AC 27-1B, Chg 1 2/12/03
(3) Fatigue evaluation of the structure, measurement of flight loads and stresses, and evaluations of fatigue strength and/or fatigue crack propagation are the subjects of this advisory circular. There is some question whether a completely reliable method for the prediction of time to fatigue crack initiation and fracture exists. Nevertheless, one engineering approach to the subject is to use theiLinear Cumulative Damage Hypothesis." This hypothesis states that every cycle of stress above an "endurance limit produces fatigue damage proportional to the ratio of cycles accumulated at the stress to fatigue "life" at that stress.
(4) Laboratory tests of this hypothesis indicate that it is reasonably valid when the loading spectrum consists of stresses that are, in effect, random. Despite the lack of an adequate theory connecting this hypothesis with more basic properties of materials, it has been successfully used in a number of applications to calculate a safelife retirement time.
(5) In addition, fatigue evaluation generally requires a method of accounting for the effect of steady loads and stresses on fatigue. Where the manufacturer does not provide other substantiating data, a Goodman diagram may be used to account for these effects.
(6) In any rational fatigue evaluation, the following factors should be considered:
Identification of the structure to be considered in the fatigue evaluation. Those elements of the rotorcraft structure that may be critical in fatigue should be identified. Typical elements include:
(A) Rotor blades and attachment fittings.
(B) Rotor heads, including hubs, hinges, dampers.
(C) Rotor drive components, including gearboxes and transmission shafts.
(D) Control system components, including control rods, servos, swashplates.
(E) Rotor supporting structure.
(F) Primary flight and ground load paths of the fuselage, including landing gear, lift frames, stabilizers and auxiliary lifting surfaces.
(ii) The loads and stresses associated with steady and maneuvering operating conditions expected in service.
Page MG 11 - 2
(III) The frequency of occurrences associated with various flight conditions and the corresponding spectrum of loading and stresses.
(iv) The fatigue characteristics of the structure, including fatigue strength, and as necessary, crack propagation and residual strength characteristics.
c. Flight Load Measurement Program.
(1) General. Subsequent to design analysis, in which aircraft loads and associated stresses are derived, the stress level and/or loads are to be verified by a carefully controlled flight load measurement program. The flight load measurement program shall demonstrate maximum and minimum loads for the entire flight envelope.
It shall also gather steady and cyclic load/strain data for use in the fatigue evaluation required by §§ 27.571 and 29.571. The parameters to be measured are primarily load calibrated strains supported by local strain measurement, accelerations, and deflections as necessary.
(2) Instrumentation.
The instrumentation system used in the flight parameter measurement program should accurately measure and record the critical parameters under operational test conditions. For critical maneuvers, the instrumentation should be capable of recording data sequences for several related channels for stationary and rotating channels. The location and distribution of the strain gauges should be based on a rational evaluation of the critical stress areas. Appropriate analytical methods should be used, such as Finite Element Modeling (FEM) and may be supplemented by other techniques including strain sensitive coatings, photoelastic methods, and thermography. Manual calculations based on the historical precedent of similar structure can also be very helpful. As much as possible, the instrumentation plan should standardize the gauge locations for each component so that all testing and other experience can be related to the same common set of measurement parameters. The gauge sensitivity, frequency response, location, distribution, and number of strain gauges must provide the strain/load spectrum and strain/load distribution for each part essential to the safe operation of the rotorcraft.
(її) The corresponding flight and ground operation parameters (airspeed, rotor RPM, center of gravity accelerations, etc.) should be recorded simultaneously and, where appropriate, as time histories. This is necessary to correlate the loads and stresses with the maneuver or operating condition during which they occurred. If the number of data parameters required exceeds the system capability, enough "carry-over" data channels should be included to reasonably relate all data for the specific maneuver when several flights are necessary.
(iii) The instrumentation system should be adequately calibrated and checked frequently during the strain survey. Ideally this would occur at the beginning AC Page MG 11 - 3
AC 27-1B, Chg 1 2/12/03 and end of a flight. Strain gauges should be temperature compensated where necessary.
(iv) In the case of calibrated structure, rational evaluation of the calibration rig procedure should be performed to demonstrate that strain-load response is representative over the range of flight maneuvers and operational parameters that are to be encountered. As a minimum, the calibration fidelity should address the loading distributions applicable to the critical failure modes of the component under evaluation and over a high percentage of the maximum expected flight load. Care should be taken to ensure that any non-linear behavior is identified and properly considered.
(3) Parts Subject to Flight Measurement. Sufficient parameters of the rotor systems, drive systems, control systems, fuselage, and supporting structure for rotors, transmissions, engines, APU's, and other dynamic components should be measured to adequately define and substantiate the loading of these components. For rotorcraft of unusual or unique design or operation or employing unusual equipment, special consideration should be given to the unique features and also the effects they may have on existing systems and structure.
(4) Flight Regimes and Operational Conditions to be Investigated.
(i) Typical flight and ground conditions to be investigated are given in Figure AC 27.MG 11-7 for conventional passenger/utility use and additional conditions in Figure AC 27.MG 11-8 for a lifting operation. For intensive lifting missions it should be recognized that lifting conditions for both internal and external loads should be investigated including landing with and without load as applicable. Figures AC 27.MG 11-2 and AC 27.MG 11-3 show flight regimes that should be investigated for power-on and power-off operation for all helicopters. Parameters, which define these regimes, are included in these figures. Other parameters such as Gross Weight and C.G. apply and should be included. The effects of temperature and of high altitude operation or altitude cycling should be investigated. As noted on Figure AC 27.MG 11-2, complete coverage at 111 percent VNE should be demonstrated for power-on operation. However, for power-off operation, Figure AC 27.MG 11-3, complete coverage at 111 percent VNE for maximum and minimum design RPM need not be obtained if points are obtained at VNE at both maximum and minimum design RPM and at 111 percent VNE at both maximum and minimum placarded RPM as indicated in the figure. Conditions arising out of special requirements such as those imposed by noise reduction and near surface operation should also be investigated.
(ii) The determination of the flight conditions to be investigated in the flight strain measurement program should be based on the anticipated uses of the helicopter. Information from similar designs and/or similar operations should be assessed and used where applicable. Flight conditions considered appropriate for the design and application must be representative of actual operation in accordance with the rotorcraft flight manual. For multi-engine helicopters, the flight conditions concerning engine out operations should be considered in addition to complete power- Page MG 11 - 4
off operation. For heavy lift and external-lift helicopters the loaded and unloaded conditions of operation should be investigated in conjunction with other important parameters. When the mission being evaluated involves the use of a long line, the flight strain survey should be flown with a long line. This will insure that the correct rotor and control loading will be duplicated particularly in hover and low speed maneuvers since the c.g. offset, drag, and inertia effect of the external load can be a factor in dynamic loads. Generally the VE when performing external cargo missions is reduced because of safety or other reasons. In addition, for these uses, and others where the operation requires frequent excursions to the proposed limitations of the rotorcraft, reasonable consideration should be given to unintentional exceedences and the development of these limits. The extent of the assessment would generally be founded on experience, but also acknowledging tolerances in instrumentation and the circumstances of the operation, particularly the level and focus of pilot workload. Similarly the effects of structurally significant maintenance should be investigated as necessary, particularly where unusually large load amplifications could be expected as a consequence and the probability of occurrence is shown or believed to be substantial based on experience.
Operating limitations and maintenance instructions may be adjusted on the basis of these investigations. For all rotorcraft the effects of ground operation should be investigated. The ground and flight conditions to be investigated should be submitted as part of with the flight evaluation program.
The severity of maneuvers investigated during the flight strain survey should be such that it is extremely unlikely that service use will be more severe. In this evaluation, flight replications should be investigated during the load survey so that normal and expected variations in achieving the specific target flight test conditions are accounted for.
(iv) The extremes of aerodynamic configuration, including operation with doors off or open and simultaneous use of equipment, should be investigated as well as the usual parameters. In addition, when the rotorcraft is equipped with externally mounted devices such as rescue hoist, spotlight, camera, infrared sensors, etc. it is necessary to evaluate the loads. These or similar devices can increase the profile drag thus increasing power required. Additionally, the airflow from these devices can impinge on the tail boom, fin or tail rotor altering the loads determined for the clean aerodynamic configuration. When these devices are installed by a modifier as a Supplemental Type Certificate (STC), it is the responsibility of the STC holder to investigate any effects the device may have alone or in combination with other externally mounted devices.
All flight conditions considered appropriate for the particular design are to be investigated over the complete rotor speed, airspeed, longitudinal and lateral center of gravity, altitude and weight (from minimum mission weight to maximum mission weight) ranges to determine that the critical loads/stresses associated with each flight condition are identified. Typical damaging flight conditions for main rotor components and suspension include, but are not limited to: high speed flight, turns, pull-ups, sideward flight, approach, autorotation, taxiing, take off and landing on slopes. For tail rotor AC Page MG 11 - 5
AC 27-1B, Chg 1 2/12/03 components, typical flight conditions to evaluate include, but not limited to: spot turns in hovering, sideward flight or level flight with sideslip. In order to account for data scatter and to determine the load/ stress levels present, a sufficient amount of data points should be obtained at each flight condition. In some instances, the critical weight, center of gravity, and altitude cases for the various maneuvers can be based on validated flight loads analysis or on past experience with similar designs. This procedure is acceptable where adequate flight tests are performed to substantiate such selections. The combinations of flight parameters that produce the most critical load/stress levels should be included in the fatigue evaluation. In addition attention should be paid to the influence of temperature through the whole range certified, especially to very cold temperature (typically below -30°C [= -20°F]). For example, the characteristics of elastomeric components in rotor assemblies are particularly sensitive to temperature change. This influence can be evaluated either in flight during a cold weather campaign or by analysis based on elastomer characteristics such as stiffness and damping measured in a cold temperature chamber.
d. Frequency of Occurrence (Usage) Spectrum.
(1) General. The frequency of occurrence spectrum (often called usage spectrum) defines the maneuvers the rotorcraft will perform in the various types of operation, and the percentage time or number of events associated with each maneuver. The diversity of rotorcraft sizes and speed together with the wide range of passenger/cargo capability requires that a comprehensive evaluation of each possible mission scenario be accomplished for each rotorcraft model. Some of the most common types of operation include transport, offshore support, traffic reporting, emergency medical services (EMS), law enforcement, search and rescue, agricultural spraying and external sling operation. Each of these operations has unique requirements in terms of maneuvers, gross weight/center of gravity and altitude.
However, each helicopter model may fly one or more of these operations throughout its operational lifetime. Replacement times should account for the worst case operation for each component, unless a method approved by the Authorities is developed which allows consideration of multiple type of operations by factoring hours or counting
(2) Spectrum Development. The frequency of occurrence spectrum should be based on information that is applicable to the mission(s) the rotorcraft is to perform. All damage that is likely to occur in actual usage should be accounted for including low cycle damage from power cycles and the ground-air-ground (GAG) cycle. This information may come from direct measurement of usage data from the same or similar rotorcraft, usage monitors, questionnaires or direct observation of the helicopter performing the mission. Design limitations established in compliance with §§ 27.309 or
29.309 and any recommended operating conditions and limitations established and specified in the rotorcraft flight manual should also be reflected in the spectrum. An example of a twin turbine spectrum is presented in Figure AC 27.MG 11-9. This table should be used only as a guide and should be modified as necessary for each particular rotorcraft. An example of the diversity in the frequency of occurrence spectrum is Page MG 11 - 6
illustrated by comparing percentages of time assigned to level flight conditions for three different rotorcraft types as shown in Figure AC 27.MG 11-1 below:
Piston Utility Turbine Utility Business Twin Turbine Transport
0.8 VNE 25% 0.8 VH 16% 0.8 VH 15%
1.0 VH 15% 0.9 VH 21% 0.9 VH 20%
1.0 VNE 3% 1.0 VH 24% 1.0 VH 38% TOTAL 43% 61% 73% FIGURE AC 27.MG 11-1: Example of the variation in time spent in level flight for three rotorcraft types.
Not only are the totals different for the different rotorcraft types but the distributions of time are also significantly different. This can become an important factor in the determination of fatigue lives, whether or not there are damaging loads in level flight, depending on how the time spent in other maneuvers is subsequently proportioned. A conservative approach to the spectrum development should be taken. It is suggested that a sensitivity study be conducted to determine the variability of component lives to different assumed percent times for level flight. This same procedure might also be used for other elements of the spectrum where significant fatigue damage is incurred.
The results from such a study may be used to influence the spectrum or the replacement time assigned to the components).
(3) External Load Operations. The unique ability of a rotorcraft to hover makes it particularly useful in moving external cargo. External load operation can be a demanding mission requiring the maximum lifting and power capability of the rotorcraft at a high rate. For example, a logging operator may use up to 50 maximum power cycles per flight hour to move logs from a cutting site to a hauling site. The power reaches a maximum limit when the load is lifted and the rotorcraft accelerates. The power reaches a minimum during the descent and will peak again if the rotorcraft decelerates and transitions to hover to release the load. Other external cargo operations with similar characteristics are fertilizer spreading, water bucket operations, and replenishment of remote oil exploration sites, etc. These power excursions are particularly critical for the rotorcraft drivetrain components. The impact of external load operation should be assessed to determine if replacement times would be affected.
(4) Management of Replacement Time.
The lowest calculated life obtained from all flight loads data and loading spectrum (including external load operations) is generally the basis for establishing the replacement time of the component(s). Regulatory maintenance and operating rules do not require recording time-in-service for different types of operations.
AC Page MG 11 - 7
AC 27-1B, Chg 1 2/12/03 However, it may be possible to adjust replacement times by counting events or by factoring flight hours for certain types of operation. Any such procedure will require the approval of the certification authority for a suitably amended airworthiness limitation section and should also consider the operational aspects involved. For example, a component where significant damage occurs during the GAG or power cycle may more appropriately be assigned a life in terms of the number of cycles of takeoffs or lifts in lieu of tracking flight time in hours. The cycles should be properly defined and related to events that are easily identified and recorded by the crew or by an approved usage monitor (e.g., takeoffs or lifts). This example procedure would retire the component sooner when it is used in an external load operation mission involving many cycles per flight hour. Conversely, an operator performing a less severe operation standard mission could leave the component in service longer since fewer cycles are accumulated. This procedure would permit the rotorcraft to be used for different types of operations and still ensure a safe replacement time for the component.
(II) Where appropriate, some of the basic usage assumptions made in the fatigue evaluation which the operator can reliably assess (such as numbers of ground air ground cycles) should be noted in the airworthiness limitations section of the maintenance manual. The intent of this would be to make operators aware of these criteria so that appropriate actions may be taken.
(iii) Should subsequent usage of the rotorcraft encompass an operation for which the original structural substantiation did not account, the effects of this new operation should be addressed, and in the interests of safety, a reassessment made.
Subsequently, if the replacement times require revision, those new times may be limited to aircraft involved in the new operation provided:
(A) Proper part re-identification is established;
(B) a rotorcraft flight manual supplement outlining limitations is approved;
(C) an airworthiness limitations section supplement is approved; (this is also required for incorporation of new methods for managing replacement times, see paragraph d4(i) above); or, (D) a combination of the above.
e. Fatigue Life Evaluation.
(1) General. Information for fatigue evaluation based on safe-life considerations leading to replacement times is provided in this section. Although there is a large quantity of information available on the fatigue strength characteristics of material specimens, built-up specimens and parts, the prediction of the strength of parts of new designs based on this information is less accurate than testing the actual part.
Consequently, for an analysis based on test data other than the actual part to be considered acceptable, additional conservatism should be used to achieve similar levels Page MG 11 - 8
of reliability and safety as obtained with a full scale test approach. However, in many cases the differences between past test specimens and the actual part (which involve such factors as stress concentration, size, and fretting) cannot be accounted for with a reasonable degree of accuracy. Therefore, it is usually necessary that the structural components be subjected to repeated load tests using information determined in the flight load measurement program. Special operational or functional characteristics that could affect the fatigue strength should also be considered in the service life evaluation.
Such factors as high blade operating temperatures due to tip jets or turbine exhaust impingement on the tail rotor should be considered as well as other special operating conditions. In addition, effects of special purpose use such as hoist and external operation, spraying, surveying, etc., should be considered if appropriate to the particular type. The fatigue strength may be evaluated using the methods outlined below, of which full scale testing is the preferred method.
(2) Analytical Methods Simplified method. This method requires that an operating boundary for stress levels be established. The following techniques that account for the effects of cyclic and steady stresses are considered acceptable for establishing the allowable stress levels:
(A) The mean endurance limit of the part should first be estimated from simple material test specimens. The test specimen material should be representative of the actual part and sufficient test data should be available to substantiate a mean endurance limit (the reference specimen endurance limit, Line AD of Figure AC 27.MG 11-4). A range of cycles from 107 to 10° may be appropriate to estimate the endurance limit dependent on the material. The estimate should account for surface conditions, fabrication methods, fretting, size and shape effects, and environmental conditions, as well as differences in stress concentrations between the test specimen and the actual part. Referring to Figure AC 27.MG 11-4, the endurance limit of the part may be represented by a straight line drawn through the yield stress (point D on the horizontal axis) and the mean endurance limit from test results, suitably adjusted to account for the considerations detailed above, at a given steady stress. The intersection of this line with the vertical axis is point B. This produces the adjusted specimen endurance limit, line BD.
(B) A factor or safety of 3 should then be applied to the adjusted specimen endurance limit so that the slope of line CD (the operating boundary) would be 1/3 of line BD.
(C) If all operating stress cycles fall below the operating boundary line (CD), no fatigue testing is necessary. When any of these stresses are above the operating boundary line, fatigue testing of the actual parts should be conducted, unless a suitably conservative approach such as that outlined in d(4)(ii) is adopted.
AC Page MG 11 - 9
AC 27-1B, Chg 1 2/12/03 (D) Caution should be exercised in the application of the analytical method above, particularly when the following items are involved:
(1) Irregularly shaped parts containing numerous or super-imposed fillets, holes, threads, or lugs.
(2) Large parts in proportion to the laboratory specimens.
(3) Parts or unique design for which no past service experience is available.
(4) Parts subject to fretting.
(5) Bolted or pinned connections.
(6) Complex castings.
(Z) Welded sections.
(8) New materials or processes without precedent of use.
(ii) Rational methods. The previous simplified method can be overly conservative, especially when only ground-air-ground cycles or very high loads associated to very low occurrences fall over the operating boundary line defined in paragraph e(2)(i) above. Consequently methods may be used which do not involve full scale testing but which apply the same principles of calculation of retirement times, based on:
(A) An S/N curve shape representative of the material of the component.
(B) A mean fatigue limit representative of the component, considering as necessary, the steady flight loads, fretting effects, and all the other influential parameters of paragraph e(2)(i)(A).
(C) An appropriate factor applied to the mean fatigue strength to produce a working limit (typically 1/3 of mean strength).
(D) Consideration of all the loads from the complete flight loads spectrum.
(E) The use of the Miner linear cumulative damage hypothesis, including both low cycle and high cycle fatigue damages.
(iii) In order to provide an acceptable alternative to the fatigue testing and simplified analytical methods, these rational methods should be based on a validated stress analysis. Finite element model correlated to strain gauge measurements for example, or previous experience of similar designs may be acceptable. The material Page MG 11 - 10
fatigue behavior should be well established for each application. It is important to apply the chosen method consistently and should any of the analyses identify the need for a replacement time, testing should be conducted to support the assumptions made. The certifying authority should approve these rational methods.
(3) Testing Methods. The fatigue strength of the flight structure may be determined in appropriate laboratory tests and evaluated in terms of the loading spectrum. The mean strength indicated by the test results should be reduced by a factor or factors such that the probability of occurrence of a lower strength part in service is very low. This conservative treatment of strength combined with a conservative treatment of both the flight loads (paragraph c.) and their frequency of occurrence (paragraph d.) must assure that the probability of failure is extremely remote. All test articles should be fully representative of the design standard selected for evaluation, including the processes used in manufacture. The test fixture should be capable of applying the loading conditions in a way that loads the component in the same manner as when on the rotorcraft. The test loads developed in the component should be correlated with those measured in the flight load survey.
S/N Curves. Constant amplitude fatigue tests should be conducted to define the mean strength. Whenever possible several S/N data points should be established for each of a number of different alternating load levels. The fatigue tests should be performed at mean stresses or loads representative of those occurring in flight. In addition, some components subjected to both dynamic and low cycle loading, for example a main rotor blade, may require the addition of a start-stop or GAG cycle testing, and the resulting fatigue damage included in the component life determination.
In order to determine the mean fatigue strength, it is necessary to test actual components. These tests will allow the construction of the mean S/N curve when combined with an established curve shape. The S/N curve shape may be derived by using a least square fit curve through coupon data or appropriate published material data. Care should be taken in the selection of curve shape, particularly when fretting is present. Then, to account for fatigue strength variability, the mean curve must be reduced to a working curve. In establishing the working curve, consideration should be given to the number of specimens tested, the variability of the fatigue results, previous test data on the same material or similar components, as well as service experience. At least four full-scale specimens are recommended, but fewer may be adequate in association with a conservative approach to establishing the working curve considering the reduction in reliability this infers. Current practice shows that when four or more specimens are used, the resulting working curve (Figure AC 27. MG 11-5) can range from 51% to 70% of the mean curve in strength for aluminum alloys and 56% to 75% for steels. The successful application of the resulting working curves will depend on the degree of conservatism shown in the flight loads and occurrence spectra. Therefore it follows that use of the least conservative of these working curves would necessitate the greatest conservatism in the flight loads and assumptions relating to likely operational use. Consideration should also be given to fatigue life reduction factors when constructing the working curve. Typical factors range from around four to ten at less than 100,000 cycles. It may be possible to determine reduction factors from a large AC Page MG 11 - 11
AC 27-1B, Chg 1 2/12/03 database of historical test data of components with similar characteristics. Care should be exercised when pooling such data to be sure that difference in failure modes and curve shapes are considered. Whatever reduction factor is selected, a rationale should be provided to substantiate it. The reduced S/N curve and the loading spectrum developed per paragraph c. and d. should be used in determining replacement times, see paragraph e(4).
(II) Spectrum tests. The establishment of replacement times based on fatigue tests in which each specimen is subjected to a spectrum of loading is to include the following considerations:
(A) Definition of the test loading spectra based on either:
(1) Load histories based on flight test data obtained for flight and ground conditions and maneuvers considered appropriate for the particular rotorcraft, and a spectrum allocating percentages of time or frequencies of occurrence to these flight and ground conditions and maneuvers, or
(2) Analysis supported by extrapolation of available load history data or prior knowledge where available.
(B) The effects of high infrequent load cycles on the test result particularly when such cycles may occur only rarely in service.
(C) The effects of omitting low load (high frequency) cycles to reduce test time should be fully established and supported by test experience to be adequately accounted for.
(D) Fatigue tests in which the loading spectra are applied such that effective randomization of loading is obtained.
(E) Assignment of replacement times. The fatigue test results should be evaluated in terms of the loading spectrum of paragraph d. if different to the test spectrum, and reduced by factors for strength and life based on similar approach to those derived for the constant amplitude tests above.
(4) Safe-Life Calculation Methodology. The key procedures for safe life determination, as shown hereafter, are based on the three basic elements of strength, loads, and usage as established in the preceding sections. These elements are reiterated below and combined according to Figure AC 27.MG 11-6 to calculate the retirement life: 1) The conservative working S/N curve developed from the mean S/N strength curve of the component using reduction factors based on the material and manufacturing variability and test parameters. 2) Loads and usage combined in an individual fatigue loads spectrum for the component determined conservatively through test and analysis. The loads may be processed by conservative methods using maximum load for the duration of each condition or by suitable cycle counting methods.
Page MG 11 - 12
GAG cycles, once per maneuver cycles, and other high load cycle events should be accounted for as the loads are combined with the established usage spectrum. The service life of the component may then be determined using Miner's linear cumulative damage rule, considering both high and low cycle fatigue damage. The calculated service life obtained for a total damage equal to one is then the maximum allowable replacement time for the component.
f. Fail-Safe Evaluation.
(1) General. The fail-safe evaluation of the flight structure is intended to ensure that, should fatigue cracks initiate, the remaining structure will withstand service loads without failure until the cracks are detected. The fail-safe evaluation generally encompasses establishing the components which are fail-safe, defining the loading conditions and extent of damage for which the structure is to be designed, conducting structural tests and analysis to substantiate that the design objective has been achieved, and establishing inspection programs to assure detection of fatigue damage.
Design features that may be used in attaining a fail-safe structure are:
Selection of materials and stress levels that provide a controlled slow rate of crack propagation combined with high residual strength after initiation of cracks.
(i¡> Design to permit detection of cracks, including the use of crack detection systems, before the cracks result in an appreciable loss of residual strength.
(#) Use of multi-path construction and the provision of crack stoppers to limit the growth of cracks.
(iv) Use of composite duplicate structures so that a fatigue crack or failure occurring in one element of the composite member will be confined to that element and the remaining structure will still possess limit load-carrying ability. It may be necessary to employ the design techniques of f(ii) above to assure effectiveness of these features.
(v) Use of backup structure wherein one member carries the entire load, with a second member available and capable of assuming the load if the primary member fails.
(2) Extent of Fail-Safe Damage. The extent of the partial failure is to be such that it would be readily detectable during the specified inspection. It may involve complete failure of a principal element, failure of more than one element, or only a partial failure of an element, depending on the rate of crack propagation, the ease of detection, and the inspection interval. Damage in inaccessible areas should extend into inspectable areas.
Typical examples of the fatigue damage that should be considered are outlined below:
AC Page MG 11 - 13