AC 35.37-1A

Subject: GUIDANCE MATERIAL FOR

FATIGUE LIMIT TESTS AND COMPOSITE

BLADE FATIGUE SUBSTANTIATION Date: 9/17/01 Initiated By: ANE-110 AC No: 35.37-1A Change:

1. PURPOSE. This advisory circular (AC) provides guidance and describes methods, but not the only methods, for demonstrating compliance with §35.37 of Title 14 of the Code of Federal Regulations, Fatigue limit tests. This AC also provides methods for the fatigue evaluation of propellers and the fatigue substantiation of composite propeller blades. Like all AC material, this AC is not, in itself, mandatory and does not constitute a regulation. While these guidelines are not mandatory, they are derived from extensive Federal Aviation Administration (FAA) and industry experience in determining compliance with the applicable regulations.

2. CANCELLATION. AC 35.37-1, Composite Propeller Blade Fatigue Substantiation, dated 5/11/93, and AC 35.37-1, Change 1, dated 9/7/93, are canceled.

3. RELATED DOCUMENTS.

a. Regulations. Sections 35.4, 35.37, 23.907, and 25.907.

b. Advisory Circulars.

(1) AC 20-66A, Vibration and Fatigue Evaluation of Airplane Propellers, dated 9/17/01.

(2) AC 20-107A, Composite Aircraft Structure, dated 4/25/84.

(3) AC 21-26, Quality Control for the Manufacture of Composite Structures, dated 6/26/89.

(4) AC 25.571-1C, Damage Tolerance and Fatigue Evaluation of Structure, dated 4/29/98.

Page 2 Par 3

c. Military Specifications. MIL-HDBK-17-1C, dated November 4, 1992.

4. DEFINITIONS. For the purposes of this AC, the following definitions apply.

a. Damage tolerance. Damage tolerance is the attribute of the structure that permits it to retain its required residual strength for a period of use after the structure has sustained a given level of damage from fatigue, corrosion, accident, or a discrete source.

b. End of life condition. The end of life condition is the physical condition of the component when it has sustained the maximum extent of damage but maintains sufficient residual strength to meet all airworthiness loading requirements. This condition is defined during certification.

c. Fail-safe. Fail-safe is the attribute of the structure that permits it to retain its required residual strength for a period of use without repair after the failure or partial failure of a principal structural element.

d. Fixed pitch wood propellers of conventional design.

(1) A fixed pitch wood propeller of conventional design is a propeller that has the following physical properties:

(a) One piece laminated wood construction;

(b) Two or four blades;

(c) Surface coatings that do not contribute significantly to the propeller strength; and

(d) Surface coatings that only provide environmental protection.

(2) A fixed pitch propeller that has a composite shell over a wood core does not qualify as conventional design if the composite shell contributes significantly to the strength and frequency response of the propeller.

(3) A fixed pitch wood propeller with a fabric or composite covering for environmental protection that does not significantly alter the structure qualifies as conventional design.

e. Flaw. A flaw is a pre-existing anomaly in the structure that is either created during manufacture or has resulted from damage to the structure introduced after manufacture.

Generally, flaws can be geometrically quantified by non-destructive inspection methods. The term flaw is widely used in fracture mechanics disciplines for assessing damage tolerance.

Par 4 Page 3

f. Hazardous propeller effects. The following conditions are considered hazardous propeller effects:

(1) A significant overspeed of the propeller.

(2) The development of excessive drag.

(3) Thrust in the opposite direction to that commanded by the pilot.

(4) A release of the propeller or any major portion of the propeller.

(5) A failure that results in excessive unbalance.

(6) The unintended movement of the propeller blades below the established minimum inflight low-pitch position.

g. Life factor. This is a factor applied to increase or decrease the number of cycles on an S-N diagram as needed to properly account for test sample size and variations in material properties, environmental effects, service deterioration, repairs, and manufacturing anomalies. This is often used in combination with a load factor, as shown in Figure 1.

h. Limit load. This term refers to the maximum load expected in service.

i. Load factor. The load factor is a factor applied to increase or decrease load or stress level on a fatigue life diagram (S-N diagram) as needed to properly account for test sample size and variations in material properties, environmental effects, service deterioration, repairs, and manufacturing anomalies. This is often used in combination with a life factor, as shown in Figure 1.

j. Principal structural element. This term refers to the element that contributes significantly to the carrying of propeller loads and whose integrity is essential in maintaining the overall structural integrity of the propeller.

k. Safe-life. The safe-life of a structure is that number of events such as stress cycles, flights, landings, or flight hours, during which there is a low probability that the strength will degrade below its design value due to fatigue damage.

l. Scatter factor. The scatter factor is a life reduction factor used in the interpretation of fatigue analysis and test results.

m. Stress ratio (R). For repeated stress cycles, this is the ratio of the minimum stress to the maximum stress, R = σmin / σmax .

Page 4 Par 5

5. DISCUSSION. Propellers are continuously subjected to steady and vibratory stresses under many different operating conditions on an airplane, both while in flight and on the ground.

Therefore, §35.37 requires a fatigue evaluation and the determination of fatigue limits. This requirement also provides data that supports installation of the propeller on an airplane. Since the rate of accumulation of stress cycles for propeller blades, hubs, and other propeller components is very high, the design goal, whenever possible, should be to show that stresses are below the component or material endurance limit. However, not all materials have a welldefined endurance limit. In addition, the stresses developed during maneuvers, ground operation, ground air ground (GAG) cycles or in other areas of the airplane operating envelope may cause damage. The accumulation of this damage should be evaluated to determine if propeller components are life limited or require mandatory inspections, or to determine if the propeller is suitable for use on an airplane.

a. Fatigue Limits. Establishing propeller fatigue limits is one step in the structural evaluation for certification and approval for use on an airplane. Fatigue limits refer to the material and structural databases used to support the fatigue evaluation of the propeller on an airplane.

Fatigue limits may take many forms, such as Goodman diagrams to assess whether or not stresses are below the endurance limit, S-N curves for safe-life evaluations, and crack growth curves (da/dn curves) and delamination growth curves for damage tolerance evaluations. The fatigue limits are developed with an understanding of the propeller component, material, failure mechanism, environment, and loading and may also include the effect of service damage. When determining the fatigue limits under §35.37, the possible airplane operating environment should also be considered. The development of fatigue limits can be supported by previous testing, past experience, and acceptable published data, when available.

b. Damage Accumulation. The fatigue limits are developed to support a damage accumulation algorithm, such as Miner’s rule for safe-life calculations or a crack growth or damage growth algorithm for damage tolerance calculations. The damage accumulation algorithm is used for the fatigue evaluation required by §35.37 and may be also used as part of showing compliance with the airplane vibration requirements in §§23.907 and 25.907. The damage accumulation algorithm should be verified by previous testing, past experience, and acceptable published data, when available.

c. Applicable Components.

(1) Components. For propeller certification, fatigue limits are established for the propeller hub, blades, and blade retention. They should also be developed for those components whose failure due to fatigue may cause a hazardous propeller effect. Examples of components whose failure may cause a hazardous propeller effect are the pitch change piston pressure cylinder (dome), counterweights, and pitch control components. For items such as bearings, which are typically part of the hub assembly, the fatigue limits are generally established for the assembly, not for the individual components.

Par 5 Page 5

(2) Regions. Each applicable component should be assessed to determine if it has multiple critical regions or if it requires different limits for different regions. Blades are an example of a component that has different limits for different regions. The blade steady stresses vary substantially from the blade root to the blade tip; the blades should have fatigue limits established for each of the stress regions. Also, metallic blades may have different material limits to account for local working effects due to shot peening or cold rolling.

d. Propeller Loads. The loads applied during the fatigue limit tests are derived from the consideration of the steady and vibratory propeller loading conditions that occur on the intended airplane and engine installation throughout its life or on a typical airplane. The loads applied during fatigue limit tests reflect the fatigue data to be generated. For example, the loads establish the stress ratios (R) for coupon specimen tests and set the amplitude and direction for full-scale testing. Also, the load magnitude and direction should be established in a manner that represents the loading the propeller will experience in service.

(1) Operating Spectrum Conditions. The applied loads are derived from an airplane operating spectrum. The operating spectrum depends on the category and operation of the airplane and includes both flight and ground operation.

(a) Flight conditions include conditions that occur with each flight such as:

1. Take-off;

2. Climb;

3. Cruise;

4. Descent;

5. Approach;

6. Landing; and

7. Reverse thrust.

(b) Transient airplane flight conditions are associated with maneuvers such as:

1. Banked turns;

2. Side-slip;

3. Pull-ups;

4. Push-overs;

Page 6 Par 5

5. Rudder kicks; and

6. Gusts.

(c) Special flight conditions are specific to a mission such as:

1. Fire-fighting;

2. Aerobatic maneuvers;

3. Emergency conditions; and

4. Training maneuvers.

(d) Ground conditions include but are not limited to:

1. Taxi;

2. Operation in cross-winds; and

3. Maintenance checks.

(e) Engine load conditions include the loads generated by the engine and transmitted to the propeller, such as firing impulses from reciprocating engines.

(2) Operating Spectrum Source. The airplane operating spectrum, when available, may be obtained from the airplane company for the intended application. When the airplane operating spectrum is not available, the spectrum may developed from design assumptions and design experience on a typical or on the intended airplane.

e. Component Degradation. The fatigue limits should account for likely service deterioration, variations in material properties, manufacturing anomalies, and environmental effects. Many methods are available to account for component degradation, such as using a life or load factor to reduce the fatigue limits or apply to fatigue test loads. Reductions in fatigue limits may be applied as life or load factors based on the manufacturer’s service experience and test database with the components, as shown in Figure 1. The reductions may also be developed by appropriate specimen tests or by testing intentionally degraded full-scale components.

(1) Manufacturing Anomalies. Propellers manufactured to a process specification have established parameters to control manufacturing variables. These parameters should be considered for their impact on structural integrity.

Par 5 Page 7

(a) For metal propellers these include, but are not limited to, surface finish and machining marks.

(b) For composite propellers these include, but are not limited to, fiber misalignment, fiber content, porosity, and delaminations.

(c) For adhesive bond joints these include, but are not limited to, adhesive thickness, voids, and disbonds.

(2) Service Damage. Propellers are exposed to and encounter a certain amount of service damage from corrosion, erosion, stone strikes, and handling damage. Composites, in addition, may be damaged by small bird strikes and hail impact. The damage to a composite should consider both visible surface damage and hidden internal damage, such as delaminations. The extent of tolerance to such damage and the method of demonstration should be considered.

(3) Environmental Degradation. Airplanes can spend their entire operational life in a severe climatic zone. Therefore, the strength degradation of the material system should be established. When assessing operational environments, the following should be considered:

(a) High temperature;

(b) Humidity;

(c) Low temperature;

(d) Thermal cycling;

(e) Ultraviolet light; and

(f) Aviation chemicals.

f. Repairs. Repairs change the condition of the propeller as manufactured. Therefore, the manufacturer should consider the impact of planned repairs on the structural integrity of the propeller.

g. Endurance limit. When conducting coupon tests, the coupons should be tested to failure under a combination of steady and vibratory loads. The steady loading should represent that anticipated in service for the propeller. The vibratory loading component should be selected to facilitate the generation of a fatigue or S-N curve, over a broad range of stress cycles, to define an endurance limit.

(1) The endurance limit for composites is normally projected to at least 500 million cycles.

(2) The endurance limit for aluminum is normally projected to at least 100 million cycles.

Page 8 Par 5

h. Accelerated Testing. The vibratory loading may be at any frequency, if internal heating does not produce harmful effects and representative failure modes are realized. Attempts to compress testing time by using high-frequency vibratory loading should be approached with caution, to prevent the introduction of non-representative temperatures, failure modes, and fatigue lives. Accelerated testing should include some form of temperature monitoring to prevent overheating the specimen. Cooling may be required to prevent overheating the component.

i. Coupon Tests. Establishing an S-N curve from coupon tests is recommended for determining fatigue strength and statistical strength distribution of propeller materials. The S-N curve should be established for both the low cycle fatigue (LCF) and high cycle fatigue (HCF) regimes. The S-N curve should be representative of the propeller’s material system. S-N curves should be developed using acceptable published data, when available, or by testing a sufficient number of specimens that represent the propeller’s material system manufacturing processes.

(1) Metals. For metals, this should include the following effects:

(a) Surface finish;

(b) Cold rolling;

(c) Shot peening; and

(d) Corrosion inhibitors.

(2) Composites. For composites, this should include the following effects:

(a) Fiber, resin and sizing;

(b) Ply stacking sequence;

(c) Ply orientation; and

(d) Manufacturing processes.

(3) Bond Joints. For bond joints, this should include the following effects:

(a) Adhesive materials;

(b) Surface preparation;

(c) Adhesive thickness; and

(d) Manufacturing processes.

Par 5 Page 9

(4) Propeller Stations. The propeller’s station-to-station variance between metallic to composite bond regions, shank sections, mid-blade section, and tip regions may warrant the testing of coupons representative of each.

j. Full-scale Testing.

(1) Test Specimen. Full-scale components should be manufactured to represent the type design and should be fatigue tested at combinations of steady and vibratory loads as needed to support subsequent evaluations. Also, to address the issues of safe-life, damage tolerance, and continued airworthiness, the specimens should be manufactured to include specific manufacturing anomalies and likely service damage. The extent of such defects/damage should be consistent with inspection techniques employed in service.

(2) Failure Criteria. Rational failure criteria should be established. This will be different for various types of construction and materials. Composites tend to fail between layers first, while metals fail through the thickness of the part and have much smaller permissible defect sizes.

(a) For composites, as shown in Figure 2, fatigue failure is markedly progressive, starting at initiation (region I) through a damage growth or delamination growth phase (region II) to an advanced damage state in which large delamination and secondary failure modes and locations form throughout the component (region III). Failure criteria such as a specific loss of stiffness in the structure, visible damage, and delamination area should be applied.

(b) For metallic components, the failure criteria may be the initiation of a fatigue crack or a specific crack length. The selected failure criteria will be a factor in establishing the component life, since the component should maintain sufficient residual strength throughout its life.

(c) When using stiffness loss as failure criteria, the effect on component resonant frequency placement should be addressed, as this may affect the continued airworthiness of the component. Vibratory loads may be magnified as component natural frequencies approach the frequency of the load.

(3) Component Monitoring. Components should be strain gaged and have load cells, as appropriate, for testing to monitor stresses and loads. When a test rig applies load with an amplitude actuator, the input amplitude may require adjustment to account for test rig and/or component wear in or degradation, to assure that the test load is maintained. The specimen should be examined regularly for cracks, delamination, or other degradation. When using stiffness as the failure criteria, it should be measured for each specimen at the outset of each test and periodically monitored throughout the test. The monitoring should be more frequent during the failure process. When appropriate, testing should be continued after the component has failed, according to the failure criteria, to demonstrate damage growth characteristics and residual strength capacity. This may include testing to the limit load conditions after LCF or HCF tests are completed to demonstrate residual strength at the end of life condition.

Page 10 Par 5

k. Data Reliability.

(1) Mean S-N Diagram. The mean trend line for the data for each critical location should be defined from the mean steady and vibratory stress for all coupons/components tested. This mean trend line should be extrapolated out to an asymptote representing the HCF endurance limit and extrapolated back to a low number of cycles. The coupon-derived S-N curve shape may be used to supplement extrapolation through the failure points when the coupon tests duplicate the failure mechanism.

(2) Reliability. As dynamically loaded components, the structural integrity of propellers is generally governed by their fatigue, rather than their static strength. Accordingly, a reliability at least as good as the “A” basis of a normal distribution should be demonstrated. That is a reliability of 99 percent, with a 95 percent confidence level.

(3) Number of Test Specimens. There is no set number of full-scale test specimens that should be fatigue tested for each critical propeller section, if the required reliability is satisfied.

(a) For “A” basis reliability, the mean endurance limit (E50) should be reduced by a factor (k), governed by the selected sample size (n) and the standard deviation (σ) data.

Specifically, the 99 percentile endurance limit (E99) may be expressed: E99 = E50 -k σ; in which

(k) is a function of the sample size (number of specimens tested). The “A” basis values may also be defined by other distributions such as Weibull or Log Normal, depending upon the fit. Refer to MIL-HDBK-17-1C for details.

(b) For the normal strength distribution, 0.95 confidence, (k) varies with sample size as follows:

Sample size Normal Distribution Reduction Factor nk 2 37.094 3 10.553 4 7.042 5 5.741 6 5.062 7 4.642 8 4.354 9 4.143 10 3.981 REF. Table 8.812, MIL-HDBK-17-1C

(c) The manufacturer’s in-house methodology, development data, and analytic expertise, along with limited full-scale testing, may be used as an alternative in substantiating the required reliability.

Par 5 Page 11

(4) Working Curve. The working curve is the design curve. It is developed from the “A” basis curve by dividing by appropriately justified load or life factors to account for the manufacturing variations, environment, and service conditions. Factors for composite materials should be developed for each new composite material. These materials include, but are not limited to, fiber, resin, coating, and adhesives.

6. COMPOSITE COMPONENT FATIGUE LIMITS.

a. Discussion. An increasing number of propeller blades are constructed of composite materials. Due to the anisotropic characteristics of composite materials, component design and verification of design differ significantly from that of metal structures; propeller blade fatigue loading differs significantly from most other structures. Past practices for metal structures or other composite structures may not be adequate when dealing with composite propeller structure and fatigue loading. Although this discussion focuses on propeller blades, the principles may be applied to any composite structure. With the added flexibility of composite component design, damage tolerance and fail safe principles should be included whenever possible. This AC assumes that when damage tolerance methods are applied the component has been designed using damage tolerance and fail-safe principles.

(1) Failure Progression. Figure 2 illustrates the typical progression of fatigue degradation and reduction in residual strength of a typical composite component. Region I is the damage initiation phase. Region II is the damage growth or delamination growth phase; this is the region in which delamination growth propagates steadily and predictably. Region III is an advanced damage state in which large delamination and secondary failure modes and locations form throughout the laminate. In region III, component stiffness and residual strength degrade rapidly.

This AC provides guidance for fatigue limits to be used in a fatigue evaluation, to assure that region III is not reached in service.

(2) Fatigue Limits. Fatigue limits should be developed to support a fatigue evaluation based on a safe-life approach or a damage tolerance approach. This AC provides guidance for both of these approaches. When appropriate, the fatigue limits should address permissible repairs that alter the as manufactured condition of the propeller.

b. Safe-Life Fatigue Limits. Developing the fatigue strength characteristics for composite components to be used in a safe-life-fatigue evaluation should be done with a two phase approach. The first phase is coupon specimen testing for both static properties and fatigue properties, to supplement the fatigue (S-N) curve shape, standard deviation, and statistical distribution. The second phase is full-scale specimen testing to establish the strength level, statistical distribution, failure location, and failure mode and mechanism for the material system and geometry of the propeller. For blades this may involve testing of metallic to composite bonds, shank, mid-blade, and tip specimens. Figure 3 illustrates some types of test rigs that have been used to assess various areas of a blade. The development of the fatigue properties may be supported by previous testing, past experience, and acceptable published data, when available.

Page 12 Par 6

(1) Loading.

(a) Two types of loading may be needed for testing full-scale composite structures.

Constant loading defines the S-N curve. Spectrum loading addresses cumulative loading effects, since damage is dependent on both LCF and HCF loads. These loads include GAG cycles, low occurrence high amplitude maneuver limit loads, and high occurrence take-off, climb, and cruise loads. The fatigue characteristics may also depend on the order in which the loads are applied;

this should be considered when establishing the test load spectrum. This effect should be addressed by performing many sets of load blocks.

(b) LCF and HCF are closely related; both of these should be addressed. Full-scale testing of composite components usually emphasizes LCF initially, since composite S-N diagrams have a shallow slope. However, unless the coupons duplicate the mode of damage, coupon tests have limited value in areas with complex geometry or metallic to composite bonds.

(c) Constant amplitude HCF failures are generated in the 106 to 108 cycle range. To induce failures in this cycle range, the alternating component of the load(s) is generally increased by a load factor from design or operating values. The steady component of load is generally not increased to induce failures. If no failure occurs, a no failure point is generated. No failure points are known as run-out points. Specimens that have completed a run-out should not be retested to generate failure points at higher load levels. The run-out specimen may have accumulated damage during previous testing that could bias the test results. A new specimen should be used for testing at a higher load level to generate a failure point.

(d) Constant amplitude LCF failures are generated in less than 106 cycle range. Failures may be induced by increasing the number of cycles by a life factor on the applied load or by increasing the vibratory component of the applied load by a load factor, or a combination of both.

The steady component of load is generally not increased to induce failures. Subsequent data evaluation should account for the cycle or load increase required to generate the failures. Use load factors with caution, as the failure mode may change while testing at increased load.

(e) Spectrum load testing is often used to verify the damage accumulation algorithm for subsequent fatigue evaluations. Miner’s rule, with appropriate consideration for scatter factor, has proven to be suitable for calculating component life. A load spectrum includes operating conditions representative of the intended operation conditions for the propeller. Within this spectrum a high amplitude low cycle maneuver load should be applied periodically. Life factors to account for scatter should be applied instead of load factors, which may alter the failure mode.

For blades, the load spectrum may include elements of a typical flight consisting of the following:

1. Start-up;

2. Taxi out;

3. Run-up to take-off thrust;