade shedding can be inferred from the design requirement:



.. It's worth noting that the instant that a blade might release from the hub, it stops developing thrust, so would be acted upon by centrifugal force primarily. ..

At the instant the blade is released from the hub there is no more centripetal force. It is now obeying Newton's first law w.r.t. the hub and will go in whatever direction it was already going. OTOH, it is subject to drag and lift from the shape of the blade and the airflow around it,so there are other forces acting on it that might cause it to accelerate in one direction or another.

[DaveReidUK]

"Centrifugal force", while non-existent, is a useful concept. :O

OTOH, it is subject to drag and lift from the shape of the blade and the airflow around it,so there are other forces acting on it that might cause it to accelerate in one direction or another.

Correct. But, as pointed out above, if the blade is going to hit the fuselage it will do so almost instantaneously, so the time during which those aerodynamic forces can alter its trajectory is very short indeed.

[EDML]

Even if the area could be reduced - the energy of a released prop blade is so high that is is nearly impossible to protect the fuselage and passengers.

There have been cases in the past were a released prop blade nearly cut the fuselage in half.

[dixi188]

Several Lockheed Electras have lost props, sometimes with considerable damage.
I know of four incidents, none of which caused injury to the occupants.
I never sit in line with the props.

[DaveReidUK]

Yes, something like that can ruin your entire day. :O

https://cimg2.ibsrv.net/gimg/pprune.org-vbulletin/325x362/n285f_electra_d2c035fa1e36d061a370e8e9f167e9165f524c9e.jpg

Note both the entry and (on top of the fuselage) exit holes made by the errant blade. Ouch.

https://www.flickr.com/photos/24101413@N03/4521923524

[lomapaseo]

Correct. But, as pointed out above, if the blade is going to hit the fuselage it will do so almost instantaneously, so the time during which those aerodynamic forces can alter its trajectory is very short indeed.

It's all relative. In order to hit the fuselage above the window line a clockwise spinning prop has to come from the outboard side of the nacele and up and over the wing (not sure which way the subject event went) Take the center of mass velocity vector and degrade it against drag and then add in the axial drag against the direction of the fuselage flight and you get a curvilinear result. As others have said the blade is also tumbling about it's center so if it is a long slender shape two impact points may be realized. followed by gash intersections between them.

I used to compute this kind of stuff and then to document the results in a Photo sketch by overlaying wet spaghetti strands (along the trajectories) on my many aircraft models that I kept in my office.

In the 90's we reviewed all the documented instances of prop sheds in an effort to update the containment regulations. After consideration it was decided the the tried and true method of designing the aircraft flight controls to best possible redundancy and separation (keel beam to crown) was the only practical way forward. (of course we always kept the pilot out of harms way)

[DaveReidUK]

It's all relative. In order to hit the fuselage above the window line a clockwise spinning prop has to come from the outboard side of the nacele and up and over the wing (not sure which way the subject event went)

The Saab's props do indeed rotate clockwise when viewed, as is the convention, from the rear.

Take the center of mass velocity vector and degrade it against drag and then add in the axial drag against the direction of the fuselage flight and you get a curvilinear result. As others have said the blade is also tumbling about it's center so if it is a long slender shape two impact points may be realized. followed by gash intersections between them.

I don't think any posters are disputing that the trajectory will be curvilinear. Just not very.

[lomapaseo]

Well I wouldn't want to sit one row behind it :)

[FrequentSLF]

Even if the area could be reduced - the energy of a released prop blade is so high that is is nearly impossible to protect the fuselage and passengers.

There have been cases in the past were a released prop blade nearly cut the fuselage in half.

Since the area could not be reduced and the energy is so high of a released prop blade, I do agree that is nearly impossible to protect the fuselage, however there is a way to protect the passenger, do not have seating on the projected trajectory of the released prop blade

[EDML]

Nice pictures showing the energy and also width of the damage. Interesting how wide the damage on the exit side of the blade is:

AVSIG: Slight C-130 Problem... (http://www.aero-farm.com/ubbthreads/showflat.php?Cat=0&Number=353294&page=0&fpart=all&vc=1)

This one crashed because the fuselage failed after a prop blade separation (CV580):

https://aviation-safety.net/database/record.php?id=19670305-0

[DType]

Used to have the toilets in the plane of the props.

[Euclideanplane]

I
And there is a good explanation for that - whatever the aerodynamic forces acting on the blade after separation, they likely have on the close order of 0.01 seconds to produce any relative motion fore or aft. Once released, the blade will be moving laterally at about 94m/s (339kph), and only has to travel about 1m to reach the fuselage (assuming it will hit it at all) on most transport turboprops.

(Math is: 2-meter blade, 1-meter half radius. At 1800 rpm the half-radius is travelling at pi x R x 1800 = 5654m/minute or 94m/sec average speed. The motion of a thrown blade is a pirouette around the blade's center of mass, tip over root, like a thrown boomerang).


It is relevant for this particular event that the propellers on a Saab 2000 stay below 1100 rpm at normal operation, though possibly marginally above that while reversing. The numbers become correspondingly lower. Although separating at collision with a rock wall makes its trajectory after separation very unpredictable, to say the least. Compare with a bullet shot out of a rifle versus a musket gun.

[dixi188]

In the interests of accuracy: although this aircraft carries Channel Express markings, it was on wet lease from Renown Aviation of Florida at the time of the incident and Channex crews had no involvement in the attempted wheels up landing.

[CargoOne]

Used to have the toilets in the plane of the props.

Many Soviet prop aircraft had either a galley or toilet or wardrobe in line of props... Like this one


https://s00.yaplakal.com/pics/pics_preview/1/6/8/5806861.jpg

[rog747]

Indeed BEA Vanguards and BMA Viscount 800's have no windows in the prop line, nor does a Britannia 300

[lomapaseo]

Probably as effective as a closed window shade

[rog747]

Funny - however as someone chipped above I think Loo's and or galleys were in-situ there back in the day

[DaveReidUK]

The certification requirement to protect the fuselage from ice that might be shed by the prop normally precludes windows in the prop plane.

[JPJP]

Indeed BEA Vanguards and BMA Viscount 800's have no windows in the prop line, nor does a Britannia 300

The humble Jetstream 31/32 (BAe 3100) had an interesting issue with the Bull Gear in the Garrett engine. It would depart the engine and travel through Row 1, then exit the other side of the fuselage. That engine was renowned for its shedding of various fast moving parts. Lovely airplane, quite a vigorous experience to fly, and passengers loved it. (Some sarcasm involved).

Regarding the Penn Saab; I found the choice of landing direction interesting given the winds. I can see why it would appear attractive given the prohibition on circling NE and the terrain. But I do wonder what the tail wind was like landing to the SE ?

[megan]

Several Lockheed Electras have lost props, sometimes with considerable damage.
I know of four incidents, none of which caused injury to the occupantsThere was a P-3 (A9-754) that ditched which had full power applied at the time. All four props separated and moved to starboard, the #2 entering the fuselage and killing one of the crew, damage also caused to the starboard outer wing leading edge, aileron and flap.

[pattern_is_full]

Regarding the Penn Saab; I found the choice of landing direction interesting given the winds. I can see why it would appear attractive given the prohibition on circling NE and the terrain. But I do wonder what the tail wind was like landing to the SE ?

Depends on how tightly a 2000 can maneuver. There is a rocky ridge rising to 1500 feet directly in line with, and about 2 miles (2 runway lengths) from, the threshold of 30.

Which is why the "missed" from an approach to 12 (was 13) is an immediate U-turn before even reaching the airport. https://aeronav.faa.gov/d-tpp/1911/06367RB.PDF

A couple views here (item #3) from the airport out along the approach path to 30: https://www.thealaskalife.com/blog/scary-alaska-airport-runways/

Not a huge deal if you can fly downwind, base and final at 70 knots - tight noogies if you have to fly it at 100kts+.

[Australopithecus]

That Electra isn’t the only example of a number 1 blade release decapitating number 2...a similar event occurred at Thunder Bay, Ont in the 1980’s (They went on to have goose shatter the captains windscreen on final approach).

I recall a story of a CV-580 putting a prop into a snow drift (12” clearance). Apparently a blade released and killed a pax.

When I flew DC-3s the joke was that if you had to ditch the captain would be happy to trade seats.

An Allison 606 prop blade (Electra, CV58) weighs a lot more than 10kgs...maybe 35 kgs.

[Pilot DAR]

An Allison 606 prop blade (Electra, CV58) weighs a lot more than 10kgs...maybe 35 kgs.

I expect more than that! I understand that the whole (Hamilton Standard) prop weighs in the range of 1000 pounds. But any prop blade releasing into the fuselage side is going through.

There is a rocky ridge rising to 1500 feet directly in line with, and about 2 miles (2 runway lengths) from, the threshold of 30.

I'm not a Saab pilot, so am not qualified to comment on an acceptable approach for that type. However, I know that Dash 8-100/200 are operated out of shorter runways, with tighter approach paths than that airport. Honnigsvag, Norway, is one example, mountains within a mile of each end of the runway. For my very limited experience flying larger planes, a 20 knot tailwind into any runway would give me the willies!

[Australopithecus]

I forgot how heavy they are. I had one mounted in a granite block in my living room (when I was single). I recall now it took three of us to get it off the truck. They are hollow ribbed steel.

[dixi188]

It's the Aero Products prop that has the steel blades. Electra prop weighed nearly half a ton.
Hamilton Standard props have Aluminium blades and weigh less.

[Airbubba]

The NTSB has issued an investigative update.

National Transportation Safety Board Washington, D.C. 20594
Aircraft Accident Investigative Update

WASHINGTON (Nov. 15, 2019) – The National Transportation Safety Board (NTSB) has issued an investigative update for the October 17 runway overrun at Tom Madsen Airport, Unalaska Alaska (Port of Dutch Harbor).

On October 17, 2019, about 1740 Alaska daylight time, PenAir flight 3296, a Saab 2000, N686PA, overran the runway while landing at the Thomas Madsen Airport (PADU), Unalaska, Alaska. The airplane passed through the airport perimeter fence, crossed a road, and came to rest on shoreline rocks. Of the 42 passengers and crewmembers on board, 1 passenger was fatally injured, and several other passengers sustained serious or minor injuries. The airplane received substantial damage. The regularly scheduled domestic passenger flight was operating under the provisions of 14 Code of Federal Regulations Part 121 from Ted Stevens International Airport (KANC), Anchorage, Alaska, to PADU.

Figure 1 View of the airplane looking northwest

The NTSB traveled to the scene of the accident, and the following investigative groups have been formed: Operational Factors, Human Performance, Systems and Structures, Powerplants, Aircraft Performance, Airports, Survival Factors, Flight Data Recorder, Cockpit Voice Recorder, and Maintenance Records.

Parties to the investigation include: The Federal Aviation Administration, PenAir and Rolls-Royce Engines.

In accordance with International Civil Aviation Organization (ICAO) Annex 13, the Swedish Accident Investigation Authority (SHK) has appointed an Accredited Representative as State of Design and Manufacture of the airplane. The Swedish Accredited Representative has appointed Saab as a technical advisor.

Investigators documented the airplane wreckage and runway marks, reviewed maintenance records, examined security camera and passenger video, and interviewed the accident crew members and other PenAir personnel from October 18 through 27 in Dutch Harbor and Anchorage.

According to the flight crew, the captain was the flying pilot and the first officer was the pilot monitoring. The first officer stated that he completed the performance calculations during cruise, before beginning descent, and prior to obtaining the weather at DUT. The flight crew indicated that they conducted a go-around during the first approach to runway 13 because they were not stabilized. On the second approach, the flight crew indicated they touched down about 1,000 feet down the runway and the captain initiated reverse thrust and normal wheel-braking. The captain stated that he went to maximum braking around the “80 knot call.” The flight crew reported that they attempted to steer the airplane to the right at the end of the runway to avoid going into the water.

The captain indicated he held an airline transport pilot certificate and had accumulated about 20,000 total flight hours of which about 14,000 hours were in the DH-8 and 101 hours were in the Saab 2000. The first officer indicated that he held an airline transport pilot certificate and had accumulated 1,446 total flight hours of which 147 were in the Saab 2000.

RUNWAY INFORMATION

Runway 13/31 at PADU is 4,500 feet long and 100 feet wide with a grooved asphalt surface.

There was a 5-inch-wide dark rubber witness mark on the runway, about 15 feet left of the runway centerline, starting about 1,840 feet from runway 13’s displaced threshold, and extending about 200 feet. Small fragments of tire were found in the area near the end of the dark rubber witness mark. The left main landing gear outboard tire was found deflated with an area that had worn entirely through the tire. There were dark rubber deposits marks from all except the outboard left main landing gear tires, in an arc to the right in the runway overrun area.

Witness marks indicated that the airplane departed the runway and overrun area, traversed a section of grass, impacted a 3- to 4-ft high chain-linked perimeter fence with evidence of left engine propeller contact, crossed a ditch, impacted a large rock, and crossed a public roadway. The left wing or left engine propeller struck a 4 to 5 ft vertical signal post on the opposite shoulder of the road and the left propeller struck a 6 to 8-ft high yellow diamond shaped road sign. There were strike marks consistent with the right engine’s propeller tips contacting the ground near where the airplane came to rest.

AIRPLANE AND ENGINES

Examination of the fuselage revealed a hole and impact damage to the left forward fuselage around the 5th window (Figure 2). A propeller blade was loosely stuck in the surrounding structure external to the fuselage and another propeller blade was found inside the fuselage All cabin seats (15 rows) were intact and secure to the floor except for seat 4A, which was displaced and damaged. The damaged area of the cabin was contained within the area on the left side between fuselage station (FS) 399 (seat 3A) and FS 488 (seat 6A), with extensive damage evident at FS 435. The left side overhead compartment (FS 399 to FS 488) partially separated from its mounts and descended about 6 to 12 inches with various brackets and debris hanging down into the seats or laying on the floor. The wall panel separated at FS 399 and displaced rearward and inward. The 4A window fuselage frame was located on the cabin floor.

Figure 2 Damage to left side of fuselage

The four main landing gear tires and brake assemblies were examined. The tire pressures were documented prior to moving the airplane. The left outboard (OB) tire was deflated, the left inboard (IB) tire was at 135 psi, and the right OB and IB tires were greater than 160 psi (maximum reading of the available gauge). A placard on the landing gear door indicated that the allowable pressures were between 165-173 psi. The left OB tire had an area approximately 11 inches in length which was ground flat all the way through the tire (See Figure 3). Because hydraulic pressure could not be applied, only an approximate measurement of the carbon brake wear pins could be made. All wear pins were in serviceable ranges. Multiple components were removed for further examination and testing.

Figure 3 Damage to left outboard tire

The airplane was equipped with two Rolls-Royce AE2001A turbo-propeller engines, and two Dowty R381/6-123-F/5 with 6 composite propeller blades. The left engine remained attached to the engine mounts although the right rear engine mount was bent and buckled. The left engine’s propeller assembly was sagging in relation to the engine’s nacelle with the propeller shaft support resting on the forward part of the nacelle. The propeller reduction gear box front housing was fractured 360° around forward of the diaphragm and in the area of the ring gear. There were broken pieces of gear box housing laying on the engine deck. The propeller shaft aft bearing was missing 3 adjacent rollers from the bearing cage. The engine was complete and did not have any indications of an uncontainment or case rupture. The left propeller hub was intact. Pieces of the blade, either part of the airfoil or just the blade butt, remained in all six blade locations on the hub. All of the propeller blades that still had parts of an airfoil remaining were in the feathered position. At those locations were the airfoil was missing, the blade retaining clamps were in place with the retaining bolts in place and safety wired. Blades Nos. 1, 3 and 4 were missing, although the base of each blade remained in the hub (Figure 4). Blade No. 2 was in place in the hub and the blade was fractured transversely across the airfoil about 28.5-inches from the disk to the fractured end. Blade No. 5 was broken about 48-inches from the disk to the fractured end. Blade No. 6 was broken about 48-inches from the disk to the fractured end. The three blades that were missing from the left engine’s propeller hub were all recovered. There was a propeller blade that was recovered from the water that was about 48-inches long. There was a propeller blade that was found hanging on the left side of the airplane from some cabin insulation that had wrapped around the butt of the blade that was about 58.5-inches long. There was a blade that was found in the cabin that was about 57-inches long.

Figure 4 Damage to left propeller

The right engine was intact and the propeller could be turned with little force. All six right propeller blade tips were broken away between 53.5-inches and 56.75-inches from the disk to the fractured end (Figure 5).

Figure 5 Right engine and propeller

FLIGHT RECORDERS

The flight data recorder (FDR) and cockpit voice recorder (CVR) were sent to the NTSB Vehicle Recorder Laboratory in Washington, DC, for readout. The initial examination of the recorder data revealed the following:

Flight Data Recorder

• More than 43 hours of recorded data was recovered, including, aircraft basic, engine, autoflight, flight controls, and warning and status parameters.
• The accident flight was recorded, and the accident was captured at the end of the recording.
• Touchdown occurred with the aircraft traveling at about 129 knots indicated airspeed and 142 knots ground speed.
• Following touchdown, the aircraft decelerated reaching a peak deceleration of -0.48 g., with the engines operating in reverse mode.
• About 25 seconds after main gear touchdown, a change in aircraft pitch and roll, and an increase in the magnitude of triaxial acceleration forces was recorded, consistent with the aircraft departing the runway surface. The engines were taken out of reverse mode and ground speed at that time was about 23 knots.

Cockpit Voice Recorder

• The CVR group convened from October 29-31, 2019 in the Recorder Laboratory to complete a transcript of accident flight.
• The flight was cleared for the RNAV runway 13 approach into PADU.
• Weather was initially reported (by the local weather observer) as winds 210 degrees at 8 knots, gusting to 14 knots, visibility 7 to 10 miles, a ceiling at 4,300ft that was broken, a temperature of 8 degrees Celsius, a dew point of 1 degree Celsius, and an altimeter setting of 29.50 inches Hg.
• A later transmission from the local weather observer to another aircraft reported the winds were 180 degrees at 7 knots, visibility 8 to 10 miles with showers in the vicinity, and a broken ceiling at 3,900 feet.
• The aircraft was configured for the approach: flaps 20, gear down.
• During the approach, the winds were reported as 270 deg at 10 knots.
• A go-around was executed, and the flight returned for a visual approach to runway 13. During the go-around, the winds were reported as 300 degrees at 8 knots.
• After the go-around, the winds were reported to be 290 at 16 gust 30 (multiple overlapping radio transmissions occurred at this time). Transmissions between the weather observer and another airplane indicated that winds favored runway 31 but could shift back to runway 13.
• The aircraft was configured again for the approach: flaps 20, gear down.
• During the second approach, winds were reported as 300 degrees at 24 knots.
• The aircraft touched down and the roll out lasted approximately 26 seconds until the aircraft departed the runway.
• The crew announced, over the PA, an evacuation out the right side of the aircraft and made a radio call for assistance
• The CVR transcript will be released when the public docket is opened.

Additional information will be released as warranted.




https://cimg7.ibsrv.net/gimg/pprune.org-vbulletin/1261x947/fig_1_c78996c5dacae6d44075e5f1e74d0af0481da3e9.jpg
https://cimg4.ibsrv.net/gimg/pprune.org-vbulletin/923x692/fig_2_copy_7a8e708339eb30a78a5feabb71a987abefcd66e3.jpg
https://cimg9.ibsrv.net/gimg/pprune.org-vbulletin/724x970/fig_3_d7cae0df97919ea7e11250151a5636233179ec36.jpg
https://cimg6.ibsrv.net/gimg/pprune.org-vbulletin/922x696/fig_4_3722d55efec99b214b623b0547619a2c08d86ca8.jpg
https://cimg5.ibsrv.net/gimg/pprune.org-vbulletin/900x675/fig_5_copy_f86c8fc2d4aabb3e39e3aec1870cc8389017b305.jpg

[Airbubba]

During the second approach, winds were reported as 300 degrees at 24 knots.

Well, at least they knew about the tailwind I suppose.

[Zeffy]

Cockpit Voice Recorder

• The CVR group convened from October 29-31, 2019 in the Recorder Laboratory to complete a transcript of accident flight.
• The flight was cleared for the RNAV runway 13 approach into PADU.

I wonder if that's a typo? Or are there Special RNP AR procedures at PADU?

[Airbubba]

The NTSB Accident Docket has been opened:

NTSB Docket - Docket Management System (https://data.ntsb.gov/Docket?ProjectID=100437)

[Pilot DAR]

It looks like the figure 4 and 5 references to left and right propellers need some correction....

[DaveReidUK]

Can't see any obvious errors - what do you have in mind ?

[Pilot DAR]

Yes Dave, You're right, comment withdrawn. If not read as "section headings" (as I did in error) the rights and lefts make sense in context.

[exlatccatsa]

As soon as I saw the photo of the Saab sitting above the water, I was reminded of Dan Air KF at Sumburgh. At least the Saab didn't enter the water and saved a lot of lives.

[Count Dracula]

I have a history with this particular airplane as I flew it between 2005 thru 2015. Overall, I flew the 2000 for 20 years. I completed the acceptance flight for it in Orebro Sweden in 2005 for my employer. Also delivered it to Bangor, ME in January 2016 for 121 modifications. I also delivered another 2000 to PenAir in PANC in late 2015. I had a long discussion with the PenAir CP about the airplane & the potential risks it posed flying into Dutch Harbor. Circling to a short RWY with terrain, wind, low weather was discussed. I didn’t feel a warm, fuzzy when I left. They had made a commitment to the airplane, had the flap 35 auto retract modification completed & steep approach mod, I believe, as well. Just wasn’t the right airframe for that airport. Great range & enroute speed, but higher Ref speeds that reduced safety margins.

[Airbubba]

Maintenance Error Leads to Fatal Runway Overrun Accident
11/2/2021

​​WASHINGTON (Nov. 2, 2021) — A passenger turboprop airplane overran a runway because its braking system was compromised by incorrectly wired anti-skid sensors, the National Transportation Safety Board determined in a public meeting Tuesday. On Oct. 17, 2019, PenAir flight 3296, a Saab SA-2000, overran the runway during a landing attempt in Unalaska, Alaska. The airplane crashed through the perimeter fence, crossed a road, and came to rest on shoreline rocks on the edge of Dutch Harbor. One passenger was killed; another was seriously injured; and eight sustained minor injuries, mostly during the evacuation. The flight crew, the flight attendant, and the other 29 passengers were uninjured.

“Even though the airplane, the pilot, the weather and federal oversight all had a role in this tragedy, it was entirely preventable,” said NTSB Chair Jennifer L. Homendy. “The brake system should have been designed to protect against human error during maintenance, the pilot shouldn’t have landed on a runway with such a strong tailwind and federal regulators should have considered the runway safety area dimensions when authorizing the airline to fly the Saab 2000 into that airport.”

A post-accident examination of the airplane revealed sensors for the anti-skid system had been incorrectly wired during an overhaul of the left main landing gear. This configuration led to the skidding and bursting of one tire and the subsequent release of brake pressure on two of the three remaining wheels. Investigators determined the loss of effective braking on three of the four main landing gear wheels prevented the flight crew from stopping on the runway.

Noting that systems should be engineered to prevent human errors that could occur during maintenance, the NTSB recommended Saab redesign the landing gear wheel speed sensor wiring to reduce the probability of a miswiring during maintenance operations. Because the captain elected to land on a runway with a reported tailwind that exceeded the airplane manufacturer’s operating limit, the airplane touched down with a higher-than-normal groundspeed. The NTSB said the decision to land with such a tailwind was “intentional, inappropriate, and indicative of plan continuation bias.”

The NTSB also found that when the Federal Aviation Administration approved PenAir to fly in and out of the Unalaska airport with the Saab 2000, they did not recognize that the safety area beyond the end of the runway did not conform to the recommended safety criteria for an airplane in that design category.

The NTSB issued six recommendations to the FAA, three to the European Union Aviation Safety Agency and one to Saab. The executive summary, probable cause, findings, and safety recommendations are in the report abstract available on the investigation web page.

The final report will be published on the NTSB website in several weeks.

The accident web page, including links to an investigative update, the public docket and related news releases, is available.

Maintenance Error Leads to Fatal Runway Overrun Accident (ntsb.gov) (https://www.ntsb.gov/news/press-releases/Pages/mr20211102b.aspx)

[punkalouver]

Video of accident……….

https://www.instagram.com/reel/Cd_0oKVAxaU/?igshid=YmMyMTA2M2Y=