Aircraft Accidents and Lessons Unlearned LV: Loganair flight 670A

On February 27, 2001, a Loganair Ltd. Shorts Brothers SD3-60 Variant 100, registration number G-BNMT, flight 670A, crashed near Birnie Rocks, Scotland, shortly after taking off from Edinburgh Airport. The aircraft, which had two Pratt and Whitney PT6A-67R turboprop engines, impacted the waters in the Firth of Forth at a 6.8˚ attitude in six meters of water.

The investigation found, “… following a selection by the crew of the anti-icing systems on the aircraft, specifically the selection of the intake anti-ice vanes, the subsequent movement of the vanes precipitated the near simultaneous engine flameouts. Interaction between the moving vanes and the residual ice, snow or slush contamination in both intake systems is considered to be the most likely cause of the engine failures.” The investigators felt that “A significant amount of snow almost certainly entered into the engine air intakes as a result of the aircraft being parked heading directly into strong surface winds during conditions of light to moderate snowfall overnight.”

Although there were six Causal Factors identified, these factors never verified Root Cause; indeed, the investigators did not even come close to determining Probable Cause using their own words. Though there stands the chance that the causal factors, as identified, led to the accident, the list of causal factors are plagued with terms, like “A significant amount of snow ALMOST CERTAINLY [capitalization added] entered …”; The flow characteristics of the engine intake system MOST PROBABLY allowed …” and “At some stage, PROBABLY AFTER engine ground running … slush ALMOST CERTAINLY migrated …” This type of analysis did/does not generate confidence in the accident report’s quality. In keeping with the effort made during this investigation, the investigators MOST CERTAINLY missed some important information and PROBABLY did not put effort into investigating the accident beyond the only questionable theory pursued, demonstrative of a languorous attempt at the fact-finding.

What was missing from the report was any effort to determine cause. Considering the aircraft and both engines were recovered in a reasonably unmolested condition – meaning no post-accident damage was encountered – the post-accident investigation and inspections of the engines alleged to have been subjected to, “… the near simultaneous engine flameouts,” any analysis quality was non-existent. The report notes found under 1.12.2.1 Powerplant did not describe any internal engine damage that led to the simultaneous flameouts; the report was absent of any information about the intake stages’ condition, even though both engines were on-site. Instead, the report delivered useless information about the crash site terrain and water temperature (at the crash site), all of which had nothing to do with the accident.

There are two reasons to look again at this flight. First, associating icing of any kind and maintenance. The second reason is, though not seen many times in more technologically advanced airplanes, aircraft like the Shorts SD3-60 are out there and we ignore them at our own peril. What the report failed to focus on, though aviation accident knowledge demanded it, were alternative possibilities of accident cause.

The British investigatory authorities, the Air Accidents Investigation Branch (AAIB) assigned four people to the investigation: the Investigator-in-Charge (IIC), an Operations investigator, an Engineering investigator and a Flight Recorders expert. It is unclear what expertise the Engineering Investigator had in both Maintenance and Engineering and, if the Engineering Investigator was an Engineer, what was his/her specialty: Systems, Powerplants, Structures or, even on the slight chance, Maintenance? This is a valid question because when reviewing accident reports, such as the Lion Air and Ethiopian Airlines B737-MAX accidents, Mechanics are called Engineers in certain countries. Either way, the Engineering Investigator failed to show any expertise in either area.

Before other possible causes are examined, it is important to understand how the aircraft lost control during the last minutes of flight, what obstacles the flight crew faced, even with a double flameout, which might have been survivable. Normally for ground operations, such as towing, many aircraft are equipped with an alternate means of hydraulic power that would not make use of the engine-driven hydraulic pumps; whether the ground hydraulic pumps are electric or reversible, they are controllable from inside the cockpit, though, they may require AC power. The accident aircraft lost power in both engines during take-off, a critical phase of flight where things tend to go wrong in spectacular ways.

The first problem was electrical; if the engine driven generators went offline together, power to ground hydraulic pumps was lost. It was unclear if the Shorts SD3-60 had a ram air turbine or air driven generator to power the pumps. Although the crew transmitted a MAYDAY, the radio could have been powered by the battery.

The Shorts SD3-60 had climbed out of Edinburgh; per the report, flaps and landing gear were retracted. Even so, these two items, with the activation of primary flight controls, are regular taps to the hydraulic pressure and quantity reserves. When the engines quit, hydraulic pumps were lost, hydraulic pressure would drop quickly, even depleting any hydraulic accumulators that may have been there to assist.

The lower an aircraft is to ground when trouble starts, the less likely successful emergency maneuvers can take place, especially with no hydraulic reserves. Starting descent from 2200 feet, the flight crew soon found room for wingtips and empennages decreased. An aircraft’s exaggerated glide ratio never achieves the advertised numbers based on conditions, especially during climb. Descent rate was sacrificed to gravity, drag, weight and balance. The loss of hydraulic power would result in a reduction in primary control authority; any tab-driven flight controls would also be difficult to ‘fly’ due to lower airspeeds at climb as opposed to cruise. This explains how quickly the flight crew could lose control of the aircraft.

The reason given above for giving this accident flight a second, more thorough, look was “associating icing of any kind and maintenance.” The Shorts SD3-60 was equipped with two Pratt and Whitney PT6A-67R turboprop engines; these engines operated on Jet-A – Jet Fuel – which has a lower density than water; anything will ‘float’ on water as long as its density is lower than that of water. The density of jet fuel is around 0.81 kg/L while that of pure water is 1.0 kg/L. Thus, jet fuel rises in water – it floats.

The Shorts SD3-60 employs a high wing with fuel tanks. Per the report, 1.6.3.4 Fuel System, “Each tank group [of which there are four] gravity feeds, via non return valves, a filter and a negative ‘g’ valve, into its own small, dedicated collector tank. Each of these two collectors incorporates its own boost pump …” To be clear, the jet fuel is gravity fed to the boost pump – not from the boost pump – before getting to the engine. The report stated that the aircraft was fueled the previous night, fifteen hours before actual flight. The aircraft sat longer than four hours, which is the average time it takes for jet fuel suspended in water, to separate into water on the bottom of the tank and jet fuel on the top; any agitation caused by the fuel pump was after the gravity feed piping.

The temperature on the field at midnight was +1˚ Centigrade and continued to drop … overnight. When landing, the aircraft passed through an ice layer, supercooling the wing … at night. The accident aircraft then sat with moderate snowfall covering the wing surface … overnight.

The wing was below freezing; any chance of the wing warming above freezing was frustrated by snow covering the wing, which blocked sunlight. A more probable cause of the accident was that the aircraft crashed because the engines were simultaneously starved of fuel due to fuel tank icing that would have fouled the gravity feed tubes, the filters or both.

Any opportunity to check the wings for water were disregarded by investigators, even though both wings were intact. The investigators also ignored the fuel station that provided the accident aircraft with fuel to check fuel integrity. Was the fuel farm supply contaminated with water? Could other aircraft have been affected by contaminated fuel? These questions will never be answered. More importantly was the wasted opportunity to generate useful and effective recommendations, such as introducing a simple fuel draining measure, called ‘Sumping’, into the Maintenance program for Loganair and other operators. A sumping program that required daily or weekly draining of fuel samples from the wings could have discovered high levels of water, if found, in the fuel tanks. That was where a Maintenance specialist would have added to the accident investigation to make the report a quality report.

As earlier stated, the second reason is, though not seen many times in more technologically advanced airplanes, aircraft like the Shorts SD3-60 are out there and we ignore them at our own peril. The solution to the fuel icing problem was one that plagues most technologically advanced aircraft as well as the older analog models and helicopters. The missed opportunity demonstrated in this report showed that it is still important to get the accident investigation right; to find out root cause and determine solutions – no matter how inconvenient it is.