AMC 25-24 Sustained Engine Imbalance

ED Decision 2009/017/R

1.       PURPOSE

This AMC sets forth an acceptable means, but not the only means, of demonstrating compliance with the provisions of CS-25 related to the aircraft design for sustained engine rotor imbalance conditions.

2.       RELATED CS PARAGRAPHS

a.       CS-25:

         CS 25.302 “Interaction of systems and structures”

         CS 25.571 “Damage tolerance and fatigue evaluation of structure”

         CS 25.629 “Aeroelastic stability requirements”

         CS 25.901 “Installation”

         CS 25.903 “Engines”

b.       CS-E:

CS-E 520 “Strength”

         CS-E 525 “Continued Rotation”

         CS-E 810 “Compressor and Turbine Blade Failure”

         CS-E 850 “Compressor, Fan and Turbine Shafts”

3.       DEFINITIONS. Some new terms have been defined for the imbalance condition in order to present criteria in a precise and consistent manner. In addition, some terms are employed from other fields and may not be in general use as defined below. The following definitions apply in this AMC:

          a.       Airborne Vibration Monitor (AVM). A device used for monitoring the operational engine vibration levels that are unrelated to the failure conditions considered by this AMC. 

          b.       Design Service Goal (DSG). The design service goal is a period of time (in flight cycles/hours) established by the applicant at the time of design and/or certification and used in showing compliance with CS 25.571.

          c.       Diversion Flight. The segment of the flight between the point where deviation from the planned route is initiated in order to land at an en route alternate airport and the point of such landing.

          d.       Ground Vibration Test (GVT). Ground resonance tests of the aeroplane normally conducted in compliance with CS 25.629.

          e.       Imbalance Design Fraction (IDF). The ratio of the design imbalance to the imbalance (including all collateral damage) resulting from release of  a single turbine, compressor, or fan blade at the maximum rotational speed to be approved, in accordance with CS-E 810.

          f.       Low Pressure (LP) Rotor. The rotating system, which includes the low pressure turbine and compressor components and a connecting shaft. 

          g.       Well Phase. The flight hours accumulated on an aeroplane or component before the failure event.

4.       BACKGROUND

          a.       Requirements. CS 25.901(c) requires the powerplant installation to comply with CS 25.1309. In addition, CS 25.903(c) requires means of stopping the rotation of an engine where continued rotation could jeopardise the safety of the aeroplane, and CS 25.903(d) requires that design precautions be taken to minimise the hazards to the aeroplane in the event of an engine rotor failure. CS-E 520(c)(2) requires that data shall be established and provided for the purpose of enabling each aircraft constructor to ascertain the forces that could be imposed on the aircraft structure and systems as a consequence of out-of-balance running and during any continued rotation with rotor unbalance after shutdown of the engine following the occurrence of blade failure, as demonstrated in compliance with CS-E 810, or a shaft, bearing or bearing support, if this results in higher loads.

          b.       Blade Failure. The failure of a fan blade and the subsequent damage to other rotating parts of the fan and engine may induce significant structural loads and vibration throughout the airframe that may damage the nacelles, equipment necessary for continued safe flight and landing, engine mounts, and airframe primary structure. Also, the effect of flight deck vibration on displays and equipment is of significance to the crew’s ability to make critical decisions regarding the shut down of the damaged engine and their ability to carry out other operations during the remainder of the flight. The vibratory loads resulting from the failure of a fan blade have traditionally been regarded as insignificant relative to other portions of the design load spectrum for the aeroplane. However, the progression to larger fan diameters and fewer blades with larger chords has changed the significance of engine structural failures that result in an imbalanced rotating assembly. This condition is further exacerbated by the fact that fans will continue to windmill in the imbalance condition following engine shut down.

          c.       Bearing/Bearing Support Failure. Service experience has shown that failures of bearings/bearing supports have also resulted in sustained high vibratory loads.

          d.       Imbalance Conditions. There are two sustained imbalance conditions that may affect safe flight: the windmilling condition and a separate high power condition.

(1)      Windmilling Condition. The windmilling condition results after the engine is shut down but continues to rotate under aerodynamic forces. The windmilling imbalance condition results from bearing/bearing support failure or loss of a fan blade along with collateral damage. This condition may last until the aeroplane completes its diversion flight, which could be several hours.

(2)     High Power Condition. The high power imbalance condition occurs immediately after blade failure but before the engine is shut down or otherwise spools down. This condition addresses losing less than a full fan blade which may not be sufficient to cause the engine to spool down on its own. This condition may last from several seconds to a few minutes.  In some cases it has hampered the crew's ability to read instruments that may have aided in determining which engine was damaged.

          e.       The information provided in this AMC is derived from the recommendations in the report “Engine Windmilling Imbalance Loads - Final Report,” dated July 1, 1997, which is appended to this NPA for information.

          f.       The criteria presented in this AMC are based on a statistical analysis of 25 years of service history of high by-pass ratio engines with fan diameters of 1.52 metres (60 inches) or greater. Although the study was limited to these larger engines, the criteria and methodology are also acceptable for use on smaller engines.

5.       EVALUATION OF THE WINDMILLING IMBALANCE CONDITIONS

          a.       Objective. It should be shown by a combination of tests and analyses that after:

i)        partial or complete loss of an engine fan blade, or

ii)       after bearing/bearing support failure, or

iii)      any other failure condition that could result in higher induced vibrations

including collateral damage, the aeroplane is capable of continued safe flight and landing.

          b.       Evaluation. The evaluation should show that during continued operation at windmilling engine rotational speeds, the induced vibrations will not cause damage that would jeopardise continued safe flight and landing. The degree of flight deck vibration[10] should not prevent the flight crew from operating the aeroplane in a safe manner. This includes the ability to read and accomplish checklist procedures.

This evaluation should consider:

(1)     The damage to airframe primary structure including, but not limited to, engine mounts and flight control surfaces,

(2)     The damage to nacelle components, and

(3)     The effects on equipment necessary for continued safe flight and landing (including connectors) mounted on the engine or airframe.

          c.       Blade Loss Imbalance Conditions

(1)      Windmilling Blade Loss Conditions.  The duration of the windmilling event should cover the expected diversion time of the aeroplane. An evaluation of service experience indicates that the probability of the combination of a 1.0 IDF and a 60 minute diversion is on the order of 10-7 to 10 -8 while the probability of the combination of a 1.0 IDF and a 180 minute diversion is 10-9 or less. Therefore, with an IDF of 1.0, it would not be necessary to consider diversion times greater than 180 minutes. In addition, the 180 minute diversion should be evaluated using nominal and realistic flight conditions and parameters. The following two separate conditions with an IDF of 1.0 are prescribed for application of the subsequent criteria which are developed consistent with the probability of occurrence:

(a)     A 60 minute diversion flight.

(b)     If the maximum diversion time established for the aeroplane exceeds 60 minutes, a diversion flight of a duration equal to the maximum diversion time, but not exceeding 180 minutes.

(2)      Aeroplane Flight Loads and Phases

(a)      Loads on the aeroplane components should be determined by dynamic analysis.  At the start of the windmill event, the aeroplane is assumed to be in level flight with a typical payload and realistic fuel loading. The speeds, altitudes, and flap configurations considered may be established according to the Aeroplane Flight Manual (AFM) procedures. The analysis should take into account unsteady aerodynamic characteristics and all significant structural degrees of freedom including rigid body modes. The vibration loads should be determined for the significant phases of the diversion profiles described in paragraphs 5c(1)(a) and (b) above. 

(b)      The significant phases are:

1        The initial phase during which the pilot establishes a cruise condition;

2        The cruise phase;

3        The descent phase; and

4        The approach to landing phase.

(c)      The flight phases may be further divided to account for variation in aerodynamic and other parameters. The calculated loads parameters should include the accelerations needed to define the vibration environment for the systems and flight deck evaluations. A range of windmilling frequencies to account for variation in engine damage and ambient temperature should be considered.

(3)      Strength Criteria

(a)     The primary airframe structure should be designed to withstand the flight and windmilling vibration load combinations defined in paragraphs 1, 2, and 3 below.

1        The peak vibration loads for the flight phases in paragraphs 5c(2)(b)1 and 3 above, combined with appropriate 1g flight loads. These loads should be considered limit loads, and a factor of safety of 1.375 should be applied to obtain ultimate load.

2        The peak vibration loads for the approach to landing phase in paragraph 5c(2)(b)4 above, combined with appropriate loads resulting from a positive symmetrical balanced manoeuvring load factor of 1.15g. These loads should be considered as limit loads, and a factor of safety of 1.375 should be applied to obtain ultimate load.

3        The vibration loads for the cruise phase in paragraph 5c(2)(b)2 above, combined with appropriate 1g flight loads and 70 percent of the flight manoeuvre loads up to the maximum likely operational speed of the aeroplane. These loads are considered to be ultimate loads.

4        The vibration loads for the cruise phase in paragraph 5c(2)(b)2 above, combined with appropriate 1g flight loads and 40 percent of the limit gust velocity of CS 25.341 as specified at VC (design cruising speed) up to the maximum likely operational speed of the aeroplane. These loads are considered to be ultimate loads.

(b)      In selecting material strength properties for the static strength analyses, the requirements of CS 25.613 apply.

(4)      Assessment of Structural Endurance

(a)      Criteria for fatigue and damage tolerance evaluations of primary structure are summarised in Table 1 below. Both of the conditions described in paragraphs 5c(1)(a) and (b) above should be evaluated. Different levels of structural endurance capability are provided for these conditions. The criteria for the condition in paragraph 5c(1)(b) are set to ensure at least a 50 percent probability of preventing a structural component failure. The criteria for the condition in paragraph 5c(1)(a) are set to ensure at least a 95 percent probability of preventing a structural component failure. These criteria are consistent with the probability of occurrences for these events discussed in paragraph 5(c)(1) above.

(b)      For multiple load path and crack arrest “fail-safe” structure, either a fatigue analysis per paragraph 1 below, or damage tolerance analysis per paragraph 2 below, may be performed to demonstrate structural endurance capability. For all other structure, the structural endurance capability should be demonstrated using only the damage tolerance approach of paragraph 2 below. The definitions of multiple load path and crack arrest "fail-safe" structure are the same as defined for use in showing compliance with CS 25.571, "Damage tolerance and fatigue evaluation of structure."

1        Fatigue Analysis. Where a fatigue analysis is used for substantiation of multiple load path “fail-safe” structure, the total fatigue damage accrued during the well phase and the windmilling phase should be considered. The analysis should be conducted considering the following:

(aa)    For the well phase, the fatigue damage should be calculated using an approved load spectrum (such as used in satisfying the requirements of CS 25.571) for the durations specified in Table 1. Average material properties may be used.

(bb)    For the windmilling phase, fatigue damage should be calculated for the diversion profiles using a diversion profile consistent with the AFM recommended operations, accounting for transient exposure to peak vibrations, as well as the more sustained exposures to vibrations. Average material properties may be used.

(cc)     For each component, the accumulated fatigue damage specified in Table 1 should be shown to be less than or equal to the fatigue damage to failure of the component.

2        Damage Tolerance Analysis. Where a damage tolerance approach is used to establish the structural endurance, the aeroplane should be shown to have adequate residual strength during the specified diversion time. The extent of damage for residual strength should be established, considering growth from an initial flaw assumed present since the aeroplane was manufactured. Total flaw growth will be that occurring during the well phase, followed by growth during the windmilling phase. The analysis should be conducted considering the following:

(aa)    The size of the initial flaw should be equivalent to a manufacturing quality flaw associated with a 95 percent probability of existence with 95 percent confidence (95/95).

(bb)    For the well phase, crack growth should be calculated starting from the initial flaw defined in paragraph 5c(4)(b)2(aa) above, using an approved load spectrum (such as used in satisfying the requirements of CS 25.571) for the duration specified in Table 1.  Average material properties may be used.

(cc)     For the windmilling phase, crack growth should be calculated for the diversion profile starting from the crack length calculated in paragraph 5c(4)(b)2(bb) above. The diversion profile should be consistent with the AFM recommended operation accounting for transient exposure to peak vibrations as well as the more sustained exposures to vibrations.  Average material properties may be used.

(dd)    The residual strength for the structure with damage equal to the crack length calculated in paragraph 5c(4)(b)2(cc) above should be shown capable of sustaining the combined loading conditions defined in paragraph 5c(3)(a) above with a factor of safety of 1.0.

TABLE 1 - Fatigue and Damage Tolerance

Notes:

1            The analysis method that may be used is described in paragraph 5 (Evaluation of the Windmilling Imbalance Conditions) of this AMC.

2            Load spectrum to be used for the analysis is the same load spectrum qualified for use in showing compliance with CS 25.571, augmented with windmilling loads as appropriate.

3            Windmilling phase is to be demonstrated following application of the well phase spectrum loads.

4            The initial flaw for damage tolerance analysis of the windmilling phase need not be greater than the flaw size determined as the detectable flaw size plus growth under well phase spectrum loads for one inspection period for mandated inspections.

5            MQF is the manufacturing quality flaw associated with 95/95 probability of existence. (Reference - ‘Verification of Methods For Damage Tolerance Evaluation of Aircraft Structures to FAA Requirements’, Tom Swift FAA, 12th International Committee on Aeronautical Fatigue, 25 May 1983, Figures 42, and 43.)

6            Maximum diversion time for condition 5c(1)(b) is the maximum diversion time established for the aeroplane, but need not exceed 180 minutes. This condition should only be investigated if the diversion time established for the aeroplane exceeds 60 minutes.

7            The allowable cycles to failure may be used in the damage calculations.

(5)      Systems Integrity

(a)      It should be shown that systems required for continued safe flight and landing after a blade-out event will withstand the vibratory environment defined for the windmilling conditions and diversion times described above. For this evaluation, the aeroplane is assumed to be dispatched in its normal configuration and condition. Additional conditions associated with the Master Minimum Equipment List (MMEL) need not be considered in combination with the blade-out event.

(b)      The initial flight environmental conditions are assumed to be night, instrument meteorological conditions (IMC) en route to nearest alternate airport, and approach landing minimum of 300 feet and 3/4 mile or runway visual range (RVR) 4000m or better.

(6)      Flight crew Response. For the windmilling condition described above, the degree of flight deck vibration shall not inhibit the flight crew’s ability to continue to operate the aeroplane in a safe manner during all phases of flight.

d.       Bearing/Bearing Support Failure. To evaluate these conditions, the low pressure (LP) rotor system should be analysed with each bearing removed, one at a time, with the initial imbalance consistent with the airborne vibration monitor (AVM) advisory level. The analysis should include the maximum operating LP rotor speed (assumed bearing failure speed), spool down, and windmilling speed regions. The effect of gravity, inlet steady air load, and significant rotor to stator rubs and gaps should be included. If the analysis or experience indicates that secondary damage such as additional mass loss, secondary bearing overload, permanent shaft deformation, or other structural changes affecting the system dynamics occur during the event, the model should be revised to account for these additional effects. The objective of the analyses is to show that the loads and vibrations produced by the bearing/bearing support failure event are less than those produced by the blade loss event across the same frequency range.

An alternative means of compliance is to conduct an assessment of the design by analogy with previous engines to demonstrate this type of failure is unlikely to occur. Previous engines should be of similar design and have accumulated a significant amount of flight hours with no adverse service experience.

e.       Other failure conditions. If any other engine structural failure conditions applicable to the specific engine design, e.g. failure of a shaft, could result in more severe induced vibrations than the blade loss or bearing/bearing support failure condition, they should be evaluated.

6.       ANALYSIS METHODOLOGY

a.       Objective of the Methodology. The aeroplane response analysis for engine windmilling imbalance is a structural dynamic problem. The objective of the methodology is to develop acceptable analytical tools for conducting dynamic investigations of imbalance events. The goal of the windmilling analyses is to produce loads and accelerations suitable for structural, systems, and flight deck evaluations.

b.       Scope of the Analysis. The analysis of the aeroplane and engine configuration should be sufficiently detailed to determine the windmilling loads and accelerations on the aeroplane. For aeroplane configurations where the windmilling loads and accelerations are shown not to be significant, the extent and depth of the analysis may be reduced accordingly.

c.       Results of the Analysis. The windmilling analyses should provide loads and accelerations for all parts of the primary structure. The evaluation of equipment and human factors may require additional analyses or tests. For example, the analysis may need to produce floor vibration levels, and the human factors evaluation may require a test (or analysis) to subject the seat and the human subject to floor vibration.

7.       MATHEMATICAL MODELLING

a.       Components of the Integrated Dynamic Model. Aeroplane dynamic responses should be calculated with a complete integrated airframe and propulsion analytical model. The model should provide representative connections at the engine-to-pylon interfaces, as well as all interfaces between components (e.g., inlet-to-engine and engine-to-thrust reverser). The model should be to a similar level of detail to that used for certification flutter and dynamic gust analyses, except that it should also be capable of representing asymmetric responses. The model should be representative of the aeroplane to the highest windmilling frequency expected. The model consists of the following components:

(1)      Airframe structural model,

(2)      Propulsion structural model (including the engine model representing the engine type-design),

(3)      Control system model,

(4)      Aerodynamic model, and

(5)      Forcing function and gyroscopic effects

The airframe and engine manufacturers should mutually agree upon the definition of the integrated structural model, based on test and experience.

b.       Airframe Structural Model. An airframe structural model is necessary in order to calculate the response at any point on the airframe due to the rotating imbalance of a windmilling engine. The airframe structural model should include the mass, stiffness, and damping of the complete airframe. A lumped mass and finite element beam representation is considered adequate to model the airframe. This type of modelling represents each airframe component, such as fuselage, empennage, and wings, as distributed lumped masses rigidly connected to weightless beams that incorporate the stiffness properties of the component. A full aeroplane model capable of representing asymmetric responses is necessary for the windmilling imbalance analyses. Appropriate detail should be included to ensure fidelity of the model at windmilling frequencies. A more detailed finite element model of the airframe may also be acceptable. Structural damping used in the windmilling analysis may be based on Ground Vibration Test (GVT) measured damping.

c.       Propulsion Structural Model

(1)      Engine manufacturers construct various types of dynamic models to determine loads and to perform dynamic analyses on the engine rotating components, its static structures and mounts. Dynamic engine models can range from a centreline two-dimensional (2D) model, to a centreline model with appropriate three-dimensional (3D) features such as mount and pylon, up to a full 3D finite element model (3D FEM). Any of these models can be run for either transient or steady state conditions.

(2)      Propulsion structural models typically include the engine and all major components of the propulsion system, such as the nacelle intake, fan cowl doors, thrust reverser, common nozzle assembly, all structural casings, frames, bearing housings, rotors, and a representative pylon. Gyroscopic effects are included. The models provide for representative connections at the engine-to-pylon interfaces as well as all interfaces between components (e.g., inlet-to-engine and engine-to-thrust reverser). The engine that is generating the imbalance forces should be modelled in this level of detail, while the undamaged engines that are operating normally need only to be modelled to represent their sympathetic response to the aeroplane windmilling condition.

(3)      Features modelled specifically for blade loss windmilling analysis typically include fan imbalance, component failure and wear, rubs (blade to casing, and intershaft), and resulting stiffness changes. Manufacturers whose engines fail the rotor support structure by design during the blade loss event should also evaluate the effect of the loss of support on engine structural response during windmilling. 

(4)      Features that should be modelled specifically for bearing/bearing support failure windmilling events include the effects of gravity, inlet steady air loads, rotor to stator structure friction and gaps, and rotor eccentricity. Secondary damage should be accounted for, such as additional mass loss, overload of other bearings, permanent shaft deformation, or other structural changes affecting the system dynamics, occurring during rundown from maximum LP rotor speed and subsequent windmilling.

d.       Control System Model. The automatic flight control system should be included in the analysis unless it can be shown to have an insignificant effect on the aeroplane response due to engine imbalance.

e.       Aerodynamic Model. The aerodynamic forces can have a significant effect on the structural response characteristics of the airframe. While analysis with no aerodynamic forces may be conservative at most frequencies, this is not always the case. Therefore, a validated aerodynamic model should be used. The use of unsteady three-dimensional panel theory methods for incompressible or compressible flow, as appropriate, is recommended for modelling of the windmilling event. Interaction between aerodynamic surfaces and main surface aerodynamic loading due to control surface deflection should be considered where significant. The level of detail of the aerodynamic model should be supported by tests or previous experience with applications to similar configurations. Main and control surface aerodynamic derivatives should be adjusted by weighting factors in the aeroelastic response solutions. The weighting factors for steady flow (k=0) are usually obtained by comparing wind tunnel test results with theoretical data.

f.       Forcing Function and Gyroscopic Forces. Engine gyroscopic forces and imbalance forcing function inputs should be considered. The imbalance forcing function should be calibrated to the results of the test performed under CS-E 810.

8.       VALIDATION.

a.       Range of Validation. The analytical model should be valid to the highest windmilling frequency expected.

b.       Aeroplane Structural Dynamic Model. The measured ground vibration tests (GVT) normally conducted for compliance with CS 25.629 may be used to validate the analytical model throughout the windmilling range. These tests consist of a complete airframe and propulsion configuration subjected to vibratory forces imparted by electro-dynamic shakers. 

(1)      Although the forces applied in the ground vibration test are small compared to the windmilling forces, these tests yield reliable linear dynamic characteristics (structural modes) of the airframe and propulsion system combination. Furthermore, the windmilling forces are far less than would be required to induce non-linear behaviour of the structural material (i.e. yielding).  Therefore, a structural dynamic model that is validated by ground vibration test is considered appropriate for the windmilling analysis.

(2)      The ground vibration test of the aeroplane may not necessarily provide sufficient information to assure that the transfer of the windmilling imbalance loads from the engine is accounted for correctly. The load transfer characteristics of the engine to airframe interface via the pylon should be validated by test and analysis correlation. In particular, the effect of the point of application of the load on the dynamic characteristics of the integrated model should be investigated in the ground vibration test by using multiple shaker locations.

(3)      Structural damping values obtained in the ground vibration tests are considered conservative for application to windmilling dynamic response analysis. Application of higher values of damping consistent with the larger amplitudes associated with windmilling analysis should be justified.

c.       Aerodynamic Model. The dynamic behaviour of the whole aeroplane in air at the structural frequency range associated with windmilling is normally validated by the flight flutter tests performed under CS 25.629.

d.       Engine Model. The engine model covering the engine type-design will normally be validated by the Engine manufacturer under CS-E 520(c)(2) by correlation against blade-off test data obtained in showing compliance with CS-E 810. This is aimed at ensuring that the model accurately predicts initial blade release event loads, any rundown resonant response behaviour, frequencies, potential structural failure sequences, and general engine movements and displacements. In addition, if the Failure of a shaft, bearing or bearing support, results in higher forces being developed, such Failures and their resulting consequences should also be accurately represented.

9.       HIGH POWER IMBALANCE CONDITION.

An imbalance condition equivalent to 50 percent of one blade at cruise rotor speed considered to last for 20 seconds may be assumed unless it is shown that the engine will respond automatically and spool down in a shorter period. It should be shown that attitude, airspeed, and altimeter indications will withstand the vibratory environment of the high power condition and operate accurately in that environment. Adequate cues should be available to determine which engine is damaged. Strength and structural endurance need not be considered for this condition.

[Amdt 25/8]