Description

On 24 October 2011, the crew of a Boeing 737-800 (EI-DPA) being operated by Ryanair on a scheduled passenger flight from Faro, Portugal to Beauvais France experienced difficulty in maintaining directional control after take off and during the climb in day Visual Meteorological Conditions (VMC). It was decided to return to land where evidence that the rudder assembly was damaged was found and the aircraft was withdrawn from service.

Investigation

When notified of the occurrence by the AAIU, the GPIAA decided to delegate the Investigation to Ireland as State of Registry and State of the Operator. The aircraft DFDR was downloaded by the Operator and a copy of the data was provided to the Investigation. The CB of the 2 hour Cockpit Voice Recorder (CVR) was tripped by the aircraft commander approximately 55 minutes after landing and was successfully downloaded. Good data was available from approximately 40 minutes prior to take-off.

It was established that the aircraft had been parked overnight on the ramp at Faro with a number of other Ryanair aircraft of the same type. At around 0400 hrs, a particularly violent storm affected the airport during which it was subsequently established that wind speeds up to 85 knots were recorded at one of the airport automatic weather stations. The incident aircraft was found afterwards to have jumped its chocks and pivoted nose left such that the nose landing gear had moved approximately 5 metres from its original position.

The engineer responsible for checking the aircraft signed off the required Check about an hour later. No Technical Log entry was made about the exposure to extreme winds or to the fact that the aircraft had jumped its chocks and moved.

The first flight of the day was not scheduled until the afternoon. The aircraft commander told the Investigation that he had been aware of the overnight storm and that during his pre-departure external inspection, he had looked carefully for any signs of damage. He advised that he had found "sand and grit in the bottom of (both) engines and some fibreglass insulation material that had blown into the back of the No. 1 engine and was located to the rear of the N1 fan". He had requested an engineer to attend the aircraft and they had agreed that he would wait for the passengers to board before removing the material. It was subsequently confirmed that this had been done.

It was established that the First Officer had been designated as PF and nothing abnormal was detected during engine start and taxi out. The initial part of the take-off roll was normal but as the aircraft rotated from runway 28, the commander saw that the aircraft was drifting slightly right of the centreline (and) at the same time we had almost full left aileron deflection. The First Officer had said that the aircraft "didn’t feel right". Given the prevailing light wind (310° /15 knots had been given by TWR with the take off clearance), the commander reported that although his first thought was they had suffered an engine failure, all engine indications were normal. During the climb out, the First Officer demonstrated how the aircraft wanted to roll to the right. Some left rudder trim was set, but it was not enough to be able to engage the AP. After take off actions were completed as the initial straight-ahead part of the SID was flown. At 6nm, the required right turn onto north was made. Once the aircraft had completed the right turn, the addition of about eight units of left rudder trim made it possible to centralise the control column and engage the AP.

Although ATC had cleared the aircraft to FL240, the crew decided to level off at FL 120 and remain in the vicinity of Faro and ATC were so advised. It was apparent that there was no applicable Quick Reference Handbook (QRH) Checklist. In the knowledge that it was at least possible that the overnight storm might have caused undetected damage and aware that aircraft control in the event of any right engine failure would require more rudder trim for directional control than remained after correcting for whatever the existing problem was, it was decided to return to Faro. This was accomplished uneventfully after a 28 minute flight and an appropriate defect entry made in the aircraft Technical Log after engine shutdown.

In order to examine the aircraft rudder, engineering staff removed panels from the vertical stabiliser to gain access to the Main and Standby rudder PCUs and it was found that the support brackets for both PCUs had sustained damage with visible evidence of cracking, fracturing and displacement. Arrangements were therefore made for a more comprehensive inspection of this area to be independently supervised.

This inspection found no additional damage to the rudder system hardware and no obvious external damage to either of the PCUs. They also "inspected the elevator and aileron systems and in both cases reported that they found no collateral damage or anomalies" and found no defect in the control column and rudder pedal forward stops, the stabiliser jack screw, the rudder pulleys and elevator pulleys or any of their associated control cables.

The principal effect of the damage that was found was that "the overall movement of the rudder remained approximately symmetrical around the neutral point, but now displaced by about 8°, trailing edge right". This effect would have created the handling abnormality experienced by the pilots.

The Inspection team also examined the nose landing gear and found that its actuators were leaking hydraulic fluid. The Operator advised that there had been no leaks prior to the extreme wind event.

It was reported that several other aircraft owned by the Operator and parked on the apron overnight had jumped their chocks and moved during the storm but no associated damage was subsequently reported on any of those aircraft.

It was noted that, as is usual in the case of powered flying controls, the 737 rudder system does not incorporate a gust-lock. Instead, a gust damping function is provided by the bypass valves for the main PCU which have flats machined into the valve spools so that a metered flow of hydraulic fluid can ensure gust damping is available when the aircraft is parked and the engine driven hydraulic pumps are not working.

In respect of the rudder, it was concluded that "the control surface loads (had probably) significantly exceeded the overload protection afforded by the maximum design gust damping provided by the Main PCU" and that this had been the result of "extreme gusts impacting the rudder from the left". Although both rudder PCUs were found to be undamaged by exposure to the loads on the system during the severe weather episode, "there was physical displacement of one of the Main PCU support brackets, where the attachment bolts had all sheared". In addition, permanent deformation and gaps were found where the other support bracket was attached to the aircraft structure and these were attributed to "overload and over stress" as were "fractures of the flanges attached to the Standby PCU support bracket and the deformation of the two brackets".

It was considered that the physical movement of one of the PCU mounting brackets and the damage sustained by the others are likely to have resulted in the PCUs themselves being displaced with the consequence that the rudder “neutral” position was itself shifted, trailing edge to the right as shown in the photograph below.

It was noted that at the time of the occurrence, the AMM did not contain any requirement for a specific check to be carried out following an extreme high wind event such as the one which occurred. It was also noted that type certification under FAR 25.415 Amendment 25-91 dated 20 August 1990 had required that “the control system must be designed…for control surface loads due to ground gusts”. The Investigation was informed by the FAA that the effect of Grandfather Rights in respect of the 737NG design, was that the regulatory maximum design wind speed was 52 knots, although Boeing confirmed that, in fact, the type actually met the slightly higher requirement of 65 knots set in 1997.

When asked by the Investigation whether consideration had been given to an increase in the FAR 25.415 wind speed above 65 knots, the FAA responded that the relevant ARAC had recommended the application of additional dynamic factors for those portions of the flight control system and surfaces that could be affected by dynamic effects rather than a blanket increase in the wind speed requirement because it was considered that the most significant contributor to damage due to ground gust was the dynamic load effect.

The nature of the extreme weather event which had occurred was reviewed with the aid of a copy of a report in Portuguese provided by the GPIAA which had been prepared by the Portuguese Government Institute of Meteorology. Based on a translation of this document into English, it was found that they had concluded that:

"In the early morning of October 24, 2011, Faro Airport and the adjoining built-up area, located east and northeast of same, were struck by a phenomenon called a microburst, in direct association with a supercell whose path was established over the mentioned area, advancing from southwest to northeast. The microburst must have affected the levels close to the ground more significantly, during the period between 04.01 hrs and 04.03 hrs UTC, and its path of destruction must have covered an approximate distance of 3 km, between the western area of Faro Airport to the city, with a variable width not exceeding 500 m…. It is evident that the supercell….core passed slightly to the northeast of the EMA (automatic weather station located on the apron) at around 1 km from it."

The annotated photograph below taken from that report, shows the location of the four automatic weather stations on or in the vicinity of the apron. The blue arrowed lines represent the maximum wind vectors at the four stations annotated with UTC times. The highest maximum instantaneous wind speed recorded during the microburst was 85 knots from 219°which occurred just after 0400 Z. The vector marked 45 m/s (87.5 knots) in the upper left of the photograph is provided for scaling purposes. The location of Stand 18 where the aircraft was parked nose in to the terminal building overnight is also shown.

It was apparent that the layout of the airport had meant that the apron was exposed to the weather which occurred and it was concluded that "the severe wind gusts probably impacted on the left hand side and rear of the aircraft and, as a consequence, the rudder control surface moved rapidly, trailing edge to the right." The recorded wind speeds meant that it was extremely likely that the parked aircraft "and specifically the rudder surface and control system was subjected to excessive loads due to ground wind speeds considerably greater than 65 knots".

It was noted that the extreme weather had not been forecast. The wind speed in the twelve hour TAF for the airport valid from 1800Z the previous day included gusts to a maximum of 50 knots with a 30% probability and the TAF valid from 0000z included gusts to a maximum of 45 knots, again with a probability of 30%.

It was determined that the Probable Cause of the incident was “undetected structural damage caused to the rudder system of EI-DPA by violent wind gusts associated with a microburst, while the aircraft was parked”.

It was further determined that there were two Contributory Causes:

• The speed of the wind gusts (to which the aircraft was exposed whilst parked out of service) at Faro considerably exceeded the figure of 65 knots specified in FAR 25.415 Amendment 25-91.

• The absence of a requirement in the AMM to comprehensively check the flight control systems of an aircraft which has been subjected to an extreme meteorological event.

Safety Action taken by Boeing subsequent to the event was noted to have consisted of the issue in October 2012 of an amendment to the Boeing 737NG AMM establishing a requirement for a specific inspection if a parked aircraft has been subjected to an “Extreme High Wind Event”. The amended text was noted to state that “examples of such events are tornadoes and microbursts” and to list as examples of evidence of such an event “situations where an aircraft has moved over its wheel chocks or where there are tyre skid marks on the ground where an aircraft was parked”. The new Inspection includes a requirement for the examination of:

• the external surfaces of flight control surfaces for damage
• actuator attachment points and their support structure
• the hinge attachment structure of flight controls.

One Safety Recommendation was made as a result of the Investigation as follows:

• that the Federal Aviation Administration should review FAR 25.415, with particular emphasis on those aircraft designs where the multiplying factors set out in the associated NPRM issued in May 2013 may not be applicable. [IRLD2014025]

The Final Report was published on 14 August 2014.

Further Reading

• Microburst
• Weather Forecast
• Flight Control Protection from Damage by Strong Winds
• Flight Crew Pre Flight External Check
• Certification of Aircraft, Design and Production
• Hydraulic Problems: Guidance for Flight Crews