The Saras PT 2 VT- XRM Accident Report- A note of dissent. Prof. Prodyut Das

Whilst collecting material for a feature on prospective Indian Civil aircraft I came across the DGCA report “The Final report on the investigation on the accident to the NAL Saras PT2 aircraft VT-XRM at Sheshagirihalli near Bidadi’ Karnataka on 6^th March 2009” report dated  6 December 2009. The full report of 75 pages is available at dgca.nic.in/accident reports. I understand that there are no other reports available to the public.

I am uneasy with the findings of the report and its recommendations because of the following:

i)                   The report does not explain the type of wreckage or the wreckage “foot print”.

In addition the report does not seem to have investigated the following:

ii)                 The configuration of the aircraft ( rear engine T tail turboprop)

iii)                “deep stall”

iv)                flat spin”

v)                  The possibility of the CG position of the aircraft having travelled beyond the rear CG limit.

Without having investigated and ruled out the above possibilities the report may have drawn erroneous conclusions and corrections.

If my prognosis is correct the aircraft may crash again if tested under the suggested corrected test parameters.

Unfortunately the report does not contain the data needed for me to check my own apprehensions. Since the Saras test flying has resumed I am taking the risk of publishing my views for a wider discussion.

Introductory Glossary:

The rear engine twin turboprop configuration:

The engine, nacelle, propeller and mountings of this class of aircraft- twin engine light commuter- is around 15% of the mass of the aircraft. If such a mass is located at the rear, as in the Saras, the CG moves backwards by a considerable account.

To “balance” the aircraft for stability the wing has to be moved back. This reduces the tail moment arm significantly and it is normal for this configuration to have either reduced yaw and pitch stability or have larger keel (fin) and stabilizer areas so as to maintain “tail volumes”. The lateral stability is worsened by the fact that there is a larger “Side area” ahead of the CG further reducing the yaw and pitch stabilities.

The location of the undercarriage is more sensitive in a rear engine layout. It can affect the stabilizer sizing (by about 30% in one case). However there is no evidence that this aspect played any immediate role in this accident and is therefore not being further discussed here.

The spin

An aircraft if yawed at near stall speeds it will inevitably flick into a spin. It happens as follows:

The aircraft is flying straight and level at near the stall speed. If yawed, say as in this case, to the left the left (port) wing’s airspeed falls and it generates less lift. This causes the wing to dip which increases the angle of attack of this left (port) wing. It is now well and truly stalled. The wing is past the inflexion of the Lift /Drag curve and so it generates less lift and more drag. Exactly the opposite happens to the starboard (right) wing. The entire wing is now subjected to a rolling couple (because of the different lifts) and a yawing couple because of the different drags. The aircraft becomes uncontrollable because of the low airspeed which makes the control surfaces forces much less than in normal flight. Recovery happens by diving the aircraft- to get the entire wing “flying” again- and correcting the yaw. A spin is a frightening situation to be in and requires preparation and mental “psyching”. Due to the lower moment arms a rear engine aircraft takes longer to recover from a spin.

Deep stall

At stalling speed an aircraft is at around an AOA of 15-18 degrees. This may increase as the aircraft begins to sink due to loss of lift. In a rear engine aircraft with a T tail the disturbed airflow past the engine nacelles “hits” the stabilizer and reduces its effectiveness partially or completely. The pilot is unable to pitch the aircraft nose down to recover from the stall. The BAC 111 prototype accident in 1963 brought this phenomenon to public attention. The Deep stall is peculiar to the rear engine layout.

The alternative construct of the accident

The starting point of this hypothesis is the examination of the impact point and the description of the wreckage:

1.      The ground is hard and stony. The wreckage should have slid along and formed the usual “wreckage trail”. This has not happened.

2.      An aircraft has fallen out of control from a considerable height (about 7000’). The scatter would have been like that of a bomb explosion with an elliptic/asymmetric scatter of debris.

These are not reported.

3.      It is reported that all the components are within a 20 mts. radius of the wreckage i.e. hardly greater than the wing span of the aircraft. This was caused by the fuel tanks exploding and not due to kinetic energy As reported rightly, this indicates there was little or no forward speed.

This type of the wreckage “footprint” happens when the aircraft has crashed whilst in a flat spin or deep stall or if the aircraft CG has travelled beyond the rear limit.

How did the spin happen?

The aircraft as per report was at around 132 knots at the beginning of left engine shutdown. The speed decayed further due to additional rudder trim drag and due to lower power from one engine. At the beginning of the engine restart point approx -1.32 minutes time before the impact the speed was around 120 kts. At that altitude and at the probable then weight of the aircraft the stalling speed was around 115 kts.

When the propeller was accidentally unfeathered, the asymmetric drag pulled the speed down close or below the stalling speed and also yawed the aircraft. The spin then became inevitable.

The spin then developed into a flat spin which was probably mistaken by the crew as a (momentary) recovery. If the aircraft was in a normal spin it would have impacted “nose down”. This is not indicated by the wreckage. Recovery from a flat spin is impossible. The crew did not have a chance.

Deep stall

According to the report the aircraft was losing height at 10,000 fpm. This works out to around 95 kts. The IAS of the aircraft had (to be noted) further decayed to around 110 kts. The relative wind was in the order of 60 degrees. Even if nose down pitch angle is considered (the report is not clear on this) the aircraft was definitely stalled and the stabilizer was in the disturbed air of the nacelles and therefore ineffective.

Rearward shift of the CG.

The pilots repeatedly complain about the excessive drag. The Saras aircraft is not per se a “draggy” layout (note 1). From the TO weight given the aircraft was not ballasted for the passengers. In a rear engine layout this results in the CG moving back. It is quite likely that the CG had started at a very rear ward position and as the fuel was consumed it moved further back beyond the permissible limit. This made the aircraft uncontrollable. An aircraft out of trim- like a car being driven with the parking brake “on” will feel “draggy “and be ready to depart control. This condition can demonstrated easily and simply.

Other factors.

1.      The pilots had over 2000 hours experience but less than 300 hours on turboprops. They were likely from the fighter stream. Test piloting is a precision engineering job.

2.      The aircraft had a three lever engine/prop control unlike the usual two lever system. This particular test required instant and instinctive reactions as well as utter familiarity with turboshaft engines- preferably not even turboprop. The Crew( apart from the Flight Engineer) were not suited for this task ( Note 2)

Conclusions:

The aircraft crashed due to problems associated with the layout, deep stall and a too rear position of the CG or a combination of all three.

If the test is repeated at the same speed but even with 15000’ AMSL altitude recovery from uncontrollability is not certain. The low speed close to the stall was definitely a factor.

The verdict of pilot error is to be reviewed and downgraded.

Recommendations

1.      The spin recovery and the deep stall characteristics of the Saras needs further study and revalidation

2.      It is presumed that prototypes were under 24X7 CCTV surveillance and monitoring. Prototype security on the ground is mandatory and standard assumption in accident investigations. IF not unauthorized tampering and or sabotage cannot be definitely ruled out.

3.      The behavior of the two pilots prior to the start of the tests, the wrong reporting of heights etc and the decision to press on despite misgivings of all three needs further investigations. It is presumed that the post mortem included checks regarding absence of harmful medication.

4.      The verdict of pilot error needs review. They were not briefed about the possibility of the aircraft spinning. Any flight test plan should be subjected to “Failure Mode Analysis” so that pitfalls are not overlooked. The failure to anticipate the possibility of proximity to stall speeds at 10,000 feet is worrisome.

5.      The report should include, in tabular form, the aircraft attitude parameters, the altitude and the control deflections, particularly for event after -1.36min.

6.      It should be mandatory that any crew member can order abort of a sortie and return to base for debrief and check. It is cheaper and faster than accident investigations and building another prototype.

7.      A sortie cannot depart until all instruments and sensors are tested for uniformity of readings.

If the root cause of the accident is not correctly identified then these men have died in vain.

Note 1.

There is no reason to surmise that the drag reported by the pilots is inherent in the design. Some views have been heard that the fuselage diameter is too large for the aircraft and that the wing aspect ratio is too low.

The fuselage L/D ratio is around 7.5:1 and is suitable for this class of aircraft and speeds.

The aspect ratio of about 8:1 is standard practice for Russian designs who use this even in their long range bombers. Interestingly the Americans use around 11:1 and suffer from structural fatigue problems. The C5A Galaxy had its entire wing replaced in most airframes and the B 52 has had so many “Band Aid” patches it is claimed that the airframe carries its own weight in these reinforcement patches. Since induced drag is a reciprocal of AR the Russian approach is technically more correct. This is a personal opinion.

Note 2

Flying safety requires quick reactions to the point of being instinctive. Wg. Cdr. Guy P. Gibson V.C.  the legendary bomber pilot of the RAF died in a crash of the DH Mosquito aircraft to which he had recently converted and was out of practice. It is surmised that when one of the fuel tanks ran dry and the engines cut delay in switching tanks because of his being out of practice may have caused the fatal crash.